Aphasia and Brain Organization

APHASIA and
BRAIN ORGANIZATION
APPLIED PSYCHOLINGUISTICS
AND COMMUNICATION DISORDERS
APPLIED PSYCHOLINGUISTICS AND MENTAL HEALTH
Edited by R. W. Rieber
PSYCHOLOGY OF LANGVAGE AND THOUGHT

Essays on the Theory and History of Psycholinguistics
COMMUNICATION DISORDERS
APHASIA AND BRAIN ORGANIZATION

Ivar Reinvang
A Continuation Order Plan is available for this series. A continuation order will bring
delivery of each new volume immediately upon publication. Volumes are billed only
upon actual shipment. For further information please contact the publisher.
APHASIA and
BRAIN ORGANIZATION
lvar Reinvang
Sumums Hospital mul
lustilulc of Psychology
UHiversi~; of Oslo
Oslo, Nmway
SPRINGER SCIENCE+BUSINESS MEDIA, LLC
Library of Congress Cataloging in Publication Data

Reinvang, lvar.
Aphasia and brain organization.
(Applied psycholinguistics and communication disorders)
Bibliography: p.
Includes index.
I. Aphasia. 2. Cognitive disorders. 3.Brain-Wounds and injuries-Complications
and sequelae. 4. Neuropsychology. I. Title. II. Series.
85-9545
6!6.85'52
RC424.7.R45 1985
ISBN 978-1-4757-9214-0 (eBook)
ISBN 978-1-4757-9216-4
DOI 10.1007/978-1-4757-9214-0
1985 Springer Science+Business Media New York

Originally published by Plenum Press, New York in 1985
Softcover reprint of the hardcover Ist edition 1985
All rights reserved
No part of this book may bc reproduced, stored in a retrieval system, or transmitted,
in any form or by any means, electronic, mechanical, photocopying, microfilming,
recording, or otherwise, without written permission from the Publisher
PREFACE
This book presents the work on aphasia coming out of the Institute
for Aphasia and Stroke in Norway during its 10 years of existence.
Rather than reviewing previously presented work, it was my desire
to give a unified analysis and discussion of our accumulated data.
The empirical basis for the analysis is a fairly large group (249 patients)
investigated with a standard, comprehensive set of procedures.
Tests of language functions must be developed anew for each
language, but comparison of my findings with other recent comprehensive studies of aphasia is faciliated by close parallels in test methods (Chapter 2). The classification system used is currently the most
accepted neurological system, but I have operationalized it for research
purposes (Chapter 3).
The analyses presented are based on the view that aphasia is
an aspect of a multidimensional disturbance of brain function. Findings of associated disturbances and variations in the aphasic condition
over time have been dismissed by some as irrelevant to the study of
aphasia as a language deficit. My view is that this rich and complex
set of findings gives important clues to the organization of brain
functions in humans. I present analyses of the relationship of aphasia
to neuropsychological disorders in conceptual organization, memory,
visuospatial abilities and apraxia (Chapters 4, 5, and 6), and I study
the variations with time of the aphasic condition (Chapter 8).
No study of aphasia is complete without an analysis of its clinicoanatomical basis. Testing the assumptions of the classical model
of aphasia, I can only partly confirm them. My analyses reveal that
V
vi
PREFACE
in many cases interachans between severallesion sites are important

in determining deficits that are often thought to have a more circumscribed clinicoanatomical basis.
In taking such a broad view of aphasia, my theoretical framework
has been influenced by concepts from general systems theory. The
theoretical chapters of the book (Chapters 1 and 9) present and develop
this type of approach sufficiently to account for the main aspects of
my findings and to suggest some new lines of investigation for the
future.
I should like to acknowledge the help and support of several
friends and colleagues. First of all, K. Sundet performed the statistical
analyses and discussed all the statistical problems with me.
P. Barenstein examined my CT scans and scored them in a standardized system. K. Willmes made available to me the system for analysis
of CT scans used in the Aachen aphasia laboratory. He also advised
me on problems of choosing appropriate statistical models. My wife,
T. Bjorg, did the artwork for the book. A special ward of thanks to
M. Taylor Sarno, who read an earlier version of this monograph and
gave me every help and encouragement to develop it for publication.
Finally, I must thank the Norwegian National Health Association
for their support of my work during several years, including the time
period during which the book was written.
IVAR REINVANG
CONTENTS
Chapter 1. Approaches to the Study of Aphasia

1.1. Clinical and Theoretical Approaches . . . . . . . .
1.2. Historical Approaches . . . . . . . . . . . . . . . . . . . . .
1.2.1. Localization Theory of Language-toBrain Relation . . . . . . . . . . . . . . . . . . . . . .
1.2.2. Criticism of Basic Premises . . . . . . . . . . .
1.2.3. Criticisms of Assumptions Regarding
the Nature of Aphasia or the Types of
Aphasie Disturbances . . . . . . . . . . . . . . .
1.3. Systems Theory Approach . . . . . . . . . . . . . . . . .
1.3.1. Applications of Systems Theory
Concepts in Neuropsychology . . . . . . . .
1.3.2. Localization of Function in Light of
Systems Theory . . . . . . . . . . . . . . . . . . . . .
1.4. The Present Study . . . . . . . . . . . . . . . . . . . . . . . .
1
5
5
6
9
11
13
16
19
Chapter 2. Operationalization of a Model

2.1. The Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.1. Broca Area . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.2. Posterior Language Area . . . . . . . . . . . . .
2.1.3. Arcuate Fasciculus . . . . . . . . . . . . . . . . . .
2.2. Aphasia Test Construction and
Standardization . . . . . . . . . . . . . . . . . . . . . . . . . . .
23
24
26
29
30
vii
CONTENTS
viii
2.2.1.
2.2.2.
2.2.3.
2.2.4.
2.2.5.
2.2.6.
2.2.7.
2.2.8.
Test Variables . . . . . . . . . . . . . . . . . . . . . .
Selection of Tasks . . . . . . . . . . . . . . . . . . .
Similarity to Other Aphasia Tests . . . . .
Choice of Normative Sampie . . . . . . . . .
Standardization . . . . . . . . . . . . . . . . . . . . .
Statistical Properties . . . . . . . . . . . . . . . . .
Relation to Background Variables . . . . .
System of Gradation . . . . . . . . . . . . . . . .
30
31
33
34
35
37
40
41
Chapter 3. Types of Aphasia

3.1. Classification System of the Norsk
Grunntest for Afasi . . . . . . . . . . . . . . . . . . . . . . .
3.1.1. Definitions of Speech Classification . . .
3.1.2. Definitions of Aphasia Types . . . . . . . . .
3.2. Subjects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3. Stability of Classification . . . . . . . . . . . . . . . . . . .
3.4. Comparison with Typology of W AB . . . . . . . .
3.5. Levels of Classification . . . . . . . . . . . . . . . . . . . .
3.6. Sampie Structure . . . . . . . . . . . . . . . . . . . . . . . . . .
3.7. Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
47
48
48
51
51
52
54
56
58
Chapter 4. Selective Aphasias

4.1. Types of Selective Aphasias . . . . . . . . . . . . . . . .
4.1.1. Modality-Specific Aphasias . . . . . . . . . . .
4.1.2. Material-Specific Aphasias . . . . . . . . . . .
4.1.3. Models of Selective Aphasias . . . . . . . . .
4.2. Studies of Sensory Mechanisms and
Language in Normals . . . . . . . . . . . . . . . . . . . . .
4.3.1. Material-Specific Deficits . . . . . . . . . . . . .
4.3.2. Modality-Specific Deficits . . . . . . . . . . . .
4.3.3. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . .
59
59
61
62
63
65
65
67
69
CONTENTS
ix
Chapter 5. Memory and Learning Deficits

5.1. Normal Memory . . . . . . . . . . . . . . . . . . . . . . . . . . 71
5.1.1. Verbal Memory . . . . . . . . . . . . . . . . . . . . . 72
5.1.2. Verbal Learning . . . . . . . . . . . . . . . . . . . . . 73
5.1.3. The Relation of Verbal Memory and
Learning to Language Function . . . . . . 73
5.2. Verbal Memory and Learning in Aphasics . . . 74
5.2.1. Verbal Memory . . . . . . . . . . . . . . . . . . . . . 74
5.2.2. Verbal Learning . . . . . . . . . . . . . . . . . . . . . 76
5.3. Nonverbal Memory and Learning . . . . . . . . . . 77
5.4. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
5.5. Present Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
5.5.1. Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
5.5.2. The Structure of Memory in
Aphasia . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8~
5.5.3. Relations of Memory to Aphasia
Group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
5.5.4. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . 93
5.5.5. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . 95
Chapter 6. Defects of Visual Nonverbal Abilities

6.1. Visual Nonverbal Functions in Aphasia
6.1.1. Apraxia . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2.1. Tests of Nonverbal Abilities . . . . . . . . . .
6.2.2. Tests of Motor Function . . . . . . . . . . . . .
6.2.3. Apraxia . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2.4. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2.5. Discussion .........................
97
99
101
101
103
103
104
106
Chapter 7. Localization of Lesion in Aphasia

7.1. Status of the Localization Model
109
7.2. New Candidates for Status as
Language Areas . . . . . . . . . . . . . . . . . . . . . . . . . . 110
CONTENTS
70201. The "Limbic System" 0 0 00 0 0 0 0 00 0 0 0 00 0

702020 The "Lenticular Zone" 0 0 0 00 0 0 0 00 0 0 0 00
702030 Medial Structures 0 00 0 0 0 0 00 0 0 0 00 0 0 0 0 0
7030 The Present Study 0 0 00 0 0 00 0 0 0 0 00 0 0 0 00 0 0 0 00
70301. Method 00 0 0 0 0 00 0 0 00 0 00 00 0 0 0 0 00 0 0 0 00
703020 Results 0 0 0 0 00 0 0 0 00 0 00 0 0 0 0 0 0 00 0 0 00 0 0 0
703030 Analysis of Test Parameters 0 00 0 0 0 00 0 0
703.40 Lesions and Their Context 0 0 00 0 0 00 0 0 0
7.40 Conclusion 00 0 00 0 0 0 00 0 0 00 0 0 0 0 00 0 0 0 00 0 0 00 0 0
110
111
112
113
115
117
119
121
128
Chapter 8 Recovery and Prognosis

0
The Recovery Process 0 0 0 00 0 0 0 0 00 0 0 0 00 0 0 0 00 129

Recovery of Nonverbal Functions 00 0 0 00 0 0 0 0 0 131
Prognosis 0 0 0 00 0 0 0 0 00 0 0 00 0 0 00 0 0 0 0 0 00 0 0 00 0 0 131
Mechanisms of Recovery 0 0 00 0 0 0 0 0 00 0 0 0 00 0 0 133
8.401. Relearning or Facilitation 0 0 0 0 00 0 0 00 0 0 133
8.4020 Reorganization of Function 0 0 0 0 00 0 0 0 0 134
8.4030 Release of Vicarious Neural
Structures and Functional Relocalization 134
8.4040 Complementary Redifferentiation of
Function 0 0 0 0 00 0 0 00 0 00 0 0 0 0 0 0 00 0 0 0 00 0 135
8050 The Present Study 0 00 0 0 00 0 00 0 0 0 0 0 00 0 0 00 0 0 0 136
80501. Recovery Pattern 0 00 0 0 00 0 0 0 0 0 0 00 0 0 00 0 136
805020 Prognosis 0 0 0 0 00 0 00 0 0 00 0 0 0 0 0 00 0 0 0 00 0 141
805030 Relations between Fundions in Acute
and Chronic Patients 0 00 0 0 0 0 0 00 0 0 0 00 0 144
805.40 Conclusions 0 0 00 0 0 00 0 00 0 00 0 0 0 00 0 0 0 00 145
801.
8020
8030
8.40
Chapter 9 The Organized Response of the Brain to Injury

0
901. Evidence for Organized Complexity 00 0 0 00 0 0 148

9020 A Proposed Systemic Model 0 0 00 0 0 0 0 00 0 0 0 00 149
90201. Abstract Model 0 0 00 0 00 0 0 0 0 0 0 00 0 0 00 0 0 149
CONTENTS
xi
9.2.2. Neural Model ....................... 151

9.2.3. Clinical Evidence on Hemispheric
Relationships . . . . . . . . . . . . . . . . . . . . . . . 152
9.2.4. Within-Hemispheric Specialization and
Differentiation in Humans . . . . . . . . . . 154
9.2.5. The Effect of Lesions and the Systemic
Basis of Recovery . . . . . . . . . . . . . . . . . . . 155
9.3. Testing the Model ......................... 158
9.3.1. Application of the Model to the Present
Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
9.3.2. General Applications of the Model .... 160
9.4. Concluding Remarks ...................... 161
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
Index ..................................................... 193
APPROACHES TO THE
STUDY OF APHASIA
1.1. Clinical and Theoretical Approaches

The study of aphasia may be motivated by clinical as weil as theoretical
considerations. It has been estimated that about 1 million people suffer
from aphasia in the United States (Sarno, 1980). In Sweden, the incidence of aphasia has been estimated at 60 per 100,000 inhabitants per
year (Broman, Lindholm, & Melin, 1967), and in Norway, Petlund
(1970) estimated the prevalence at .09%. The most frequent cause of
aphasia is strake, which is itself a common disease in an elderly
population. Whereas the risk of strake in the fifth decade of life is
.2%, the corresponding risk in the seventh decade is 2.0% (Marquardsen, 1969). Add the fact that 20% to 25% of strake patients are
initially aphasic (Brust, Schafer, Richter, & Bruun, 1976), and the
magnitude of the clinical problems becomes striking. In this context,
the need for practical and reliable methods of testing is apparent. A
classification system with knowledge of associated neurological and
neuropsychological deficits, prognosis, and underlying pathology is
a prerequisite for sound treatment.
From a theoretical point of view, aphasia has, since the time of
the founding papers of Broca (1861) and Wernicke (1874), presented
a unique opportunity to study the relationship of the brain to higher
mental functions. The theoretical problern may, however, be
1
CHAPTER 1
fonnulated in different ways, and different methodological approaches

may be chosen.
A clinically based research strategy includes the following steps:
1. Observing naturally occurring associations and dissociations
of symptoms. These groupings are referred to as syndromes, with
the understanding that they are clinically useful "fictions."
2. Distinguishing between theoretically meaningful associations
(phenomena associated because they reflect the same function) and
theoretically meaningless associations (associations produced by theoretically uninteresting combinations of functions). The methods used
include psychometric analysis, post hoc control of the lesion variable
(as the size of the lesion is believed to be the most significant factor
in producing spurious associations of symptoms), and experimental
control of the task variable. The result is a structural analysis of the
function involved.
An alternative research strategy is based on assumptions about
the nature of language in the normal case.
It is fair to say that, recently, theoretically oriented efforts have
had the goal of analyzing (decomposing) the cognitive-linguistic process into constituent subfunctions and assigning neural correlates to
these subfunctions. In order to attain the goal of accounting for processing, the internal computational steps of subfunctions and their
ordering must be specified for a given type of task.
Progress in linguistics has led to models of the subcomponents
of the language function, and to possible rules for relating linguistic
symbols to each other. The structural school of linguistics has influenced aphasiology through the works of Jakobsan (1971), whereas in
more recent times the transformational generative model of linguistics
presented by Chomsky (1965) has been influential.
The tenn neurolinguistics (Hecaen & Dubois, 1971; Whitaker, 1971)
stands for an interdisciplinary study of aphasia based on neurology
and linguistics.
In an influential paper, Arbib and Caplan (1979) argued that
neurolinguistics must make an effort to give a computational account
of processing and that this can be approached by converging efforts
of neurolinguistics, psycholinguistics, artificial intelligence, and
neurophysiology.
In the summary of Caplan (1982), the first steps are
APPROACHES TO THE STUDY OF APHASIA
1. The Ievel at which the nature of computation is expressed. With

respect to human language, Marshall suggests that generative transformational theories of grammar provide a characterization of the structures
relevant to language, that is, a characterization of the features of the
mental object attained.
2. The Ievel at which algorithms that implement a computation are
characterized. Marshall suggests that work on parsing strategies, both
implemented and based on the results of psychological experimentation,
provides an example of the beginnings of a characterization of the psychological steps which are operative in the attainment of the linguistic
structures of Level 1.
3. The Ievel at which an algorithm is committed to particular mechanisms, which has been "the traditional preserve within psycholinguistics
of the aphasiologist." (p. 423)
Clinical and theoretical motivations have been closely wedded

in the history of the study of aphasia. It has been assumed that
theoretical inferences could be drawn with confidence on the basis
of clinical observation of the association and dissociation of phenomena, and that the syndrome is a significant unit for theoretical analysis.
Only recently has the closeness or fruitfulness of this alliance been
questioned. According toMarshall (1982),
there will be some models ofbrain organization within which the demands
of clinical diagnosis and theoretical understanding pull in diametrically
opposed directions. (p. 404)
For the neurolinguist, the unit of analysis is language, and

assumptions about a language function are independently motivated
from studies in linguistics and psycholinguistics. It does not follow,
however, that aphasia in toto or subsets of aphasia phenomena are
wholly interpretable as a failure in subfundians or processing stages
of the language function. lt is interesting to follow the increasing
divisions of clinical and neurolinguistic studies. In the 1970s there
was optimism that the major clinical syndromes of aphasia could be
given a neurolinguistic analysis referring to breakdown in major blocks
of linguistic subfunctions (Caramazza & Bemdt, 1978). The more recent
attitude is that only selected aphasic symptoms, including agrammatism and some forms of dyslexia and agraphia, can be usefully
studied, and then preferably in selected cases with '1 pure" defects.
Taking the stand that there are only two approaches to aphasi(lthe clinical, which takes the patient as an unanalyzed whole as its
unit of study, and the theoretical, which takes the language function
CHAPTER 1
and nonclinical models of it as its units of study-grossly disregards

a third approach.
This approach may be termed neuropsychological and takes the
syndrome as its unit of analysis. It follows the step of the clinical
research procedure as outlined above. When the outlines and divisions of a functional domain have been established with gross neurological correlates, then one of two options may be chosen. One is
to say that this is as far as one can get in clinical group studies, and
selected patients with more specific deficits offer the only opportunity
to advance the study by clinical material or methods. Another option,
however, is to say that syndromes are meaningful indicators of the
multidimensional response of the brain to localized injury. They are
indications of the organizing principles at work in the efforts of the
whole brain to maintain optimal functioning, as much as they are
indications of the contributions of the missing parts.
Although neurolinguistic analysis is strong on detailed functional analysis and specification of processing, it has difficulty justifying its selection of study material and its notion of relevant case. We
have no theoretical metric for measuring the "pureness" of functional
deficits, and it may well be that sharply delineated behavioral symptoms are the consequence of highly complex functional interactions.
The findings from analyses of pure deficits can be applied to more
complex cases only if one assumes that pure deficits can be concatenated without giving rise to strong interaction effects. (See Shallice,
1979, for a discussion of the problems and advantages of single-case
studies.)
Neuropsychological analysis is weak on detailed function-and
process analysis. The existence of some alleged syndromes may be
questioned on empirical grounds (e.g., the criticism of Benton, 1961,
of the Gerstman syndrome). Neuropsychological analysis has the
advantage of not requiring an independently motivated theoretical
model and provides much of the necessary framework for more
sophisticated analysis by giving a normative background for gradation
of performances. If the phenomena under study are interactive, then
syndromes may reflect the emergent properties of factors combined
in a larger system, and a description of the relationship of the variables
for different parametric values is a necessary part of a complete theoretical analysis.
1.2. Historical Approaches

The present conceptions of aphasia date back to the continental
European neurological tradition before and around the turn of the
century. This tradition has a main stream, represented by Wernicke
(1874) and Lichtheim (1885) and several tributaries with supplementary approaches (Marie, 1906; Jackson, as summarized by Head, 1915,
1926; Goldstein, 1948; Luria, 1970; Jakobson, 1971; Hecaen & Dubois,
1971). To a surprising degree, the mainstream of thinking around the
turn of the century is still a dominant mode of thought (Benson &
Geschwind, 1977). Because all classical thories are centered on concepts of localization of function, it is useful to give a more general
characteristic of localization theory before discussing the controversies.
1.2 .1. Localization Theory of Language-to-Brain Relation

No author can be taken as the foremost representative of localization theory. The following is the present author's summary of the
essential features of the theory implied by authors who use terms like
"language area" or "speech center" to describe the neurological basis
of the language functions:
1. The brain contains areas with specialized functions, beyond
the sensory and motor areas. Normally, one cerebral hemisphere
contains all the structures necessary and sufficient for language. This
hemisphere is said tobe dominant (for language). Normally, the left
hemisphere is dominant, but in some instances, these structures may
be distributed between the hemispheres or may be located entirely
in the right hemisphere.
2. Within the dominant hemisphere, there is also specialization,
so that some areas are of critical importance to the language function
and some are not. The structures necessary for language (language
areas) are commonly believed tobe cortical, and tobe located in the
temporal and frontal Iobes. There are, however, different theories
and formulations of which specific areas are important and how far
their functions are differentiated.
3. Different parts of the language areas are specialized for different functions. Differently localized lesions in the language areas
give rise to varied clinical syndromes. By focusing on the features
CHAPTER 1
that show the most consistent relationships to the locus of injury, a

definition of types of aphasia can be given. It is not assumed that all
pathological performances in aphasia show a lawful relationship to
the locus of injury. Alternative classifications built on principles other
than clinicopathological correlation may be chosen but would have
to prove their advantage for special purposes. Again, there are different alternative formulations of which are the major and minor
aphasic syndromes and what is their specific relationship to the locus
of injury.
4. Language areas have fiber connections with one another and
with other areas. In the classicallocalization theories, these Connections are believed to have very simple functions of transmitting stimuli, thereby triggering the activity characteristic of the area receiving
the stimulus. More complex information on the results of previous
stages of analysis may also be transmitted, thereby "adding" or integrating the activity of several connected areas before a motor response
is emitted. This simple conception of the functioning of connecting
fibers has led to their being named association fibers and to their areas
of convergence being called association areas. Although it has not been
done in classical localization theory, it is entirely possible, without
abandoning localization theory altogether, to explore the hypothesis
that association fibers have more complex functions than believed.
The localization theory has been criticized both on general
conceptual grounds and with respect to some of its more specific
statements about the nature of aphasia and the types of aphasic
disturbances.
It is worthwhile to pause and note that none of these criticisms
question the existence of a correlation between the type of aphasia
and the locus of the lesion. Even authors often identified as antilocalizationalists, like Jackson (see Head, 1915), Marie (1906), Head
(1926), and Goldstein (1948), never denied the existence of clinicopathological correlation.
1.2.2. Criticism of Basic Premises

The criticisms most often advanced may be summarized under
three points:
1. It is impossible or unacceptable to try to localize normallanguage, a criticism stated forcefully by Jackson (see Head, 1915).
2. The mixture of behavioral and neurological terms of classifications is ill-conceived and confusing. This criticism, too, is
closely connected with the work of Jackson (see Head, 1915).
3. The general form of the theory (connectionism or associationism) is outdated and has been shown tobe inadequate.
Both Head (1926) and Pribram (1971) have stated this argument forcefully.
Regarding the alleged nonlocalizability of normallanguage, it is
appropriate to stress the difficulty of using observational clinical data
as a basis for inference about normal processes. In particular it is
unwise to name "centers" on the direct basis of lesion locus and
symptom description. This is no more than to say that phrenology is
outdatedas a model of neuropsychological research. On the basis of
observation that patients with certain lesions have difficulty in naming
objects, we would be unwise in inferring that the locus of the lesion
is normally the locus of object names. But assume that characteristics
of this naming difficulty can be teased out further by experimental
variation of conditions and can be shown to deviate from normal
performance by certain parameters. We would have then a basis for
hypothesizing an underlying process, which can then, again hypothetically, be related to a given neurological structure. The hypotheses
may very well have implications that could also be tested on normal
individuals by means of behavioral measures.
The criticism is correct if it is reformulated to say that no hypothesis assigning normal processes to given neurological structures should
be accepted on clinical evidence alone. The declaration that language
cannot be localized, however, seems to be an arbitrary conceptual
decision that any function with a definite relation to a neural locus
cannot be called language.
Mixing behavioral and neurological classification was called
"psychoneurology" by Jackson (see Head, 1915). There is danger of
tautological reasoning if concepts from one category are used to define
those from another. If frontal aphasia is defined as the type of aphasia
resulting from frontal injury, then the question of the frontal localization of this syndrome has already been settled by definition.
CHAPTER I
However, if care is taken to define behavioral categories in behavior

terms and neurological categories in anatomical terms, then there
should be no objection to studying the relationships between the two.
With the advent of more sophisticated statistical techniques, the
question of the behavioral validity of aphasia types can be raised: Are
there naturally occurring clusters of aphasia symptoms, and if so, do
they correspond to the classically described types of aphasia? It must
be recognized that the answer to this question depends on the patient
group studied. The agent of injury may be such as to produce diffuse
lesions, as in metabolic or anoxic lesions. It may also produce discrete
lesions, but of several anatomically distinct structures not known to
have a common function. This combination of lesions may weil occur
in cerebrovascular disease, where structures may be damaged together
by virtue of having a common blood supply. In penetrating head
injuries, again, the lesion may be discrete and weil defined but may
not follow the demarcation lines drawn by anatomy. Rather than
injuring one well-defined anatomical structure completely, it may
incompletely injure three.
Modern statistical studies started with Weisenburg and McBride
(1935) and continued with Jones and Wepman (1961) and Schuell,
Jenkins, and Carroll (1962). All these studies rejected classical classification schemes but are open to the criticism of Iack of control of
localization of the lesion. Recent studies by Goodglass and Kaplan
(1972) and Kertesz and Phipps (1977) indicate that an extension and
refinement of classificatory schemes within the framework of a classical clinicopathological model are a likely development.
I agree wholeheartedly with the critics of classical localization
theory that associationism is an inadequate theory for explaining the
complex activity of the nervous system. Hughlings Jackson was aware
of this point. From his studies of epilepsy, he described a certain dass
of symptoms as "release" symptoms, that is, symptoms caused by a
loss of inhibition. In his hierarchial model, the alleged loss of the
propositionallevel of function and the emergence of autornahe speech
are the primary example ofthistype of deficit in aphasia. In modern
times, several authors, among them Pribram (1971), have rejected
associationism and have proposed more complex theoretical models.
As noted above (p. 6), localization theory does not presuppose
associationism, although the two theories have been closely linked
historically. The inadequacy of an associationist model of the brain,

tagether with the criticisms discussed above, should not lead to abandoning the concept of localization of function. Localization theory
should, however, be modified and modernized. The specific content
of an adequate theory is largely unknown. It is, however, of some
interest to discuss what general features an adequate theory must
have.
1.2.3. Criticisms of Assumptions Regarding the Nature of Aphasia

or the Types of Aphasie Disturbances
The following criticisms will be discussed:
1. Language cannot be distinguished from intelligence. Aphasia
is symptomatic of a more general intellectual disturbance

(Marie, 1906; Bay, 1962).
2. Different forms of aphasia do not exist; only aphasia with
different additional disturbances exists (Bay, 1962; Schuell,
Jenkins & Jimenez-Pabon, 1965).
Jackson (see Head, 1915) proposed that language is integrated
in several levels of mental functioning. Aphasia is not a disturbance
of an anatomically localized language mechanism or process; rather,
it reflects mental regression from a "propositional Ievel of functioning
to lower levels. Speech in emotional context is preserved, but propositional speech is lost. Other influential thinkers supported this position. lt was adopted by Head (1926), who echoed Jackson's statement
that an aphasic is in a certain sense "lame in his thinking, and by
P. Marie (1906}, who said that aphasia is "a special sort of intelligence
defect. In modern times Bay (1962) has been a strong advocate of
the view that a conceptual disturbance is inherent in aphasia.
The position of Wernicke (1874) on this issue was clear:
11
II
II
The spoken and written narne of an object is not a new attribute of

the object. It is thus clearly different frorn the actual sensory rnemory
irnages of the object. Only the latter make up the concept of the object.
Disturbance of the concepts of things with which we deal in the process
of thinking are always disturbances of intelligence. Disturbances of speech,
on the contrary, cause difficulties only in the use of the conventional
means of representation of the concepts. (p.63)
10
CHAPTER 1
The consequence of this issue for research seems to be to question whether a consistent defect in "intelligence" can be found in
aphasics. If so, it is necessary to postulate an inherent link between
thought and language beyond the plausible assumption that the language disturbance makes an instrument for thought less available. A
way of demonstrating a defect of intelligence is to show that, given
a defect in performing a task with language material, it is possible to
demonstrate the defect even if the verbal elements of the task are
removed.
The available research on hemispheric asymmetry only partly
supports the notion of material-specific functions of the two hemispheres (Milner, 1974; Gazzaniga & Ledoux 1978), and differences in
the cognitive mode of operation of the two hemispheres must be
considered. (For review, see Bradshaw and Nettleton, 1981.) Research
on intelligence in aphasia, summarized in Lebrun and Hoops (1974),
indicates some reduction in specific nonverbal tasks, but the role of
the size of the injury in explaining such defects is uncertain. The
evidence on the issue is not strong enough to lead us to abandon the
theory of localized language function. The facts and their interpretation are discussed further in Chapter 6.
The second challenge to localization theory and the clinicopathological model is the question of whether different types of aphasia
exist. The position taken by antilocalizationists is that different syndromes exist after differently localized lesions, but they should not
be called different forms of aphasia. They should ratherbe seen as
aphasia with different, added disturbances. Marie (1906) stated that
Broca aphasia is the combination of aphasia and anarthria. This is the
holistic interpretation of aphasia, which has also been popular in
modern times through the work of Schuell et al. (1965). There is no
doubt that, in aphasia, variations in performance can often be observed,
so that some patients have disproportionate difficulties with speaking,
writing, reading, or auditory analysis. Sometimes such variations
determine the classification of the type of aphasia. Whether or not
such disturbances of performance should be called disturbance of
language is partly a conceptual question. Benson and Geschwind
(1977) defined language as "perception of verbal sensory stimuli, integration of these stimuli with prior knowledge, and activation of verbal
11
response-meehanisms" (p. 2). This definition obviously allows variation in performanee with different sensory modalities or response
modes to be classified as language disturbanees. Opponents would
presumably restriet their definition of language to a eognitive meehanism and would exclude pereeptual and response faetors. Data from
aphasiology ean eontribute to a resolution of this eoneeptual question.
If it ean be shown that speeifie neural cireuits exist for programming
speeeh or analyzing language, then it would seem natural to Iet the
definition of language, at least from the physiologist's point of view,
include the funetion of those cireuits. If, on the other hand, auditory
language pereeption eannot be distinguished from auditory pereeption in general, and programming of speeeh eannot be distinguished
from programming of other eomplex motor behaviors, then there
seems little reason to include pereeptual and response meehanisms
in the definitions of language. Rather, they would have tobe viewed
as tools for implementing language. This diseussion is taken up in
Chapter 4.
The evidenee from psyeholinguistics, from the work on speeeh
pereeption and diehohe listening (Studdert-Kennedy & Shankweiler,
1970), and from language pathology in eonnection with dyslexia (Marshall & Neweombe, 1973) seems to favor a view of specialized pereeptual meehanisms for language. The holistie or purely eognitive
eoneeption of language does not find support in these studies.
1.3. Systems Theory Approach

Generalsystems theory (von Bertallanfy, 1948) views organisms
as organized wholes. It is antireductionistie in the sense that it agrees
with the traditional Gestalt slogan "The whole is more than the sum
of the parts." On the other hand, it is not holistic in a traditional
sense. It regards analysis of wholes into eonsistent parts and their
mutual relationship as a fruitful undertaking. It does aim to transeend
and integrate the traditional positions of reductionism versus holism.
General systems theory has found applieations in several scienees,
among others in brain scienee.
12
CHAPTER 1
Conditions for considering a systems theory type of analysis are

present when there is evidence of organized complexity, that is, a
systematic relationship between many factors without simple one-toone correspondences. It is especially suited to a situation in which
different phenomena mutually interact without being ordered in a
causal chain. Weiss (1969) suggested statistical criteria for the presence
of systems effect, one criterion being that, in a system, the variance
of the whole is less than the sum of variances of the components,
indicating that coupling of components constrains the output of the
ensemble. Denenberg (1978) advocated the use of systems theory
when investigations based on ANOVA designsshowprominent interaction effects and an absence of main effects.
Systems theory may be applied to the neural Ievel of analysis
by considering the possibility that developing or maintaining the function of a piece of neural tissue is dependent on interactions with other
units of neural tissue.
Systems theory may also be applied to the behaviorallevel when
the influences of several factors are highly interactive, so that a decomposition and parceling out of causative factors becomes less meaningful than describing the different possible functionaJ states of the
system and their necessary conditions.
Von Bertallanfy (1948) introduced some general concepts that
will be of use in a further discussion of brain function. The brain
shares with other living systems the property of openness, that is,
the property of relating to an environment. The brain (and other open
systems) tends to continually reorganize the relations of its components to achieve more optimal functioning, and this process may be
under predominantly primary regulation (determined by the structure
of the brain itself) or under secondary regulation (determined by feedback from the environment). Development may go in the direction
of less interaction and greater degree of independent functioning of
components, called segregation. This is accompanied by a process of
mechanization in which the functional mode of a component becomes
fixed.
An advantage of systems theory is its suitability to conceptionalizing and organizing complex data. It thus has considerable heuristic
value in structuring a complex field of study. Its disadvantages are
~ts emphasis on description rather than explanation and the ensuing
13
difficulty of deriving testable predictions. General systems theory is

not a testable theory but only a framework for developing more detailed
theories within a specific field.
1.3.1 Applications of Systems Theory Concepts in Neuropsychology

A comprehensive review of this topic is beyond the scope of this
chapter. Few authors refer explicitly to systems theory itself, and very
many use concepts that are related to systems theory in one way or
another. Here I focus especially on statements emphasizing the
dynamic, interactive nature of brain function and the nature of symptoms as an organized response of the whole brain after injury.
One context in which the dynamic nature of function-tolocalization relationships has been much debated is the study of the
ontogeny of the cerebrallateralization of the language function. Lenneberg (1967) summarized the evidence and concluded that language
is gradually lateralized after being initially bilaterally represented.
Since then, it has been recognized that this view is overstated. There
is evidence of early specialization of the left hemisphere for language
(Dennis & Whitaker, 1977), and for several aspects of language, there
are no indications of continuing lateralization. As pointed out by
Moscowitch (1977), current tests may tap only very low levels of
linguistic processing, and "whether higher order linguistic processes
do indeed become progressively more lateralized with age is open to
debate" (p. 204).
Seines (1974) gave an extensive review of the role of the corpus
callosum in establishing hemispheric specialization. With poorly
developed cortico-cortical Connections, reciprocal specialization
develops only so far as structural asymmetries allow. With callosal
agenesis, there is an indication of greater likelihood of bilaterallanguage representation and a higher incidence of retardation in language development. The development of hemispheric specialization
for language in normals may be related to relatively late myelinization
of callosal fibers (Yakovlev & Lecours, 1967). Seines (1974) commented:
The evidence in support of the view that CC (corpus callosum) may
be responsible for the establishment of language lateralization is thus not
strong, but at least it has the advantage of being relatively easy to confirm
or disconfirm. This view differs from the "traditional" inhibitory theory
14
CHAPTER 1
in that the latter ascribes to the CC the roJe of a more or less permanent
mediator of inhibitory influences from the dominant hemisphere, while
the present theory views the CC as instrumental only for the establishment of language lateralization. Once this has been accomplished, there
should no Ionger be any need for inhibitory influences. This view does
not exclude, of course, that the CC is functional in transfer of information,
in particular visual information, and learning between the two hemispheres, and also in securing mental unity. (p. 132)
This distinction drawn by Seines between establishing and maintaining language lateralization is important in explaining differences
between the effect of lesions on mature and immature nervous systems.
In the mature nervous system, patients with callosal section
preserve left-hemisphere lateralization of speech control. It has been
noted, however, that the right hemisphere in these patients has good
comprehension of auditory verbal stimuli. This finding is in apparent
contrast to the global aphasia with poor comprehension resulting from
massive left-hemisphere injury in stroke patients. It may therefore be
that, even in the mature nervous system, some change in preserved
tissue (reduced differentiation) takes place with loss of callosal input.
Denenberg (1981) reviewed the evidence for hemispheric specialization and differentiation in animals. Numerous examples of differences in the effects of left- and right-sided injuries can be cited. Of
even greater significance in the present context is the evidence that
the two hemispheres are systemically coupled, that is, that the function of an intact brain is not simply the sum of activities in the two
isolated hemispheres. Some of the actions of one hemispherie on the
other are inhibitory. Add evidence that hemispheric specialization is
sensitive to early experience, and the need for assuming a dynamic
component both in establishing and maintaining hemispheric specialization is strongly supported:
lt is hypothesized that homologaus brain areas and their connecting
callosal fibers must be intact at birth, and must be intact throughout
development for Iateralization to reach its maximum Ievel. If there is either
hemispheric darnage or callosal darnage the brain will be less specialized
with respect to hemispheric differences. The hypothesis specifies two
homologous brain areas and their connecting fibers as the "unit" for the
development of lateralization. This is based on the assumption that such
a unit will act to maximize neural heterogeneity (i.e., Iateralization) because
of hemispheric competition. (Denenberg, 1981, p. 18)
Brown (1979) was concerned with the mechanisms of symptom

formation and of recovery and concluded that
15
in pathology Ievels in language production appear as symptoms. A symptom reveals a stage in language production that is traversed in the realization of the normal utterance. Brain darnage has the effect of allowing
symptoms-contents from more preliminary levels--to come to the fore.
There may also be a regression to a more preliminary Ievel. Accordingly,
a brain lesion does not disrupt a mechanism or a center where that mechanism is situated. Rather, it involves that structurallevel through which
the (pathological) content is normally elaborated. (p. 141)
The idea of a small set of Ievels that organize the basic phenomena
of perception, action, reaction, and language is clearly related to earlier ideas of a hierarchical organization of the brain (Jackson, 1878,
on propositional vs. autornahe language). In the model of Brown,
Ievels are also hierarchically organized, and the output of one is the
input into the next:
There is a resubmission of ernerging abstract content at each hierarchical Ievel to the same reiterated process-in other words, one process
at multiple Ievels, rather than multiple processes at the same Ievel. (Brown,
1979, p. 142)
This is called a micro-genetic process.

Kinsbourne advanced the concept of functional cerebral space
(Kinsbourne & Hicks, 1978), which makes the assumption that the
brain is a highly linked neuronal network:
The programming of a particular continuous activity involves not only
the cerebral locus at which the programming is accomplished, but also
involves, by spread of activation, a !arge proportion of the total cerebral
space, the amount occupied being greater the more closely the operator's
performance approximates the maximum of which he is capable. (p. 346)
These authors explored the dimensions of functional cerebral space

by studying the performance of concurrent tasks:
According to this model, if a single cerebral programme is being developed, the completed programs are facilitated to a greater extent at functionally closer loci than at functionally distant loci and, successively, transfer
of training is greater to the closer locus. In contradiction, unrelated motor
programs can be run concurrently most effectively if based on neural
activity in loci functionally remote from each other. (p. 347)
The concept has been used in studies of developing lateralization

of language by examining the interference between speech and leftor right-handed activities. This research has concluded that lateralization of spoken language in right-handers is established before 3
years of age (White & Kinsbourne, 1980).
CHAPTER 1
16
The insistence on the interactive, dynamic properties of neural

functions is in agreement with a systems theory approach. This
approach is further developed in an eloquent statement of his position
by Kinsbourne (1982):
There are no discontinuities in the brain. No independent channels
traverse it, nor is its territory divisible into areas that house autonomaus
processes .... No Simulation of human behaviour, however impressively
successful in impersonating its model, is capable of revealing how the
human mind arrives at the same outcome, unless it is based on a network
mechanism. (p. 412)
1.3.2 Localization of Function in Light of Systems Theory

It is possible to classify neuropsychological theories according
to their position on the two dichotomies of localization versus nonlocalization and systemic versus nonsystemic. I will first describe the
dichotomy between nonlocalization and localization views and then
show how these views are modified by introducing the concept of
systemic functioning.
1. Nonlocalization, nonsystemic. The theory says that functions are
diffusely represented in a structurally relatively undifferentiated brain.
Neural networks have been described as an alternative to localized
functional centers. The alternative, however, seems hard to reconcile
with the highly specific structures and patterns of connections found
in modern neuroscience.
The original idea of the neuron network as a continuum of nerve cells
of Standard shape and isotropic (random or geometrically determined)
connectivity properties has all but disappeared from our image of the
centers of the higher animals. (Szentagothai & Arbib, 1975, p. 43)
Therefore, because of the known specifity and diversity of anatomical

structures in the brain, this theory cannot be maintained for the functioning of the brain as a whole. It may, however, be considered for
certain functions in relation to limited brain regions. The thesis that
lesions within the language areas give rise to aphasias of varying
severity, but not of varying type, exemplifies this proposition.
2. Localization, nonsystemic. This theory states that the brain has
highly specific and diversified anatomical structures with equally specific and diversified functions. The extreme example is phrenology,
17
which says that the brain is a collection of independently working

organs.
The advantage of classical clinicopathological theory over phrenology is that it adds the possibility of integrative action by postulating
connections between neural centers, as weil as the building up of
more complex functions by association. The theory is still nonsystemic, so long as the presence of connection does not modify the
operations of localized functional centers. The Wernicke-Lichtheim
model uses the concept of localized functions to describe and explain
loss of function but adds aphasia syndromes (conduction aphasia,
transcortical aphasias) caused by isolation or disconnection of language areas.
An even more sophisticated step in analysis is taken when localized centers are connected in temporal sequence and shifting combinations. These "functional systems" (see Luria, 1973) are seen as
underlying normal performances. They arestill not systemically organized in the sense of the present discussion, because the functioning
of a component is not modified by other components, barring the
special case of disconnection.
3. Nonlocalization, systemic. Although acknowledging the highly
interconnected nature of neural tissue, I have already rejected the
neural net as a sufficient model of the human nervaus system. As an
alternative to neural nets, Szentagothai and Arbib (1975) described
more modern concepts, based on the idea of "modules" of neuronal
organization. Although recognizing anatomical specificity, still the
similarity of neuronal building blocks, called modules, throughout the
cortex is stressed.
To make such a model systemic, it would be necessary to assume
that the pattern of interaction between neural elements (modules)
determines function, whereas this pattern can be set up anywhere in
the brain, or at least in the cortex. Although such formulations are
more theoretically acceptable than simpler concepts in the nonsystemic version of nonlocalization theory, the problern is to show how
localized injury to a brain thus organized could result in a differential
deficit in the language function, with relative sparing of other functions. Hence, the introduction of a systemic dimension does not make
nonlocalization theories better able to explain the empirical findings.
Although rejecting this alternative as a model for the brain as a whole,
18
CHAPTER 1
it may be that the language areas function in this way in relation to

some language functions. In this model, small lesions may lead to
slight or no defects, and lesions beyond a critical size may lead to a
nonlinear increase of severity of deficit in several functions.
4. Localization, systemic. Whereas a nonsystemic localization theory holds that the function of a given area is determined by its anatomical structure alone, systemic localization would mean that the
function of a given area is determined by its relationship to other
areas within the bounds determined by its anatomical structure. This
qualification is important in view of the well-known high degree of
structural specificity of different anatomical regions in the brain. lt
may be, however, that given a ground plan of anatomical structure
and connections, the pattern of function-to-structure relationship
becomes more highly specific with time. If the ground plan is disrupted by injury, then some degree of rearrangement of the functionto-structure relationship may be possible. This flexibility in the development and the adjustment of the localization of function would have
to be attributed to the interplay between the anatomical structures
preserved at any given time. Such interplay must be mediated by
neural pathways. Accepting the idea of a systemic element in the
localization of function also entails a wider conception of the role of
neural connection than that which is postulated in associationist
schemes of brain function.
This model has the potential not only of explaining selective
deficits with localized injuries, but also of explaining change or recovery of function in patients with structurally stable brain lesions. To
my knowledge, no one yet has worked out a systemic localization
theory of aphasia. The work of Luria ( e. g., Luria, 1966), despite looking
at symptoms or performances as based on functional systems, does
not discuss systemic explanations of the phenomena of cerebrallocalization of function as such.
In a brain thus organized, the effects of lesions would be specific
according to locus, but not additive with composite lesions. The effect
of a lesion would be the result of a preinjury pattern of localization
and the nature of the systemic response to injury in preserved tissue.
In summary, then, the selectivity of deficits with differently
localized lesions isthebest criterion for adopting a localization theory,
whereas the additivity of effects in limited versus composite lesions
19
is the main clue to the systemic organization of cerebral representation. In nonsystemic organization, the additivity of effects is preserved, but not in systemic organizations. Finally, the degree of
changeability in the performance-to-structure relationship in recovery
can be used as additional relevant information.
1.4. The Present Study

What one undertakes as a research project is determined by one' s
interests, practicallimitations, and, most important, what one regards
as reasonably well established.
During the period of time in which the material for this study
was collected, the general services offered aphasic individuals in Norway were unsatisfactory. Speech therapy services were not organized
and were available only in large cities or communities. No counseling,
social support, or information pertaining specifically to the problems
connected with aphasia were given to families, and no training in the
care and treatment of aphasia patients was offered to hospital personnel. As a first step to improve this situation, the Institute for
Aphasia and Stroke was established in 1973 by a donation from the
National Health Association, a private organization with the fight
against coronary and other vascular diseases as one of its goals.
The Institute for Aphasia and Stroke is a test laboratory located
in the Sunnaas Rehabilitation Hospital, which is one of the municipal
hospitals of Oslo, Norway. The staff of the institute consists of one
neuropsychologist, one technician, and one research associate. The
hospital has 226 beds and admits patients with several kinds of functional disturbance with organic etiology, offering physiotherapy,
occupational therapy, speech therapy, and, in addition, social and
medical services. Patients may be admitted for evaluation only, or for
full treatment. The hospital offers only inpatient services; thus, the
patients admitted for treatment have severe physical handicaps,
whereas patients with lighter physical defects are referred to other
institutions with outpatient services.
Because of its unique position, the institution received applications for admission from the entire country during the period of
this study. An attempt was made to see as many as possible of the
20
CHAPTER 1
patients for evaluation and testing, so as to get a survey of the population referred for treatment. Although no exact figure can be given,
it can safely be stated that more than 90% of the patients referred
were tested.
The decision to create an aphasia registry was motivated by the
desire for a systematic registration of all available information pertinent to the description and evaluation of the patient group. The registry should serve primarily as an instrument for clinical research,
concentrating on the connection of aphasia with other symptoms and
on the development of aphasia with time. The results of tests performed at the Institute for Aphasia and Stroke form the main content
of the registry, with medical and general background information
added.
As reviewed above, a summary of the consensus in 1978, when
this study started, runs as follows:
The clinically defined syndromes of aphasia are stable entities
with a well-defined pathological substrate. Because aphasia is a linguistic deficit, a more refined linguistic analysis of language performances in the major syndromes will allow us to replace the static
traditional descriptions of functions as unanalyzed wholes with
dynamic processing concepts approaching the ideal of complete computational specification with neural correlates.
My difficulties with accepting the position just summarized were
based on both methodological and conceptual worries. First of all, I
worried about the loosely defined procedures for testing and defining
aphasic syndromes. A necessary first step for clinical research-and
a step that must be taken anew in each different language communityis to define strict and quantifiable procedures for testing and classification. The system of myself and my colleagues is described in
Chapters 2 and 3.
Second, I worried about the seemingly innocuous assumption
that aphasia is a linguistic deficit. Remernhering the papers by Teuber
and Weinstein (1956), by Weinstein (1964), and by others showing
an association of aphasia with some visual reasoning and learning
tests, as well as the many exiting papers by Kimura (see Kimura, 1979)
on the close association of language and higher order motor functions,
I thought it more appropriate to define aphasia at the outset as a
linguistic-cognitive defect. Although I in no way wish to question the
21
reality of a language function separate from other cognitive functions,

it may still be the case that aphasia does not reflect an isolated disturbance of this function. If it turnsout that only a few selected cases
demonstrate pure disturbances of language, then a scientific approach
to the great majority is needed. Maybe these cases can be viewed as
just "mixed" and can be explained as additions of defects observed
in isolation in the pure cases. But it may also be that important interactions are at work, so that mixed syndromes becomes a misnomer for
unanalyzed complexity.
Third, I worried about the generally simplistic approach to the
effect of brain injuries evident in a deficit-oriented analysis. It seemed
to me that the multidimensional response of the brain to injury, as
weil as the variations over time of this response, was what had to be
described and accounted for. I was (and am) disturbed by the tendency to stress regularity and to dismiss variability in the response of
the brain to injury as "noise." Commenting on the relative success
of syndrome classification, Wernicke (1874) stated:
Only a particular period in the course of the disease should be considered if one is to diagnose aphasia correctly. On the one hand, the
general phenomena which accompany the onset of aphasia, as they do
that of most localized lesions of the brain, must have disappeared. On
the other hand, however, the conditions ought not to have Iasted so long
that the possibility of compensation by the other hemisphere is already
present. (p. 69)
Here, the motivation of the neurological diagnostician to ignore information not pointing to the locus of the injury is clearly seen.
Poeck (1983b) echoed the same opinion:
lt cannot be denied that a certain number of vascular aphasias (approximately 15%) cannot be classified in terms of standard or nonstandard
syndromes. The main reason, in our experience, is that the examination
is done too early, prior to the establishment of a weil defined syndrome,
or at the late stage of recovery, with or without the effects of speech
therapy. (p. 80)
On the basis of these reflections, and having standardized the

necessary tools for measuring and classifying aphasic phenomena
(Chapters 2 and 3), I have therefore undertaken a broadly conceived
program of testing aphasics with neuropsychological tests. Information on lesions with CT-scans has been recorded when possible, and
repeated testing has been performed in order to chart as far as possible
22
CHAPTER 1
the extremes of parametric values in aphasia, conceived of as an

experiment of nature. I will attempt to give a systems-theory-oriented
account of the complexity of the aphasic condition. lt must be recognized that the account falls short of the goal of an experimental
analysis of causally significant factors, if such an analysis can be given.
lt also falls short of the ideal of a specification of the actual processing
stages behind the performances observed. I would still claim that, at
the very least, this type of analysis of the organized complexity of
linguistic-cognitive phenomena under a set of extreme conditions is
a valuable complement to other sources of information about the
underlying system.
OPERATIONALIZATION
OF AMODEL
2.1. The Model

The goal is to select a model that, tothebest of our current knowledge,
captures the significant dimensions of clinical syndromes. It should
also account for the associations of parameters that are useful for
defining syndromes while leaving reasonable space for withinsyndrome variations.
The Wernicke-Lichtheim model underlies the terminology and
the clinical classification systems most frequently used today. It is a
model within the localizationist tradition (Wernicke, 1874; Lichtheim,
1885), and it identifies two cortical areas important to the language
function, the Broca and Wernicke areas. The fibers associating these
areas are assumed to run in the arcuate fasciculus. In addition, Lichtheim assumed that transcortical fibers, via a hypothetical "concept
center," can mediate information between the language areas. Different forms of aphasia follow from lesions of different neurological
structures.
The localization of language areas is shown in Figure 2.1, and
the types of aphasia resulting from differently located lesions are
shown in Table 2.1.
A more detailed review of brain regions and the associated aphasia
types is given below. The review does not limit itself to statements
23
24
CHAPTER 2
Figure 2.1. Localization of language areas. Legend: (1) anterior language area (Broca);
(2) Wernicke area; (3) supramarginal gyrus; (4) angular gyrus.
by classical authors but intends to sketch the present-day status of

this model.
2.1.1. Broca Area

The Broca area is located in the third transverse frontal convolution, which can be divided into three parts: the orbital, the triangular, and the opercular. lt is the opercular part of the convolution
Table 2.1. Assumptions of the Wernicke-Lichtheim Model
Type of aphasia
Broca
Wernicke
Global
Conduction
Anomic
Isolation syndrome
Transcortical motor
Transcortical sensory
Locus of lesion
Broca area
Wernicke area
Broca and Wernicke areas
Arcuate fasciculus
Angular gyrus
Extensive neocortical, sparing Broca and
Wernicke areas
Frontal, sparing the Broca area
Parieto-occipital, sparing the Wernicke area
OPERATIONALIZATION OF A MODEL
25
that makes up the Broca center (Baily & von Bonin, 1951). This conclusion is corroborated by the results of electrical Stimulation during
local anaesthesia in operations for epilepsy (Rasmussen & Milner,
1975).
The Broca area is designated Area 44 in Brodman's classification.
According to recent anatomical evidence summarized by Galaburda
(1982), Area 44 can be distinguished from surrounding cortex and
represents an intermediate degree of architectonic differentiation
between premotor cortex and primary motor cortex. Galaburda cited
evidence that interhemispheric asymmetries can be shown for parts
of the frontal operculum. The Broca area, like surrounding frontal
and lorbitall cortex, has evolved out of a proisocortical zone located
in the anterior insular region, and it maintains connections with this
moreprimitive zone.
Wernicke (1874) believed that the Broca area receives sensory
inputs from the musculature. It has the function of storing memory
("images") of performed movements. These images can be aroused
via association fibers from other cortical areas, thus giving rise to
speech. In the later literature, there has been recurring controversy
about the importance of the Broca area for language. The controversy
is at least partly conceptual. Some would assign the Broca area a
purely motor function and name the effect of a lesion of the Broca
area anarthria (Marie, 1906). Others would assign to it a special role
in the programming of speech movements but prefer to dass the
resulting defect as a form of apraxia (Liepmann, 1915). Finally, some
would hold that the Broca area is essential for the activation of response
mechanisms in language, but that the deficits resulting from failure
can be distinguished from arthric and apraxic disturbances, and must
be properly classified as aphasic (Benson & Geschwind, 1977). Some
authors would deny that the Broca area has any function at all in
relation to language or speech (Pribram, 1971).
Broca aphasia is characterized by nonfluent speech, that is, speech
made up of poorly articulated short phrases produced with hesitations
and effort, particularly in initiation. Auditory comprehension is good,
but not completely normal. Ability to repeat and name is impaired,
but often better than the ability to produce words in spontaneaus
speech. Reading comprehension is relatively good, but writing is
always impaired. (Goodglass & Kaplan, 1972; Benson & Geschwind,
26
CHAPTER 2
1977; Mohr, 1976; Kerschensteiner, Poeck, Huber, Stachowiack, &

Weniger, 1975).
2.1.2. Posterior Language Area
The posterior language area is composed of parts of the temporal
neocortex, the gyrus supramarginalis, and the gyrus angularis. The
Wernicke area is defined in this monograph as the temporal part of the
posterior language area.
There seems to be a general agreement about the practical rule
of thumb followed by neurosurgeons that the anterior part of the
temporallobe up to Labbe's vein can be excised without dire consequences for the language function. In neuroanatomical terms,
Heschl's gyri are often given as the anterior Iimit of the Wernicke
area. In regard to the posterior, it is generally agreed that this area
is continuous with the supramarginal and angular gyri. The main
disparity between diverse statements and diagrams seems to be that
some regard only the superior temporal convolution as relevant to
language, whereas others include the middle, and some authors even
the inferior, temporal gyrus (Bogen & Bogen, 1976). The results based
on electrical cortical stimulation in local anaesthesia vary among early
reports (Penfield & Roberts, 1959), which seem to indicate a more
extensive area, and later reports (Rasmussen & Milner, 1975; Fedio
& van Buren, 1975), which find a more restricted area. I have adopted
the definition that the Wernicke area consists of the posterior part of
the superior and middle temporal gyri.
In the opinion of Galaburda (1982), the Wernicke area shows a
degree of anatomic differentiation, judged by architectonic criteria,
closely similar to that of the Broca area. He noted:
In fact, architectonic similarities between anterior and posterior language areas and the overlap in their connectional organization make it a
somewhat surprising finding that lesions in either region produce such
different aphasic syndromes. (p. 443)
It should be noted, however, that the studies referred to have

been performed on rhesus monkeys! The posterior language area
has evolved out of proisocortex located in the temporal and posterior insular region. It contains regions of varying cytoarchitectonic
27
differentiation, from primary auditory sensory cortex to more generalized neocortex, found in the inferior parietal lobule and the
temporo-occipital junction.
Wernicke (1874) believed that the Wernicke area is a store of
auditory ward images (Klangbilder). The condition after injury is therefore characterized by difficulties with auditory language perception
(total or partial ward deafness) and disturbances of speech (because
the appropriate auditory images for stimulating motor representations
are disturbed).
Geschwind (1979) stated:
Much new information has been added in the past 100 years, but the
general principles Wernicke elaborated still seem valid. In this model the
underlying structure of an utterance arises in Wernicke's area. lt is then
transferred through the arcuate fasciculus to Broca's area where it evokes
a detailed and coordinated program for vocalization. (p. 187)
Wernicke aphasia is characterized by fluent, paraphasic speech

and reduced auditory comprehension. Speech is produced without
effort and has complex grammatical structure. Informational content
is deficient (Goodglass & Kaplan, 1972; Huber, Stachowiack, Poeck,
& Kerschensteiner, 1975). The term jargon aphasia is sometimes used
for cases in which speech is totally incomprehensible, but jargon is
not confined toWernicke aphasia (Benson & Geschwind, 1977). Repetition is usually disturbed to the same degree as auditory comprehension, whereas naming performances may vary. Reading and writing
are usually severely disturbed, but in some cases, they are preserved
(Lecours & Rouillon, 1976).
The supramarginal gyrus is continuous with the superiortemporal
gyrus. Wernicke (1874) regarded the supramarginal gyrus as part of
a continuous perisylvian gyrus, anatomically and functionally continuaus with the superior temporal convolution. The stimulation data
seem to support this view (Penfield & Roberts, 1959; Rasmussen &
Milner, 1975). The reason for giving special consideration to this gyrus
isthat the probability of an auditory-language-comprehension defect
is markedly lower with a lesion of the supramarginal gyrus alone than
with a lesion of the superior temporal gyrus (Luria, 1970), whereas
the probability of reduced fluency of speech with misarticulation and
phoneme substitutions increases, possibly because of the proximity
to the primary somatosensory cortex. The possibility of finding a
28
CHAPTER 2
syndrome corresponding to conduction aphasia may therefore exist with

lesions of the supramarginal gyrus. Cytoarchitectonic studies by Galaburda, LeMay, Kemper, and Geschwind (1978) indicate that the part
of the supramarginal gyrus immediately adjacent to the superior temporal gyrus may belong to the auditory association cortex, whereas
more peripheral portions do not.
The angular gyrus is conventionally defined and is anatomically
continuous with the middle temporal gyrus. According to Henschen
(1922), only five cases with selective involvement of the angular gyrus
had been published up tothat time. The patients were all alexic and
agraphic but had no auditory comprehension defect. Lesions of the
angular gyrus occur often with more extensive involvement of the
posterior language area. Authors who use an extended concept of the
Wernicke area usually include the angular gyrus (e.g., Marie, 1906;
Dejerine, 1914; Penfield & Roberts, 1959). Wernicke (1874) accorded
no status to the angular gyrus in connection with reading and writing
but assumed that association fibers from the occipital lobe to the
Wernicke area were necessary for reading. The assumption of a special
importance of the angular gyrus for reading and writing is widely
adopted today. Some authors also accorded it a special function in
word retrieval (naming) (Geschwind, 1967b; Luria, 1970). Anomic aphasia
is characterized by fluent speech with marked shortage of content
words. There is little paraphasia as such, but there are attempts to
substitute circumlocutions and vague descriptions for content words.
Comprehension and repetition are good, but there are severe problems in reading and writing (Goodglass & Kaplan, 1972; Kertesz,
1979).
Alesion encompassing both Broca and Wernicke areas produces
a global aphasia. This may be regarded as a composite form of aphasia
that should not be classified as an independent type, but clinically it
has distinct features. It is characterized by severe loss in alllanguage
modalities, but the patient is usually not mute. Often, he or she has
verbal stereotypes, consisting of conventional phrases (swearing) or
meaningless syllabic combinations. In auditory comprehension, there
is also some ability to react to concrete words, particularly if they are
emotionally significant for the patient (Stachowiak, Huber, Kerschensteiner, Poeck, & Weniger, 1977).
29
2 .1.3. Arcuate Fasciculus
Very precise descriptions of this cortico-cortical fiber bundle are hard

to find. lt has a compact middle portion sweeping around the insula
parallel to the circular sulcus. The ends fanout and connect the inferior and middle frontal convolutions with large parts of the convexity
of the temporal lobe. The existence of a direct projection from auditory
areas to a homologue of the Broca area has been confirmed in rhesus
monkeys by Pandya and Galaburda (1980).
This fiber bundle is one among several structures assumed to
be of functional importance in connecting the posterior language area
with the Broca area. Wernicke believed the insula to have this function, but this possibility is now considered unlikely, and in the neoclassical literature, the arcuate fasciculus is accepted as the major
functional connection. Lesions of the fascicle result in a conduction
aphasia with relatively fluent speech and good comprehension, but
repetition difficulty. Fluency may be less than in Wernicke aphasia,
because the patient makes frequent attempts to correct literal paraphasias (phoneme substitution errors). He or she may go through a
series of approximations in attempting to correct his or her production
("zeroing in"). In addition to good auditory comprehension, there is
also often good reading comprehension (Benson, Sheremata, Bouchard, Segarra, Price, & Geschwind, 1973; Green & Howes, 1977;
Benson & Geschwind, 1977).
Lesions outside the language areas mentioned above may produce aphasia. This means not that the areas injured have language
functions, but that the language areas normally interact with surrounding areas when language is integrated in complex behaviors.
The essential feature of such lesions producing aphasia is that they
disconnect or isolate parts of the entire language areas from the surrounding cortex.
Total isolation produces an isolated-speech-area syndrome. The
patient has no spontaneaus speech but responds to questions. The
response ,is afmost fil~ways a direct repetition of the question. Speech
is weil aiticulated ~d tb:e patient repeats even long sentences . ' He or
she shows no sign of comprehension and fails in all language tests
except repetition (Geschwind, Quadfasel, & Segarra, 1968).
30
CHAPTER 2
Transcortical motor aphasia is characterized by an excellent ability

to repeat and a sparse, but well-articulated, speech. Comprehension
both of speech and of print is adequate, whereas writing is defective
(Rubens, 1976).
Transcortical sensory aphasia has fluent, paraphasic speech and
poor auditory comprehension. Unlike in Wernicke aphasia, however,
the ability to repeat even long sentences is preserved. Reading and
writing are usually severely defective (Kertesz, 1979).
2.2. Aphasia Test Construction and Standardization

In developing a suitable test methodology, the following consiJerations are important:
The test must cover the variables necessary and sufficient for
classifying patients into types of aphasia in the Wernicke-Lichtheim
version of a clinical pathological approach to aphasia.
The tasks selected for operationalizing these variables must be
similar to the tasks used by other investigators in the same tradition.
A system of gradation must be developed, so that a comparison
between individuals and between different performances within one
individual is possible. This comparison presupposes that scores with
acceptable statistical reliability are employed.
Some statistical justification must be given for grouping tests
under a common heading. The scores must be checked for their sensitivity to extraneous variables, and appropriate corrections for such
unwanted influences must be developed.
2.2.1. Test Variables

The variables necessary for classifying patients in the clinicalpathological tradition can be deduced directly from the description
of aphasia syndromes by classical authors (Wernicke, 1874; Lichtheim,
1885). The necessary variables are the following:
Fluency of spontaneous speech
Auditory comprehension
Repetition
31
Naming
Reading comprehension
Reading aloud
Writing
Only the first four variables are critical to classification. The
classical authors were uncertain about the frequency with which disturbances of reading and writing accompany the different aphasic
syndromes and about the mechanism producing them. From the point
of view of the controversy between "localizationists" and "holists,"
the inclusion of reading and writing among the basic variables affords
an opportunity for comparing sense modalities (auditory and visual)
and response modes (oral and graphic). All modern aphasia tests
include tests of reading and writing.
In 1973, it was decided to construct a test battery comprising the
above main variables for the purpose of classifying and grading aphasic
disturbances. This work resulted in the publication of the Norsk
Grunntest for Afasi (NGA) (Reinvang & Engvik, 1980b).
2.2.2. Selection of Tasks
A description of the type of tasks used to operationalize each
variable follows. For details of the procedure, the Appendix to this
volume must be consulted.
2.2.2.1. Spontaneaus Speech. It is evaluated in response to specific
questions ("What is your occupation?") and to open questions ("Tell
me about your family"). The interview is tape recorded.
Three aspects of speech are evaluated:
1. Communicative function (0-4 rating)
2. Qualitative disturbance (0-3 rating)
Litera! paraphasia
Complex paraphasia
Visible effort
Hesitations, pauses
Stereotypy
Dysarthria
Self-correction
3. Quantity of speech
32
CHAPTER 2
Words per minute (0-200)

Words per utterance (1,0-10,0)
Quantity of speech is scored on the basis of a transcript of the interview.
2.2.2.2. Auditory Comprehension. In the NGA, there are three
classes of stimuli: objects, body parts, and language material. There
are also three types of responses: pointing, complex acts, and yesno choice. They are combined in the following tasks (the number of
items is given in parentheses):
1.
2.
3.
4.
5.
6.
7.
8.
Body parts, point to the item named (1);

Body parts, point to the item described (6);
Body actions, carry out spoken instructions (10);
Objects, point to item named (11);
Objects, point to item described (6);
Objects, carry out spoken instructions (10);
Comprehension of ideas, respond yes or no (14);
Comprehension of relative statements, respond yes or no (4).
2.2.2.3. Repetition. The NGA includes the following types of tasks:

1. Word repetition (20)
2. Repetition of nonsense syllables (8)
3. Repetition of sentences (12)
Words are varied in terms of number of syllables, content (numbers and content words), and articulatory difficulty. These features
arenot scored separately. Nonsense syllables have been included as
being representative of extremely low-probability words. They vary
in number of syllables and stress pattern. Sentences vary in length
(number of words) and content. They include examples of sentences
loaded with function words (" Aldri annet enn om og men," meaning
"Never anything but if' s and or' s") and of low-probability nonsense
sentences ("Baaten sank i hytt og vaer," which can be translated as
"The ship sank as the wind blows").
2.2.2.4. Naming. The material used in the auditory comprehension test is used again in the naming test. The following types of
tasks are included:
1. Body parts, name (11);
2. Body actions, describe (5);
33
3. Objects, name (10);

4. Objects acted on, describe (5);
5. Responsive naming (10).
2.2.2.5. Reading and Sentence Construction. Printed stimuliareused,
with letters, words, and sentences tagether with some objects from
the test sample to investigate reading. Responses are oral (reading
aloud), matching, pointing, or performing an action. Stimuli and
responses are combined in the following tasks:
1. Reading comprehension, tested with recognition (6), word rec-
ognition (6), word-object matehing (6), and printed instructions (5).

2. Reading aloud, tested with letters (6), words (10), and sentences (5 items, scored 2-1-0).
3. Syntax, sentence construction, tested with sentence fragments, tobe arranged in correct order (6).
2.2.2.6. Writing. In the NGA, writing is tested with 10 items,
including writing of the patient' s own name, word copying, word
dictation, written naming, and sentence dictation.
2.2.3. Similarity to Other Aphasia Tests
In comparing the methodology with other modern operationalizations of the same model, Benson and Geschwind (1977) are considered an authoritative source on recommended practice for clinical
neurological investigation. Among modern aphasia test-batteries, reference is made to the Boston Diagnostic Aphasia Assessment (BDA)
(Goodglass & Kaplan, 1972), the Western Aphasia Battery (WAB)
(Kertesz & Poole, 1974; Kertesz, 1979), the Aachener Aphasietest (AAT)
(Huber, Poeck, Weniger, & Willmes, 1983), and the Neurosensory
Center Comprehensive Examinabon for Aphasia (NCCEA) (Benton,
1967). It may be doubtful whether the NCCEA is intended as an
operationalization of the Wernicke-Lichtheim conception of aphasia.
Benton (1967) gave as one of the purposes of the test
To include specific tests in the battery which could be employed to
investigate current questions in the field of aphasia (e.g. the reality of the
clinical pictures of conduction aphasia, central aphasia and transcortical
aphasia). (p. 41)
34
CHAPTER 2
This statement must mean that the test comprises the necessary information for making a classification.
There is a high degree of overlap in the types of tasks included
in these tests, and they are in significant agreement with clinical
neurological recommendations of suitable tasks. The greatest variation is found in the procedure for evaluating fluency of speech. The
NCCEA has no procedure for registering fluency. The BOA, the AAT,
and the WAB use rating procedures, based on the same sort of qualitative observations as in the NGA. The NGA seems to be the first
test to use quantified measures of speech (words per minute, utterance length) in clinical aphasia testing.
2.2.4. Choice of Normative Sampie

The NGA has been standardized on a sample of 161 consecutive
referrals with aphasia to a rehabilitation hospital (Reinvang & Engvik,
1980b). The reason for referral was mainly a request for treatment or
for an evaluation of the indication for treatment. No study of other
standardization samples has been performed.
In the Boston diagnostic test, the standardization was also limited to aphasics referred to the hospital (a VA hospital). lt was feit
that the selectivity of the group could be counterbalanced by developing separate norms for different severity-groups in the sample. This
procedure was found, however, to affect only the Ievel of the resulting
profilesandnot their form, and it was therefore dropped.
In the Western Aphasia Battery, control groups of nonneurological and neurological nonaphasic patients were used. For the groups
with no involvement of the dominant hemisphere, the overall scores
were above the 97th percentile for the aphasic group. In a group with
diffuse involvement of the cerebrum, including the dominant hemisphere, but with no clinical diagnosis of aphasia, the score was at the
91st percentile for the aphasic group, but here one might question
the criterion for the presence of aphasia in the control group. The
results confirm the impression that nonaphasic individuals rarely fail
the sort of items included in aphasia tests.
In connection with standardization of the NCCEA, Benton (1967)
developed an elaborate procedure for age- and education-adjustment
scores. He included anormal group in his standardization procedure
but found that there was very little overlap between the aphasic and
35
the normal populations. A reference group of aphasics was therefore

used for developing norms for intra- and interindividual comparisons
of aphasics. For the AAT, the main interest is to classify clinical syndromes, but as a part of standardization, nonaphasic controls were
included in the same proportion as in clinical referrals. The tests,
excluding speech rating, discriminated aphasics from normals with
above 90% accuracy. Had spontaneaus speech been included, the
discrimination would have been near perfect (Willmes, Poeck, Weiniger, & Huber, 1980).
It may be concluded that the sort of tests described above, presumably including the NGA, discriminate poorly within the normal
group. Altering the content of the tests to improve on discrimination
among normals would lead to loss of discriminatory power among
aphasics, as they would all be compressed into a narrow segment of
the normal range or would fall totally below the normal range. A test
constructed for the purpose of registering the type and degree of
aphasic disturbances must therefore be standardized on a sample of
aphasic patients.
The standardization sample is admittedly selective. One of the
greatest clinical problems in connection with the clinical management
of aphasia today is to find criteria for the selection of candidates for
treatment and for differentiating the form of treatment among those
candidates selected from the !arge population of individuals with
unmistakable and often severe aphasic difficulties who seek such
treatment. I believe that the intersection of selective pressures originating from the patient, his or her relatives, and the system of referral
in Norway has produced a group that is representative in a limited
sense, namely, representative of the type of aphasic who is currently
considered a possible candidate for treatment. A test that is useful
for establishing criteria for differentiation and prognosis in this group
meets a significant clinical need.
2.2.5. Standardization
The standardization sample consisted of a subset of the total

sample used in the rest of this study. It consisted of all aphasics
admitted to Sunnaas hospital between 1974 and 1978, a period of
about 4 years. This sample included 161 patients.
It was not regarded as necessary to restandardize the test for
36
CHAPTER 2
Table 2.2. Mean, Standard Deviation, Range, and Rehability of Subtest Scores
N
Mean
so
139
1.79
1.1
132
124
56
133
131
133
51
.70
.56
.55
1.56
.52
.86
1.12
.9
.8
.8
1.1
.9
1.0
1.1
70
72
48.89
3.92
39.7
2.4
Body parts, identify

Body :parts, describe
Body1actions
Objects, identify
Objects, describe
Objects, actions
Ideas, meaning
Ideas, relations
Total
161
75
161
73
75
161
77
77
73
8.60
3.59
6.37
9.1.9
4.61
6.70
11.20
3.52
53.4
3.4
1.8
3.1
3.0
2.0
3.2
3.1
1.3
17.5
Words
Nonsense syllables
Sentences
Total
104
106
106
104
14.42
4.96
6.78
26.10
7.1
3.0
4.6
14.1
Body parts
Body actions
Objects
Objects, action
Responsive
Total
161
161
73
161
161
73
6.59
1.91
6.73
3.30
6.12
26.70
4.4
1.7
3.7
2.5
4.2
15.2
Letters
Words
Sentences
Total
161
161
161
161
5.04
9.92
3.19
18.10
1.7
3.6
2.1
6.7
Letters
Words
Sentences
Total
161
75
160
75
4.47
6.43
5.59
16.90
2.2
3.7
4.3
9.6
~26
.91
.94
.95
.98
Main variable subtest

Communication
Quality of speech
Fluency
Repetition
Naming
Reading
comprehension
Reading
aloud
Litera! paraphasia
Complex paraphasia
Visible effort
Hesitation
Stereotypy
Articulation
Self-correction
Words per minute
Utterance length
Range reliab.
0---4
~3
~3
~3
~3
~3
~3
~3
~11
~5
~10
~11
~
~10
~14
0---4
9-71
~20
~
~12
0---40
~11
~5
~10
~5
~10
0---41
~
~12
~5
~23
~
~10
~10
.93
.87
.87
.92
.89
.89
.84
.66
.98
.97
.91
.95
.98
.95
.89
.94
.90
.95
.98
.87
.95
.93
.97
Syntax
Sentence
arrangement
71
2.48
2.2
.83
Writing
Total
103
5.20
3.2
~10
.88
68
150.80
61.1
Aphasia coefficient
14-217
.995
37
the total sample of the present study, as the test had been shown to
have generally satisfactory statistical properties. To avoid unnecessary
detail, the specific composition of the standardization sample is not
shown. It contained 76% cerebrovascular cases, of whom two thirds
were men. The mean age was 50. Furtherdetails can be found in the
handbook for the test (Reinvang & Engvik, 1980b). A Danish edition
of the test has appeared (Reinvang & Engvik, 1984), and a Swedish
translation is used informally.
2.2. 6. Statistical Properties
The performance of the standardization sample of subjects on

the test is shown in Table 2.2 The subjects showed a wide range of
performance, from almost no correct responses to perfect performance. The number of subjects varied somewhat because of revisions
in the content of the test made during the time of standardization.
All the main variables (total scores) have a very high reliability
measured by the alpha coefficient, a measure of internal consistency
(Nunally, 1967). The test-retest correlations have been determined
(Reinvang, 1981). They are reproduced here as Table 2.3 to confirm
the impression of the high reliability Ievels given by the measure of
internal consistency. The testing was clone in a clinical context, which
means that the tests and the retests were performed with lang intervals. Recovery processes thus may have effected the results. The tests
Table 2.3. Test-Retest Correlations
Repetition
Naming
Reading aloud
Syntax
Writing
Aphasia coefficient
"Not statistically significant.
Acute
Chronic
.93
.90
.94
.93
.89
.96
.66
.89
.98
.so
.62
.56
.70
.44"
.58
.82
CHAPTER 2
38
Table 2.4. Test-Retest Correlations
Communication
Litera! paraphasia
Camplex paraphasia
Visible effort
Hestiation
Stereotypy
Articulation
Self-correction
Words per minutes
Utterance length
Acute
Chronic
.87
.84
.74
.40"
.95
.51
.94
.79
.69"
.86
.86
.68
.54
.87
.75
.60
.80
.49
.95
.98
"Not statistically significant.
are divided into acute (test and retest within 6 months after onset of
aphasia) and chronic (later tests).
For the rating scales and quantitative measures of speech, no
study of internal consistency could be made. In Table 2.4, the testretest correlation coefficients have been given.
The correlations show some variability and are generally lower
than the test-retest coefficients for objective scores. The quantitative
measures (words per minute and utterance length) compare favorably
with the rating scales. The reason that no study of intertester reliability
was performed is that, at the time of standardization, very few persons except the author had been trained in administering the test.
The homogeneity of the main variables can be evaluated by
inspecting the table of intercorrelations of subtests contributing to the
same main variable (Table 2.5).
Principal component analyses with varimax rotation were performed to evaluate the loading of each subtest on the main variable
to which it contributes. The loadings are generally very high and
indicate that further splitting up of themainvariables is not motivated
by the standardization data.
The total scores have a high loading on a common factor (principal component analysis with varimax rotation), and this justifies the
introduction of the sum of total scores on main variables, the aphasia
coefficient, as a valid measure of the severity of aphasia (Table 2.6).
39
Table 2.5. Intercorrelations and Loading an First Factor for Subtests"

Intercorrelations
2
Loading
on 1st
factor
Body parts, identify
Body parts, describe
Body parts, action
Objects, identify
Objects, describe
Objects, action
Ideas, meaning
Ideas, relations
78
86
54
71
79
76
74
81
57
73
74
76
34
48
68
82
82
45
83
70
52
22
83
72
27
75
43
38
90
88
91
74
88
92
87
48
Repetition
Words
Nonsense syllables
Sentences
91
86
96
96
94
85
Naming
Body parts
Body parts, action
Objects
Objects, action
Responsive
89
88
91
90
76
91
85
84
84
92
96
93
90
97
95
Letters
Words
Sentences
64
67
87
90
91
67
Reading aloud
Letters
Words
Sentences
80
77
86
"Decimal points are omitted throughout the table.
92
95
94
40
CHAPTER 2
Table 2.6. Intercorrelations and Factor Loading for the Main Variables"
Intercorrelations
Variable
Auditory
comprehension
Repetition
Naming
Reading aloud
Syntax
Writing
Loading on
Ist factor
89
68
85
84
72
57
65
80
57
78
47
58
75
78
61
62
77
63
69
62
64
65
82
91
88
89
75
80
"Decimal points are omitted throughout the table.
2.2.7. Relation to Background Variables
Studies with the NCCEA (Benton, 1967) have indicated that age
corrections should be employed for some aphasia variables, and the
work of McGlone (see McGlone, 1980) indicates that aphasia may be
less severe in females than in males. No studies have been found,
apart from Benton (1967), showing a relationship between the type
of variables included in the aphasia test and education. In general, it
is expected that the minimal overlap between the aphasia population
and the normal population in language performance makes it unlikely
that any strong relationship with education should be found. The
relationship of aphasia test variables to age, sex, and education is
summarized in Tables 2.7, 2.8, and 2.9. More detailed tables are given
in the handbook for the test.
Differences were tested for significance by one-way analysis of variance and relationships at p ~ .05 have been given as
significant.
There are few significant relationships between age and performance on different parts of the test. In general, the results indicate
that separate norms should not be used for separate age groups. It
must be added, however, that children were not represented in our
standardization sample. The variables showing a weak interaction
with age (repetition and naming) do not show a consistent trend of
decreasing performance with increasing age. Hence, there is no reason to suggest systematic age-dependent adjustments of scores.
41
Table 2.7. Mean of Test Performance in Different Age Groups

Age group
Variable
Words per min.
Utterance length
Auditory
comprehension
Repetition
Naming
Reading
comprehension
Reading aloud
Syntax
Writing
19
and
under
20-40
41-50
51--60
61-70
70+
62.8
4.9
54.3
31.7
3.4
56.2
45.4
2.8
42.3
52.4
4.2
55.8
61.8
4.6
55.6
45.0
3.4
50.6
n.s.
n.s.
n.s.
28.6
29.7
21.7
28.6
29.2
19.8
15.4
12.8
17.2
29.3
31.1
17.0
29.9
28.1
18.1
23.2
27.4
16.9
.05
.05
n.s.
23.7
3.3
5.3
16.8
2.8
5.9
11.0
1.5
4.7
18.5
2.5
4.9
18.3
2.5
4.8
14.2
2.8
6.2
n.s.
n.s.
n.s.
It may be concluded on the basis of Table 2.8 that there are few
and unsystematic relationships between educational or professional
Ievel and performance on the aphasia test. The results indicate that
separate norms for different educational groups are not motivated.
In the one case of a significant relationship (reading comprehension),
the tendency was for subjects with higher education to have poorer
reading comprehension. This is probably an accidental finding.
The results in Table 2. 9 indicate that there are few relationships
between sex and performance on the aphasia test and that separate
norms for males and females are not motivated. The observed difference on reading comprehension is probably an accidental finding.
In conclusion, the statistical studies of the standardization data
indicate that, in this sample, the NGA measured aphasic performance
with a high degree of reliability and consistency, and that performances showed very little dependence on age, sex, or educationallevel.
2.2.8. System of Gradation

In choosing a system of grading results, one finds that the two
most likely systems are z scores and percentile values.
The z score is based on the assumption that the underlying
Words per minute

Utterance length
Repetition
Naming
Reading aloud
Syntax
Writing
Variable
61.7
4.6
59.6
29.0
33.4
21.8
24.2
3.2
6.1
In school
58.4
4.4
56.4
24.8
28.9
20.0
18.8
2.4
4.6
Unskilled
Iabor
46.0
3.3
58.9
29.8
31.9
18.8
19.5
3.5
5.7
Skilied
Iabor
42.1
3.8
49.4
23.4
22.9
17.1
14.8
2.0
4.8
Artisan
54.5
4.6
52.3
29.8
26.0
18.4
17.0
2.6
5.2
High school +
additional
Table 2.8. Mean of Test Performance in Different Educational Groups
5.6
61.5
4.6
38.8
27.5
19.8
13.2
10.3
College,
university
n.s.
n.s.
n.s.
n.s.
n.s.
.05
n.s.
n.s.
n.s.
::0
I)
:r
>-
;e
43
Table 2.9. Mean Test Performance in Malesand Fernales

Variable
Male
Fernale
Words per minute

Utterance length
Repetition
Naming
Reading aloud
Syntax
Writing
48.9
3.9
52.5
25.5
26.1
17.4
16.5
2.5
5.0
49.2
4.1
54.9
27.3
27.8
19.7
17.5
2.5
5.6
n.s.
n.s.
n.s.
n.s.
n.s.
.05
n.s.
n.s.
n.s.
from the mean in number of standard deviations, and the sign of the
score tells if the deviation is positive or negative. The BOA assessment
uses z scores to represent the results. These scores are advantageaus
for further statistical treatment but are open to the criticism that the
underlying distribution of scores for most tasks used with aphasics
is not normal.
Percentile values tell what proportion of scores falls above or
below a given value. If a raw score of 10 correct responses corresponds
to a percentile value of 25, that means that 25% of the standardization
0 1 2 3 4 5 6 7 8 9 10
WPM
Fflien<J
'WPU
!lud Comprehef1Sion
RepetW.on
**
I
to
H
~
'
ltt:::~~
p
~l"!f-"
ltp~
fo
l~lP
~ 1 7 1 0 I~
15
~
f
5I..
~I'
20~
f'
~)~J
z
7
I2~
Word.s per mi.n.
= Word.s
Oral Readil]g
* WPM
J~
~~~
Readlllg Comprehension
** WPU
~( 3
NaJTUfl!]
Sentence Constructlon
Writin!J
flphtisiJJ CoejJLcient
1~( )J
I
I.
I 'I 2 5 2 9 122
0 1 2 3 't 5 6 1 8 9 10
per utterance
Figure 2.2. Example of aphasia test result.
44
CHAPTER 2
population scores below this value and 75% score above. Percentile
values are less suitable than z scores for various statistical treatments
but have the advantage of not presupposing any form of the underlying distribution. Percentile values are used in the NCCEA and in
the WAB.
In the NGA, percentile values are used to represent scores. The
main usage of percentile values is to make individual judgments in
the form of test profiles, whereas for statistical studies on groups the
untransformed raw scores are used.
An example of a test result is given in Figure 2.2. The raw scores
are circled, and the scale indicates that, for example, a raw score of
38 on naming is at the 65th percentile of the distribution of aphasics.
How to proceed from the test profile to the determination of
aphasia type is described in the next chapter.
TYPES OF APHASIA
The nomenclature and criteria for types of aphasia in this study are
based on the Wernicke-Lichtheim model, as described in Chapter 1.
Within this tradition, different test batteries have different rules or
guidelines for determining the aphasia type after converting raw scores
to derived scores and examining the resulting test profiles.
The Boston Diagnostic Aphasia Assessment (Goodglass &
Kaplan, 1972) gives only guidelines for interpretation tagether with
an indication of the range of variability in scores within a given type
of aphasia. Clinical judgment is recognized in addition to scores as a
valid basis for classification.
The Western Aphasia Battery (WAB) gives more strict quantitative definitions with exact cutoff values.
The classification system and criteria are shown in Table 3.1.
The scores are percentage scores, except for fluency, which is
rated directly on a 10-point scale.
In the system based on the Aachen Aphasia Test (AAT), clinical
judges have classified a reference group of aphasics. A computer
program decides the likelihood that new patients will be assigned to
any subcategory of the reference group, and a probability of at least
80% is necessary to accept a classification (Willmes et al., 1980).
In choosing between a system with a strict quantitative criteria
and one with room for clinical judgment, one can argue that a study
that intends to explore the relationship of aphasia type to locus of
lesion and to neurological signs must strive for maximally specific
definitions of aphasia type, as otherwise the risk of the judgment's
45
CHAPTER 3
46
Table 3.1. Aphasia Classification System of WAB

Criteria for classification
Aphasia
Global
Broca's
Isolation
Transcortical motor
Wernicke's
Conduction
Anomic
Fluency
Comprehension
Repetition
Naming
0--4
0--4
0--4
0--4
5-10
5-10
5-10
5-10
0-3.9
4--10
0-3.9
4--10
0-6.9
0-6.9
7-10
7-10
0--4.9
0-7.9
5-10
8-10
0-7.9
8-10
0-6.9
7-10
0-6
0-3
0-6
0-8
0-9
0-9
0-9
0-9
Note. From Aphasia and associated disorders: Taxonomy, localization and recovery (p. 58) by
A. Kertesz, 1979, New York: Grune & Stratton. Copyright 1979 by Grune & Stratton,
Inc. Reprinted by permission.
being contaminated by information on the dependent variable would

be present (i.e., the fallacy of psychoneurology). For research purposes, a strictly quantified system therefore seems necessary.
A classification system like that in the WAB has the property of
being exhaustive; that is, any patient with a complete aphasia-test
result is assigned to one of the groups in Table 3.1. It may be argued
that a classification system based on a clinicopathological model should
allow for unclassifiable or "mixed" cases. This conclusion seems a
natural consequence of choosing a model that stresses selective deficits in patients with limited lesions. In clinical practice, some patients
must have composite lesions, and hence mixed symptomatology. The
proportion of such cases depends on the selectivity of action of the
agent of injury, as well as the selectivity of the sample. For research
purposes, it may be justified to exclude unclassified cases from a study
because they tend to obscure the findings. Clinically, however, they
are common and must be characterized by some diagnostic term.
Both in the Boston diagnostic test and the Western Aphasia
Battery, the rules for classification are a priori, that is, based on arbitrary definitions of cutoft values. lt would be possible to study empirical classifications based on purely statistical criteria so that optimal
cutoft points for predicting a given outcome (e.g., locus of lesion)
could be established. This procedure is followed in the AAT, where
TYPES OF APHASIA
47
the criterion variable is a dinical dassification. Another way to derive

an empirical dassification is to study dusterings of scores within a
sample of aphasia test performances by advanced statistical methods.
In previous research by Kertesz (1979), both the methods of a priori
quantitative definition and of empirical, statistically derived dassification have been used. On the basis of these studies, it seems that a
likely development is a further refinement within the framework of
a dassical dinicopathological dassification system. A dustering analysis performed on the AAT by Willmes et al. (1980) gives groups that
overlap extensively with the dinical dassifications.
3 .1. Classification System of the Norsk Grunntest
for Afasi
In the Norsk Grunntest for Afasi (NGA), strict, quantitative criteria for division into types are suggested, so that, given a test profile,
the type designation follows automatically. The cutoff values chosen
for this study have been determined by the author' s experience and
judgment. The dassification adopted by the NGA is not exhaustive;
hence, undassifiable cases occur. No patientortest is exduded from
the study because of "mixed" findings or other peculiarities of the
test result. The definitions of aphasia types proposed here differ from
those used by Kertesz et al. mainly in being relational. Rather than
focusing on the absolute Ievel of, for example, comprehension in Broca
aphasia, it is stressed that comprehension must be better than fluency
and that the difference must exceed a certain cutoff score. This type
of rule is intended to allow a characteristic configuration of performances to be designated by a given name although the Ievel of performance might improve. Relational definitions are most appropriate
with Broca, Wernicke, conduction, anomic, and transcortical aphasias, where dinical descriptions all note relational features. All aphasia
types should not be defined relationally, however. In global aphasia,
the uniform severity of the deficit across performances should be
stressed. The termglobal aphasia refers only to patients with nonfluent
speech. I have chosen to indude jargon aphasia as a separate type,
and to define it as uniformly severe aphasia but with mixed or fluent
speech. By having separate terms for all the most severe aphasics,
regardless of fluency, confounding of type and severity of aphasia
48
CHAPTER 3
may be reduced in further analyses comparing performance in different types of aphasia.
3.1.1. Definitions of Speech Classification

If a tape recording exists so that quantitative evaluations can be
made, then the rules are the following:
Nonfluent speech: Words per minute below 40 and utterance length

below 4.0.
Words per minute above 80, utterance length
Fluent speech:
above 5.0 and presence of paraphasia (literal or
complex).
The cutoff values follow the values suggested by Kerschensteiner,
Poeck, and Brunner (1972). The cutoff point for the upper Iimit of
nonfluency is also close to the 50th percentile in the empirical distribution (see Chapter 2).
Intermediate: The criteria for neither nonfluent nor fluent speech are
satisfied, and speech is neither normal nor predominantly dysarthric.
Dysarthric:
Fluency cannot be determined because of a strong
dysarthric component in the speech. Dysarthria does
not preclude that there is also an aphasic disturbance.
Normal:
Speech is quantitatively in the fluent range, but without
presence of paraphasia or other qualitative signs.
If a tape recording is not available, then nonfluent speech is
diagnosed on the basis of a high rating on hesitation, visible effort,
and stereotypy. Fluent speech means the presence of complex paraphasia or literal paraphasia, with a relative absence of visible effort
or hesitation. If one dass of ratings does not clearly predominate over
the other, then fluency is mixed.
3.1.2. Definitions of Aphasia Types

The most characteristic features of different types of aphasia are
outlined in Table 3.2, whereas precise, quantitative definitions are
given in the following sections.
3.1.2.1. Types of Aphasia with Intermediate or Fluent Speech. Anomic
Fluent
Low
Low
High
Low
Low
Others
High
High
Others
Low
Low
Low
Low
Mixed nonfluent
Isolated speech area syndrome

Global
Transcortical motor
Mixed, with mixed fluency
Jargon
Conduction
Mixed, fluent
Anomic
Conduction
Jargon
Wernicke
Low
High
Low
High
Low
Low
Low
Aphasia type
Naming
Repetition
High
Others
Low
Note. For precise definitions, see text.

"The terms high and low may be defined conditionally.
Nonfluent
Intermediate
High"
Fluency
Auditory
comprehension
Table 3.2 Characteristics of Different Aphasia Types
:t
[J)
>s;;
::c
>
.."
[J)
tTl
-<
....,
50
CHAPTER 3
aphasia has fluent speech. Naming is more than 20 percentile points

worse than repetition and fluency.
Conduction aphasia has intermediate or fluent speech and comprehension is more than 20 percentile points better than repetition.
Transcortical sensory aphasia has fluent speech, and repetition
is more than 20 percentile points better than auditory comprehension.
Jargon aphasia has fluent or intermediate speech. Auditory comprehension, repetition, and naming are below the 20th percentile.
The aphasia coefficient is below the 20th percentile.
Wernicke aphasia has fluent speech, and auditory comprehension and repetition are more than 20 percentile points below fluency.
The aphasia coefficient is above the 20th percentile.
Mixed aphasia with fluent speech is the group of remaining
patients with fluent speech; they have test profiles not classifiable as
jargon, Wernicke, transcortical sensory, anomic, or conduction aphasia.
This is not regarded as a type of aphasia in the classical literature.
Mixed aphasia with intermediate speech is the group of patients
in whom spontaneaus speech cannot be classified as either fluent or
nonfluent, and the test profile is not classifiable as jargon aphasia or
conduction aphasia. This group is not homogenic and must be regarded
as u'nclassifiable cases.
3.1.2.2. Types of Aphasia with Nonfluent Speech. Transcortical motor
aphasia has nonfluent speech, and auditory comprehension and repetition are more than 20 percentile points better than fluency.
Broca aphasia has nonfluent speech and auditory comprehension more than 20 percentile points better than fluency. Naming is
more than 20 percentile points better than fluency.
Isolated-speech-area syndrome has nonfluent speech and repetition more than 50 percentile points better than auditory comprehension and naming.
Global aphasia has nonfluent speech and scores below the 20th
percentile for auditory comprehension, repetition, and naming. The
aphasia coefficient (AC) is below the 20th percentile.
Mixed aphasia with nonfluent speech is not a type of aphasia
in conventional classifications. The group consists of the remaining
patients with nonfluent speech who cannot be classified as having
global, isolation syndrome, Broca, or transcortical motor aphasia.
51
TYPES OF APHASIA
Table 3.3. Characteristics of the Sampie

Characteristic
Mean
Aphasia coefficient (AC)

Age at test (in years)
Time from illness to test (in days)
129
50.2
260
Median
SD
135
62.6
53.1
134
15.9
372
Range
1-216
11-80
2-2,131
3.2 Subjects
The classification rules and other empirical questions were tested
on a main sample described in the following (Table 3.3). It consisted
of 249 patients included on the basis that a complete aphasia test and
scores on key neuropsychological variables were recorded. Patients
with predominantly dysarthric speech were excluded.
There were 161 men and 88 women in the sample. In comparison
with the standardization sample (Chapter 2), it may be noted that the
aphasia as measured with AC was on the average more severe in the
main sample. There were 84% cerebrovascular patients, the large
majority of thromboembolic origin. Head injuries were represented
by 11% in the sample and miscellaneous other diagnoses by 5%.
The results of applying the classifications rules for aphasia types
to this sample are shown (Tables 3.4 and 3.5). When more than one
test was performed, the first was used for classification.
3.3 Stability of Classification

When more than one test has been clone, the stability of the
classification can be studied. This is not strictly a reliability measure,
as the influences of the recovery process are at work when the time
span between tests is several months. Still, under these conditions,
it should be pointed out that the test-retest coefficients of the
Table 3.4. Speech Fluency

Nonfluent
Fluent
Intermediate
Normal
129
34
78
8
(52%)
(14%)
(31%)
( 3%)
52
CHAPTER 3
Table 3.5. Aphasia Type

Global
Isolation syndrome
Transcortical motor
Broca
Jargon
Wernicke
Anomic
Conduction
Nonfluent, unclassifiable
Others
37
3
19
12
11
6
8
(15%)
( 1%)
( 1%)
( 8%)
( 5%)
( 4%)
( 2%)
( 3%)
67
61
(27%)
(24%)
23
( 9%)
quantitative test variables are very high (see Tables 2.3 and 2.4; the
tables on the stability of classification are Tables 8.2 and 8.3).
Significant stability is present in a statistical sense; that is, the
probability of being classified in a syndrome is not independent of
previous classification. Still, the probability of reclassification is quite
sizable (36% ), and this seems somewhat undesirable. Rather than
having a complex set of intersyndrome movements with improvement, it would be desirable that a patient retain his or her classification. If improvement involves highly specific patterns of recovery,
however, then syndrome reclassification may be accepted as a genuine finding.
3.4. Comparison with Typology of WAB

With respect to the WAB, the similarity in test procedure is high
enough so that a comparison of outcomes can be made. On the basis
of the classification criteria of Kertesz, a group of patients tested with
the NGA can be classified according to the criteria used in WAB. The
relationship of the two classification systems is shown in Table 3.6
(from Sundet & Engvik, 1984).
The NGA is more restrictive in assigning a classification, but of
the patients who are classified, 85% are given the same classification
by the WAB. Same of the apparent disagreement is spurious. The
WAB does not use jargon aphasia as aseparate category. Given this
premise, it is reasonable that the 8 patients called jargon asphasics
in the NGA should be classified as Wernicke aphasics in the WAB.
In all, it is reasonable to regard patients With a given syndrome
Broca's
59
(31 %)
44
(23%)
(4%)
(2%)
38
(20%)
16
16
(8%)
1
6
(3%)
12
"0
1::
J-;
f-
"'
1::
u
;:;
1::
..
(10%)
19
(2%)
4
193
(20%)
(28%)
55
39
(2%)
3
3
12
(10%)
(2%)
19
(4%)
(4%)
8
7
(2%)
(1%)
1
3
(12%)
(15%)
23
28
-<
1::
60
West Germany. Reprinted by permissionof the authors.
Note. From The validity of asphasic subtypes by K. Sundet and H. Engvik, 1984, June. Paper presented at INS-European Conference, Aachan,
Mixed, fluent
Mixed, mixed
Mixed, nonfluent
1
Conduction
Anomic
3::
(!)
J-;
Trans. sens.
<Ii
"2
J-;
f-
"'
<Ii
1::
Wernicke's
21
"'
.9
(!)
<Jl
<Ii
1::
<Jl
'(!)
29
20
J-;
'co
0
E
Jargon
Trans. mot.
1
28
Global
Isolation
Norwegian Basic Aphasia Assessment
~
..0
<Jl
Western Aphasia Battery
V1
VJ
:;;:
(fl
::r:
)>
>-;:)
)>
.."
(fl
tr1
-<
>-;:)
--l
54
CHAPTER 3
designation in the NGA system as a subset of patients with the same

diagnosis in the WAB system.
Sundet and Engvik (1984) also performed a duster analysis after
the guidelines suggested by Kertesz and Phipps (1977) and described
the 9 most prominent clusters (Table 3.7). The corresponding table
from Kertesz and Phipps (1977, Table 2) is reproduced as Table 3.8.
In interpreting the clusters, it is apparent that Clusters 1 and 2
comprise severe aphasias with nonfluent speech, whereas 4 and 5
represent less severe nonfluent patients, including some whose speech
rate placed them in a borderline zone.
Cluster 3 clearly captures severe aphasias with fluent speech
output, whereas 6, 7, and 8 have the less severe fluent patients, with
the qualification that Clusters 7 and 8 have a high proportion of cases
with less-than-fluent but not nonfluent speech.
3.5. Levels of Classification

On the basis of these quantitative analyses one might question
the desirability of a fine-grained classification system for all purposes.
Table 3.7. Cluster Composition of Patients in the NGA System
Cluster
Number of
patients
Percentage of aphasia types
25
96% global
20
15% global
85% mixed nonfluent
16
50% jargon
25% Wernicke
19
74% mixed nonfluent
32
31% Broca
63% mixed nonfluent
29
31% conduction
52% mixed intermediate
19
32% conduction
11% anomic
53% mixedintermediate
23
57% Broca
9% trans. mot.
22% mixedintermediate
66% conduction
33% anomic
TYPES OF APHASIA
55
Tab/e 3.8. Cluster Composition of Patients in the WAB System

Cluster
Number of
patients
Percentage of clinical aphasia types
30
97% global
II
15
86% Broca's
III
12
25% global
25% Broca's
25% isolation
IV
13
54% Broca's
23% isolation
23% transcortical motor
100% transcortical sensory
VI
12
58% conduction
VII
11
100% Wernicke's
57% conduction
VIII
IX
18
63% anomic
20
100% anomic
Note. From "Numerical taxonomy of aphasia" by A. Kertez and J. B.

Phipps, 1977, Brain and Language, 4, pp. 1-10. Copyright 1977by Academic
Press, lnc. Reprinted by permission.
Rather than saying, "Aphasics should be grouped in this way, because

that is how they are grouped in nature," it seems possible to adopt
both a fine-grained and a coarse-grained classification system depending on the objective.
It is questionable if a syndrome division as derived from the
Wernicke-Lichtheim model is fruitful for all purposes. The empirical
results so far suggest that a fourfold division of aphasias into mild
versus severe deficit and nonfluent versus relatively fluent speech
captures the main divisions of the empirical structure. This seems to
be the view also of the Aachen group, who operate with four main
syndromes: global, Broca, Wernicke, and anomic aphasia. Poeck
(1983b) stated:
Aphasie syndromes are expressive syndromes. They do not have a clearly
defined receptive aspect, although language comprehension is compromised in all aphasic patients, even though to a different degree. (p. 85)
Comprehension is, next to fluency, the most important criterion

for classification in the Wernicke-Lichtheim model.
A fourfold classification system has also been used in many
CHAPTER3
56
reports from the Milan group (Basso, Vignolo, and others; see Basso,
Capitani, & Vignolo, 1979; Basso, Capitani, Luzzati, & Spinnler, 1981).
The possible contribution of time to classification must be
acknowledged. Even at the Ievel of coarse-grained analysis, some distinction between acute and chronic symptomatology should be made.
A coarse-grained classification system can be used, then, when
general structures underlying performance are analyzed with psychometric methods. For the analyses presented in this study, a classification system with three dimensions is used: fluency, severity,
and chronicity. Each of these variables is dichotomized so as to prepare the ground for psychometric analyses with ANOVA designs and
opportunities for measuring interactions.
The cutoff point for fluency is set so that all nonfluent patients
form one group. The cutoff point for severity is at the median value
of the aphasia coefficient of the sample. For classifying chronicity, I
follow the same practice as Kertesz and Phipps (1980), regarding
aphasia as chronic when 6 months or more have elapsed since onset.
This is somewhat Ionger than the median of the sample.
The resulting structure of the sample is shown as a quasi-threedimensional figure (Figure 3.1).
3.6. Sampie Structure

An analysis of the relation of the classification system to sex,
age, etiology, and education was performed to ascertain the presence
of any systematic bias in the groups. The relationship was measured
with chi-square tests when the dependent variable was discrete (sex,
non]Went
fllient
se~
mild
Figure 3.1. Sampie structure.
57
TYPES OF APHASIA
education, and diagnosis) and with ANOVA for age. Only in the
latter analysis could interactions be tested, and these were found to
be insignificant (see Table 3.9).
The age variable was weakly related to fluency, and the finding
was that nonfluent patients were younger than fluent patients (48.8
vs. 50.0 years). More importantly, the severe aphasics were older
than the mild aphasics (52.2 vs. 46.6 years). The one significant finding
on education is probably not important: twelve subjects were still in
school, and of these, 10 had a mild degree of aphasia.
With diagnosis, the relation to fluency was just marginally significant (p = .05). Traumatic patients who tended to fall in the fluent
group account for the trend. More important is the relationship to
severity, which isthat traumatic patients tended to fall in the mild group.
The tendencies support earlier reports that head-injured patients
have a higher probability of being fluent than vascular cases. This
finding has been reported for closed head injuries by Heilman, Safran,
and Geschwind (1971).
There is also a tendency for fluent aphasics to be older than
nonfluent aphasics, as reported by bler, Albert, Goodglass, and
Benson (1978). This tendency is also present in the series reported by
Kertesz and Sheppard (1981), who reviewed alternative explanations
of the finding. It may be related to the changing organization of the
aging brain, as suggested by Brown and Jaffe (1975); to different
survival probabilities for patients with different aphasia syndromes;
or to differences in the etiology of cerebrovascular disease with age,
as suggested by Kertesz and Sheppard (1981).
The conclusion must be that, in analyses showing a difference
of severe versus mild aphasics, the age bias must be corrected for,
unless the variable in question is known tobe uncorrelated with age.
For some analyses the separation of vascular from nonvascular groups
should be considered also.
Table 3.9. Summary of Relation of Aphasia Classification to Background Variables
Age
Sex
Education
Diagnosis
Nonfluent vs. fluent

Severe vs. mild
Acute vs. chronic
Note:- = no significant difference; * = significant difference at p < .05; **

= significant difference at p < .01.
58
CHAPTER 3
3.7. Conclusion
In summary, a simplified classification system based on the
dimensions fluency, severity, and chronicity is justified by empirical
findings and offers advantages for the purpose of psychometric studies. This system is exhaustive and classifies all patients. The reality
of selective aphasia syndromes, some of them rare, is confirmed. The
criteria applied here are consistent with, but stricter than, the criteria
used in another major operationalization of the Wernicke-Lichtheim
model, that of Kertesz (1979). The system is exclusive and did not
classify 51% of the cases in this sample. The possible advantage was
that the resulting groups were homogeneous. Traditionally, the relation of classification to locus of lesion has been the center of interest.
The strongest argument for traditional typology is that it is claimed
to predict lesion localization. As noted in Chapter 1, even the critics
of a clinical-pathological model or a localizationist approach have
accepted that the traditional typology has a predictive value. The value
of the present well-defined, but restrictive, system remains tobe seen
(see Chapter 7).
More recently, the use of clinically defined types for neurolinguistically oriented analyses of language subprocesses has been
debated. The current opinion goes in the direction that only studies
of single cases or of very homogeneaus small samples are of value in
such studies (see Schwartz, 1984). Thesequestions cannot easily be
resolved, and for the time being, we must live with different classification systems, differing bothin type andin degree of inclusiveness,
and must evaluate their usefulness in relation to the objective of study.
SELECTIVE APHASIAS
4 .1. Types of Selective Aphasias

Under this heading is usually discussed a varied set of conditions that
have the common feature of only affecting verbal performance in a
given modality (auditory or visual, oral or graphic) while the general
Ievel of language functioning is unimpaired. The latter condition is
very rarely fulfilled since at least a mild aphasic disturbance is usually
present. The existence of deficits specific for comprehension or naming of some categories of words but not of others have also been
reported.
4 .1.1. Modality-Specific Aphasias

The types of modality-specific aphasias discussed in the Iiterature include agraphia (selective disturbance of writing), alexia (selective
disturbance of reading), aphemia (selective disturbance of speech), and
auditory verbal agnosia (pure word deafness), meaning selective disturbance of auditory language comprehension. In these cases, it can
be shown that whatever sensorimotor disturbances are present are
not sufficient to account for the verbal deficits, because other patients
with the same sensorimotor deficits do not have the same verbal
deficits. All these forms of aphasia are rare.
In aphemia, the patient has severe speech difficulty both spontaneously and on imitation. The presence of intact swallowing and
tongue motility distinguishes the condition clearly from anarthria,
59
60
CHAPTER 4
although the patient may show clumsiness and inaccuracy in imitating

oral movement. Articulation is impaired even for automatized
sequences (e.g., counting). In the recovery phase, the patient speaks
slowly and with effort but typically produces full sentences with no
sign of agrammatism. In the conceptual scheme of WernickeLichtheim, this form of aphasia is termed subcortical motor aphasia,
with the implication that the lesion disconnects output from the Broca
area from motor effector organs.
Pure agraphia is a very rare syndrome but has been described
by Exner (1881); more recent studies have been summarized by Laine
and Marttila (1981). Some of these cases have evident problems with
the motor aspects of writing and should be viewed as cases of an
apraxic nature. Some patients, however, have symptoms that are
more unequivocally of a language nature; one such case was described
in detail by Bub and Kertesz (1982). This patient produced semantic
paragraphias, writing concrete names from dictation but notbeingable
to write pronounceable pseudowords. No consistent locus of lesion
has been found for patients with pure agraphia.
Pure alexia (alexia without agraphia) has been studied in a series
of cases starting with Dejerine (1892). Geschwind and Fusillo (1966)
gave impetus to renewed interest in the syndrome, which they
explained as a disconnection syndrome (i.e., visual input is cut off
from the language areas). Patients with pure alexia have lesions of
the posterior left hemisphere, almost invariably resulting in right-side
hemianopsia. They typically read isolated letters or numerals with
some success but fail to read words or, indeed, to identify letters in
a word context. In addition to hemianopsia, inability to name colors
(color anomia) is frequently present. Patients may show varying degrees
of visual agnosia (i.e., a general inability to identify or name visual
stimuli). Although the disconnection explanation of pure alexia has
been widely accepted, opposition to it has been voiced by Hecaen
and Kremin (1976) and Levine and Calvanio (1982), both of which
groups have found evidence of a left-hemisphere mechanism for the
identification of letter groups. The destruction of this mechanism,
rather than the disconnection of right-hemisphere visual input from
left-hemisphere language processes, is involved in pure alexia.
In addition to pure alexia (alexia without agraphia), aphasic alexia
(alexia with agraphia) may be noted as a semipure defect. In these
SELECTIVE APHASIAS
61
cases, speech and auditory comprehension is relatively intact, whereas

disturbance of reading and writing may be severe. Several patterns
may be exhibited. The relative preservation of letter reading with
impaired word reading has been mentioned.
In the syndrome of deep dyslexia (see Coltheart, Patterson, &
Marshall, 1980), the patient produces semantic paralexias, reads isolated letters and meaningless letter combinations poorly, and reads
concrete nouns with fair success. Hecaen and Kremin (1976) found
sentence alexia tobe a distinct variety of alexia.
Auditory-verbalagnosia (pure word deafness) isarare condition
found in some cases of bilateral temporal-lobe pathology with signs
of more general auditory agnosic defects (Ulrich, 1978). A case without
pathological verification, but very likely only left-hemisphere involvement, was reported by Gazzaniga, Velletri Glass, Sarno, and Posner
(1973). In Wernicke aphasia, recovered cases may report the experience that the speech of others sounds like a foreign language; as one
patient said, "I can hear, but the sound doesn't come all the way
through." These cases may demonstrate relatively intact reading
comprehension and thus may approximate the condition of auditory
verbal agnosia.
4.1.2. Material-Specific Aphasias

There have been a few reports of deficits specific to certain categories of words. Goodglass, Klein, Carey, and Jones (1966), comparing data for comprehension and production (naming), made the
observation that the hierarchy of difficulty for categories of material
varies strongly between the conditions. Goodglass and Kaplan (1972)
remarked on the frequent occurrence of comprehension failure for
names of body parts. Dennis (1976) found a specific inability to produce or understand body part names in a young girl after resection
of the left anterior temporal lobe. The author concluded that the
patient' s problern could not be explained by a disconnection of sound
and meaning but involved lexical selection within the category of body
parts.
Color-naming defects are found primarily in connection with
symptoms of dyslexia or visual agnosia (de Renzi & Spinnler, 1967;
Kinsbourne & Warrington, 1964; Geschwind & Fusillo, 1966). The
62
CHAPTER 4
suggestion that there is a second form of color anomia occurring in

the context of aphasia and lesion of the language areas (Oxbury,
Oxbury, & Humphrey, 1969) has not won much support.
Geschwind (1967b), in his classical paper on naming errors, considered the hypothesis that the usually observed generalized naming
impairment is an agglomeration of category-specific naming defects.
He pronounced the issue to be of great theoretical importance but
found no empirical basis for resolving it. Poeck and his associates
(Orgass, Poeck, & Kerschensteiner, 1974; Poeck & Stachowiack, 1975)
rejected the notion of category-specific naming or comprehension
deficits. They pointed out that, in unselected aphasic patients, comprehension and naming of different materials are highly correlated
(on the order of .6-.7). This is obviously not a telling argument because
lesion size may weil account for correlatins ofthat magnitude. Specific color-naming defects are found only in patients with callosal
syndromes. These researchers concluded that the more interesting
differences between aphasics are in the use and comprehension of
linguistically defined categories (e.g., word dass), and this has been
the dominant research trend for the last 10 years.
4 .1. 3. Models of Selective Aphasias

Although rare in themselves, the pure aphasias are interesting
for the light they may throw on the mechanisms for integrating sensorimotor and language processes. Historkai discussions centered on
the question of whether all the "pure" cases could be explained by a
disconnection of the language areas from input or output. The alternative was to assume separate "centers" for acquired languagedependent skills, and proposals were made for locating centers for
both writing and reading. The second frontal convolution and the
angular gyrus were the respective candidates. This thinking originated in a strict associationistic and hierarchic processing model in
which all perceptual processes take place bottom-up (from sensory
to symbolic Ievels) and all the motor control processes in top-down
fashion.
Recent neuropsychological analyses of dyslexia and agraphia
have led to models emphasizing parallel and alternative coding systems for graphic material. Assurne that reading can be accomplished
SELECTIVE APHASIAS
63
both by a letter-to-sound translation process and by a whole-wordto-meaning matehing procedure. The selective impairment of one of
these routes for reading predicts essential features of the known dyslexia syndromesrather well. The two-routes-of-reading hypothesis as
an explanatory model for dyslexia was proposed by Marshall and
Newcombe (1973), and the analysiswas taken up and extended in a
later book (Coltheart et al., 1980).
In the case of writing, Friederici, Schoenle, and Goodglass (1981)
proposed two independent encoding systems at the word level, one
phoneme-to-grapheme conversion system and one word-to-graphicpattern conversion system. In line with this proposal, Bub and Kertesz
(1982) described the syndrome of deep agraphia in an analogy to deep
dyslexia, which may not be explained as failure of the phoneme-tographeme conversion mechanism.
What these models do is to reject the hypothesis that sensory
systems interact with symbolic systems (language) only at very low
levels of the linguistic code. This condusion may have wide implications, and before discussing it further, I examine some data on
normals.
4.2. Studies of Sensory Mechanisms and Language

in Normals
The main interest of psycholinguists and cognitive psychologists
has been to show the influence of higher levels of information processes on lower level perceptual processes. In the well-known dick
experiments, Fodor and Bever (1965) showed that the perceived temporal location of a dick presented during a sentence was influenced
by the syntactic structure of the sentence.
In the proofreading situation, it is commonly assumed that
attending to the semantic content of the text makes it more difficult
to detect misprints and spelling errors. The Iiterature on memory in
an information-processing context usually makes the assumption that
sensory-based memory (preserving the sensory qualities of the stimulus) is of short duration and that moredurable memory is based on
verbal-symbolic recoding in which information on physical characteristics is lost. (More on memory in the next chapter.)
64
CHAPTER4
If higher Ievel processing interacts with perceptual processes,

there is no a priori reason that the interaction may not be bidirectional.
In that case, the notion of hierarchically organized Ievels becomes
harder to maintain.
An initial indication that linguistic information is coded with
reference to sense modality comes from experiments on the so-called
priming effect. The effect is that words that have been presented
before are easier to recognize than unfamiliar words. It has been found
that, in order to facilitate visual recognition, the repeated words have
tobe presented visually. Increased familiarity on the basis of auditory
presentation does not facilitate visual recognition. This finding led
Morton (1979) to postulate separate visual and auditory wordrecognition units in his Iogogen model. Further evidence for modalityspecific learning in the processing of text comes from studies by Kolers
and coworkers (Kolers, Palef, & Stelmach, 1980) and from Levy (1983).
The latter found that familiarity with a text improves the ability to
detect printing errors. The improvement takes place only in the case
where familiarity is acquired by reading the same text in the same
script. Neither scrambled words in the same script, the same text in
a different script, nor the same text presented auditorily Ieads to any
facilitation in detecting printing errors.
Hasher and Zachs (1979) suggested that some aspects of stimuli
are encoded automatically. Automatie encoding is characterized by
an occurrence independent of variations in age or type of instruction
and by a minimum of effort. There is some evidence that frequency
of occurrence and spatiotemporal context are automatically encoded
stimulus aspects. The work on priming effects referred to above suggests that sensory modality may also be automatically encoded. Lehman (1982) studied recall for word lists given in mixed auditory or
visual presentation. There was uniformly high recall for the modality
of presentation regardless of age and whether the testing for recall
of modality was expected or unexpected. Instructions ensured that
the words were processed for meaning. Even the subjects' recalling
the words themselves by grouping them taxonomically across modalities did not interfere with recall of modality. The author concluded
that the modality of presentation is automatically encoded in longterm memory, and that the importance of modality in the organization
and retrieval of information is unknown.
SELECTIVE APHASIAS
65
Finally, some evidence of the mutual independence of the processes underlying reading and writing in normals may be cited. Speike,
Hirst, and Neisser (1976) found that subjects could be trained to read
a text for meaning while writing from dictation, without interference
between the two activities.
All these studies indicate that the representation of sensory and
linguistic information is highly specific and integrated. The problern
of the activation and retrieval of linguistic information based on input
into a sensory modality may not best be seen as that of connecting
two separate representations but as that of coordinating aspects of a
unitary representation. Unitary in this context means that the linguistic and nonlinguistic codes are integrated. Codes are multiple or
distributed in the sense that the same "chunk" of linguistic information may be coded in different nonlinguistic contexts.

The occurrence of selective (material- or modality-specific)
aphasias in the present patient sample was studied.
For material-specific deficits, the question of the selective impairment of body part comprehension and naming was studied. The
relevant parts of the aphasia test were analyzed with principal factor
analysis to determine if a material-specific factor was present. If so,
the results were analyzed further with relation to aphasia classification. Statistical Package for the Social Seiences (SPSS) (Nie, Hadlai Hull,
Jenkins, Steinbrenner, & Brent, 1975) was used for performing the
analyses.
For modality-specific deficits, factor analysis is not a suitable
method. As in the classification of types of aphasia, an a priori system
was used, based on differences between performances exceeding a
predetermined cutoff point.
4.3.1. Material-Specific Deficits

The relevant parts of the aphasia test are summarized in Table 4.1.
A factor analysis was performed on the results of these tests and
factors with eigenvalues above 1.0 were rotated with the varimax
66
CHAPTER 4
Table 4.1. Selected Aphasia Subtests

Material
Body parts
Objects
Auditory
comprehension
Identify from name (BP-Cl)

Identify from description
(BP-C2)
Perform action (BP-C3)
Identify from name (0-Cl)

Identify from description
(O-C2)
Perform action (O-C3)
Naming
Confrontation naming
(BP-Nl)
Naming of action (BP-N2)
Confrontation naming (0-Nl)

Naming of actions (O-N2)
procedure to determine the loadings of the different tests. The hypothesis was that a factor with loadings from all the tests referring to body
parts could be found.
The factor analysis yielded two factors accounting for 83% of
the variance. The loading of the subtests on these factors is given in
Table 4.2. It is evident that the two factors separated naming and
comprehension rather than body parts and objects. The hypothesis
was thus not confirmed.
Table 4.2. Factor Analysis of Tests with Body

Parts and Objects
Factor 1
BP-Cl
BP-C2
BP-C3
BP-Nl
BP-N2
0-Cl
O-C2
O-C3
0-Nl
O-N2
Factor 2
.78
.70
.70
.88
.91
.84
.84
.76
.87
.91
67
SELECTIVE APHASIAS
4.3.2. Modality-Specific Deficits

The tests used in the assessment of reading and writing are
summarized in Table 4.3. The predetermined criteria for diagnosing
a modality-specific defect used a minimum difference of 20 percentile
points as a cutoff score. The criteria were of the same type as those
used in the definitions of aphasia type (Chapter 3). The definitions
and frequencies for the different categories are given in Table 4.4.
The breakdown with respect to type and severity is shown in
Table 4.5. The association of each selective deficit with the type and
severity of aphasia cannot be tested with a statistical method giving
a measure of interaction. The association with severity is less interesting because the opportunity for detecting selective deficits is not
present in very severe cases (floor effect). Therefore, chi-square tests
of association were performed on the association of nonfluent or fluent
speech with alexia, agraphia, auditory verbal agnosia, and hypergraphia. No significant associations were found.
Bearing the weak statistical basis in mind, I offer some comments
on the apparent trends in the data. Alexia and agraphia were found
more frequently in mild, fluent cases, a finding consistent with traditional notions of an association of these deficits with anomic aphasia
(see Benson & Geschwind, 1977). The trend was more marked for
agraphia than for alexia, reminding us that alexia may also be found
Table 4.3. Tests of Reading and Writing

Reading
comprehension
Point to letters
Point to words
Match word to object
Perform instruction
(6
(6
(6
(5
items)
items)
items)
items)
Reading aloud
Name letters
Say object names
Say polysyllabic abstract words
Say sentences
(6
(6
(4
(5
items)
items)
items)
items)
Writing
Own name
Copy words
Word dictation
Object naming
Sentence dictation
(2
(2
(2
(2
(2
items)
items)
items)
items)
items)
68
CHAPTER 4
Table 4.4. Definition and Incidence of Modality-Specific Deficits

Type
Definition
Alexia without
agraphia
Reading compreh. < aud. comprehension"

Reading aloud < repetition and naming
15
(6)
(2)
Alexia with
agraphia
Agraphia
Writing < repetition and naming
16
(6)
Auditory verbal
agnosia
Reading compreh. > aud. comprehension
17
(7)
Hyperlexia
(pure)
Reading aloud > repetition and naming
10
(4)
Hypergraphia
(pure)
Writing > repetition and naming
31
(12)
(4)
102
(41)
Hyperlexia with
hypergraphia
"Difference exceeding 20 percentile points.
Table 4.5. Relation of Modality-Specific Deficits to Type and Severity of Aphasia
Alexia without agraphia

Alexia with agraphia
Agraphia
Auditory verbal agnosia
Hyperlexia
Hypergraphia
Hyperlexia with
hypergraphia
Total
Severe
nonfluent
Mild
nonfluent
Severe
fluent
Mild
fluent
0
2
1
3
3
1
2
2
1
0
1
4
4
13
7
3
10
8
3
9
17
17
43
25
SELECTIVE APHASIAS
69
in patients with nonfluent speech (Benson, 1977). Hypergraphia was

found in all groups, but the occurrence of about 40% of the cases in
the severe nonfluent group is interesting. One report (Mohr, Sidman,
Stoddard, Leicester, & Rosenberger, 1973) emphasized dissociation
of oral and written responses in global aphasia. Auditory verbal agnosia was found more frequently in the fluent group. Localization of
lesions is discussed in Chapter 7.
4.3.3. Conclusion
One likely reason for these divergent findings can be found in

the literatme reviewed initially. Impairments of more than one mechanism can cause disturbances of function, and statistical criteria as
used here cannot distinguish between disturbances caused by different mechanisms. This is an area in which experimental single-case
studies have been valuable, and the present approach must limit itself
to a descriptive outline.
The findings do not support a notion of material-specific comprehension or naming mechanisms for body parts. Modality-specific
deficits defined by a relative criterion were found in all major groups
of aphasia. This result would argue against viewing these deficits as
a product of only one mechanism, disconnection of the language areas
from input or output. It is not denied that this mechanism exists. It
is, however, also necessary to take into account the integrated nature
of the sensory and linguistic representations and the possibility of
selective deficits of this integrated code.
MEMORY AND LEARNING

DEFICITS
5.1. Normal Memory

What kinds of memory are there? According to Tulving (1982), there
are at least three kinds: procedural, episodic, and semantic. Procedural
memory is concerned with the performance of skills and corresponds
to "knowing how." Episodic memory is memory for events in a spatiotemporal context, whereas semantic memory is conceptual knowledge.
The last two forms of memory are, in principle, different from the
first because the question of veridicality can meaningfully be asked
about them. You believe that you saw John in the stolen car, but is
it really true? You know that all cars have wheels, but is it really so?
Memory may be of a short-lived or a morepermanent kind. The
type of encoding processes performed on stimuli are important in
determining their memorability. The distinction between sensory
memory (iconic or echoic), short-term memory, and long-term memory served originally to delineate memory systems with different decay
functions (Sperling, 1960; Atkinson & Shiffrin, 1968). Major coding
and recoding operations are the transfer from sensory to short-term
memory by phonetic articulatory recoding (Conrad, 1964) and the
transfer from short-term to long-term memory by rehearsal (Atkinson
& Shiffrin, 1968). This by-now-classic theory places great emphasis
on verbal-articulatory mechanisms in coding. lt has been modified in
several ways. Sensory memory that is not verbally recoded is not
71
72
CHAPTER 5
necessarily short-lived. Nonverbal codes have their own long-term

representations (see Crowder, 1976, Chapter 3). Even verbal information may enter long-term or semantic memory without going by
way of phonemic recoding, as discussed in connection with reading
in Chapter 4. A notion of processing memory or working memory
has also gained access to memory theories. It is a memory mechanism
holding information from different memory stores plus information
on the current environment with a view to making decisions. It is fair
to say that memory theories have become less strictly hierarchical and
less inclined to view memory as exclusively stimulus-driven (see
Broadbent, 1984, and the ensuing discussion for a review).
There are no pure tests for memory components, but different
experimental paradigms have been developed to study memory. Different task analyses have been proposed for these paradigms, and
many experimental variations have been used to elucidate these
hypotheses. For convenience, paradigms may be subsumed under
the headings of memory and learning tasks without implying that mechanisms responsible for performance respect these boundaries.
5.1.1. Verbal Memory

Immediate memory is measured in clinical practice by having
the patient repeat a string of digits, letters, or words and noting the
patient's success with different list lengths. This procedure teils something about the capacity for storage, but not about the duration of
memory. Reproduction after variable delays may be introduced to
study information decay, and to prevent rehearsal in the delay interval, a distractor task can be used. Immediatememory does not directly
reflect short-term memory. Reproducing six or seven digits probably
involves more than one memory mechanism. The initial items have
probably been coded in a long-term memory storage, whereas the
last item(s) is coded in short-term memory (STM), the very last item
even in precategorical acoustic storage (Crowder & Morton, 1969).
Glanzer (1972) argued for this multiple-memory-systems explanation
of the position effects in verbal free-recall paradigms, where both the
initial and the final items are better recalled than those in the middle
positions.
MEMORY AND LEARNING DEFICITS
73
5.1.2. Verbal Learning

The paradigms frequently used are the learning of word lists
and paired associate learning. Tulving (1972) pointed out that these
paradigms are not good models for the acquisition of genuinely new
knowledge. Known words with their category membership and associations are used, and hence, what the subject is mainly required to
learn is the particular pattern of co-occurrence of items. Although
stimulus-and-response difficulty may also be manipulated as experimental variables with significant effects, a component of episodic
memory requirement is thus present. Further evidence for factors
other than the use of semantic or conceptual knowledge comes from
studies of serial learning, in which the stimulus sequence is fixed.
Ebenholz (1972) has shown that the serial position of items is coded
and facilitates the learning of a new list in which some old items are
presented in the same serial position as compared with different serial
positions.
5.1.3. The Relation of Verbal Memory and Learning to Language
Function
There is no consensus that memory processes are impaired in
the syndrome of amnesia (see Hirst, 1982). These patients have problems recalling day-to-day events and learning new material when
learning requires several trials. Skill learning is preserved, and the
patients profit from retrieval cues. Episodic memory is believed to be
involved, whereas semantic memory is largely intact. Amnesie patients
are not regarded as having language defects. People with poor conceptual knowledge or bizarre beliefs may be regarded as stupid, ignorant, or insane, but they are not usually characterized as language
deficient.
Another variation of clinical memory defects is the rare cases
with very limited auditory verbal immediate memory described by
Warrington and Shallice (1969) and Basso, Spinnler, Vallar, and Zanobio (1982). The injuries producing such deficits usually involve the
language areas, but the patients are only mildly aphasic.
There are three possible hypotheses about the relation of verbal
memory to aphasia:
74
CHAPTER 5
1. A verbal memory deficit causes the language impairment or

certain aspects of it.
2. The language impairment causes deficient memory because
encoding into the verbal code is impaired. Such encoding is
important for durable memory.
3. Verbalmemory impairments exist independently of language
impairment.
The first hypothesis is unlikely in view of the above remarks on clinical
memory disturbances and language. The second hypothesis seems
almost self-evident in view of the importance of verbal coding processes in theories of memory, but the memory mechanisms in on-line
language processing may only partly overlap with those tested in
memory experiments. The third hypothesis therefore has a fair chance
of success.
5.2. Verbal Memory and Learning in Aphasics

5.2.1. Verbal Memory
Aphasics have poor immediate memory as assessed by their
ability to repeat digits or letters, pointing to digits or letters spoken
or shown, pointing to pictures corresponding to spoken words, or
recognizing words previously heard (Goodglass, Gleason, & Hyde,
1970; de Renzi & Nichelli, 1975; Cermak & Moreines, 1976). To account
for these deficits, some authors have considered the second hypothesis above.
The experimental investigation of rehearsal effects was started
by Heilman, Scholes, and Watson (1976), who hypothesized that the
inability to repeat of Broca and conduction aphasics was the cause of
their poor performance on tasks of immediate memory. The absence
of an interference effect between acoustically similar stimuli had previously been noted by Goodglass, Denes, and Calderon (1974). This
was taken to indicate the nonuse of covert verbal mediation by aphasics in a memory task that involves such mediation in normals. Rothi
and Hutehinsan (1981) used a paradigm that allowed an independent
assessment of immediate memory (repetition) and informational decay
75
over filled and unfilled intervals. Of all groups studied, nonfluent

aphasics showed the poorest immediate reproduction, but no loss of
information with distracting tasks (counting). Fluent aphasics, on the
other hand, showed somewhat better (but subnormal) immediate
reproduction and an interference effect of the distracting task. This
finding leads to the conclusion that all aphasics show a deficit in verbal
memory, but fluent aphasics make some efficient use of a verbal
rehearsal mechanism, whereas nonfluent aphasics do not. As nonfluent aphasics demonstrate significant retention, the question of which
memory mechanisms they use is an intriguing one. Rothi and Hutehinsan (1981) discussed the possibility of direct semantic coding on the
basis of sensory information, the slowness of this process accounting
for the information loss.
The conclusion seems to be that a defect in verbal rehearsal
leading to an inefficient "working memory" (Baddeley & Hitch, 1974)
is present in nonfluent aphasics, but that it does not account for all
findings.
In exploring the third hypothesis, that different memory impairments may coexist with and interact with aphasia, the possibility of
a memory impairment peculiar to nonfluent aphasics has already been
stated (they do not have the type of memory with the decay characteristics of STM).
The presence of an acoustically based memory defect is central
in the interpretation by Luria (see Luria, 1973) of the syndrome of
Wernicke aphasia. In his terminology, it is an "acoustico-mnestic
aphasia," and the basis of it is an auditory perceptual defect not
specific to verbal material. In connection with studies of auditoryverbal short-term memory deficits, Shallice and Warrington (1974)
found normal memory for environmental sounds and concluded that
the deficit studied by them was modality- and material-specific. Gordon (1983) found decrement in auditory immediate recall for both
digitsandtonal stimuli in aphasics with lesions of the Heschl's, superior, and middle temporal gyri, as well as the inferior parietallobule.
Both auditory-verbal short-term memory deficit (Warrington, Logue,
& Pratt, 1971; Warrington & Shallice, 1969) and selective auditoryperceptuallanguage deficit (pure word deafness) (Gazzaniga et al.,
1973) can exist independently of aphasia, but it would not be surprising if any one of these deficits could be associated with aphasia.
76
CHAPTER 5
The presence of a recency effect may give clues about which is the
most common type in association with Wernicke aphasia.
Specific impairment of auditory precategorical storage (PAS)
(Crowder & Morton, 1969) is thought to explain the recency effect in
verbal learning, the effect being that items in the final position of
auditorily presented lists are recalled better than items in middle or
prefinal positions. When verbal material is presented visually, the
effect is not present.
The Iiterature on sequential errors refers to studies by Efron
(1963) and Swisher and Hirsh (1972). These have shown that, in order
to reliably judge the order of two successively presented auditory
stimuli, the aphasic patient requires a time separation one order of
magnitude greater than the normal subject. Moreover, it is patients
with posterior lesions who show this deficit phenomenon to the most
extreme degree. The deficit in perceptual ordering of nonlanguage
auditory signals shows no direct correlation with auditory language
comprehension. Tzortzis and Albert (1974) extended the study of
ordering deficits to language material. Three conduction aphasics
showed retention for content but not for order of words in a shortterm memory task. The authors suggested that this deficit underlies
the repetition deficit in conduction aphasia. The study by Heilman,
Schales, and Watson (1976) opposed this conclusion and found no
qualitative and quantitative differences in memory scores between
Broca and conduction aphasics.
The failure to encode temporal characteristics of the stimulus
sequence may be relevant to the form of memory referred to as episodic
and may be different from the encoding of stimulus content.
In conclusion, it seems that aphasics show distinctive verbal
memory deficits that rhay be specific to the type of aphasia, although
the findings are by no means conclusive.
5.2.2. Verbal Learning

Given that aphasic patients have poor short-term verbal memory
this does not totally determine their capacity for verballearning. Current theories of memory admit of direct coding into long-term memory
77
without short-term storage, and even with limited short-term memory, the possibility of alternative learning strategies may cause differences in the efficiency of learning.
There is a conceptual and methodological difficulty in interpreting deficiencies in verbal learning as memory difficulties. It is
likely that aphasia leads to some alterations in the structure of the
premorbid lexicon (Zurif & Caramazza, 1976). Because new information interacts with already-coded information in long-term storage,
it is difficult to pinpoint the source of a deviance resulting from this
interaction. It is reasonable to describe the findings without making
strong theoretical claims.
Carson, Carson, and Tikofsky (1968) described results for aphasics and controls with common verballearning paradigms including
serial learning. They concluded that the aphasics showed normal
learning curves, but with a generally lower level of achievement than
the controls. Howes and Geschwind (1964) described two groups of
aphasics, Types A and B, of which Type B show disturbances in word
associations. It is natural to assume that they have disturbances relating to semantic coding that interfere with learning that requires
semantic grouping or association. The prediction was apparently confirmed by Beauvois and Lhermitte (1975), who measured immediate
memory span for words and the learning of eight-word lists. Patients
with semantic paraphasia showed normal immediate memory and
severe learning impairment. Patients with exclusively phonemic paraphasia showed reduced immediate memory span but normallearning.
5.3. Nonverbal Memory and Learning

Tests of nonverbal memory and learning have been used in
research that contrasts patients with left- and right-hemisphere
pathology. A series of studies from the Montreal Neurologkai Institute have shown the resulting impairments to be material-specific;
that is, patients with left-temporal-lobe pathology showed deficits
with verbal but not with nonverbal material, and patients with righttemporal-lobe pathology showed the reverse pattern. These patients
were only minimally aphasic or nonaphasic (see Milner, 1974).
78
CHAPTER 5
Oe Renzi and Nichelli (1975) used some of the methods developed in the above research to study a broader range of patients with
left- and right-hemisphere pathology. Using the Corsi block-tapping
test (see description in Section 5.5.1), a nonverbal analogue of digit
span, they found that patients with right-hemisphere lesions were
inferior to patients with left-hemisphere lesions, who, in turn, were
inferior to controls. The presence, type, or degree of aphasia was not
important in explaining the results. The authors are inclined to believe
that the deficit in the left-hemisphere group was related to a disturbance of attentional factors present mainly in patients with posterior
lesions. Oe Renzi, Faglioni, and Previdi (1977) extended the previous
research by including a learning task with block-pointing sequences
of supraspan length. Again, it was the patients with visual field defects
who did poorly, especially those with a right-hemisphere lesion. The
presence or the type of aphasia was not reported in this study. The
authors reported another study in which a subspan sequence of three
blocks was reproduced after filled or unfilled intervals of 6 or 18
seconds. All groups showed some loss of information even without
interference, but the loss was increased with a verbal interference task
(counting). The patternwas the samein controls andin patients with
right- and left-hemisphere injuries.
Cermak and Tarlow (1978) tested memory for words, pictures
of objects, and nonsense shapes in a continuous recognition paradigm. Although severely impaired for words, the aphasics showed
normal memory for pictures. The nonsense shape task proved too
difficult for even the control group. It was concluded that the memory
deficit showed by the aphasics was material-specific, and not related
to any attentional or perceptual difficulty with the stimuli. The aphasics all had nonfluent speech.
5.4. Conclusion
The above discussion leaves many unresolved questions. First
of all, there are the questions of the specificity of the memory disorders
for verbal material. For short-term or immediate memory tasks, it has
been well documented that specific deficits exist that are neither a
79
direct correlate of the severity of the aphasia nor an aspect of a generalized memory disorder.
To the extent that some of the memory defects reported in aphasia
refer to the spatiotemporal encoding of verbal sequences, there seems
to be no reason why they should be specific to verbal material. It is
consistent with a material-specificity hypothesis that, as long as tasks
are matched for difficulty, the encoding of the nonverbal aspects of
the task or the use of response strategies should not be specific to
material type. If the material-specificity hypothesis is unfounded, then
the alternative hypothesis is that task types are organized as integrated wholes cerebrally, and not composed of constituents that can
be varied independently. The learning tasks may serve to throw light
on this problem.
Second, there are the questions of subtypes. Above it has been
hinted that at least two types of short-term-memory deficits may exist,
one material-specific and another modality-specific. To the extent that
these are based, in the one case, on a close association with defective
rehearsal and, in the other, with defective precategorical acoustic
storage, a study of serial position curves may be useful in dissociating
them. Research has focused on the association of short-term-memory
defect with specific forms of aphasia, especially with conduction
aphasia. In view of the pervasive difficulties that aphasics have with
such tasks, it does not seem likely that the difficulties are restricted
to specific aphasia types, although this question is still open. The
careful definition of type as independent of severity of aphasia seems
crucial in this context.
5.5. Present Study

The study compares the results of different procedures and paradigms, described in more detail below, to throw light on the questions raised about the material specificity and the nature of verbal
memory and learning problems in aphasics. A distinctive feature of
the approach was the attempt to apply parallel tasks of a verbal and
nonverbal nature to the study of these problems.
The subjects for the study are the population of 249 patients
described in Chapter 2. For tasks demanding verbal performance, it
80
CHAPTER 5
was reasonable to exclude patients with insufficient verbal ability to

comply with the instructions, and their exclusion led to a lass of cases
in different analyses. To avoid unnecessary detail, this exclusion is
not discussed for every test, but the requirements for inclusion are
stated; for the most demanding tests, 35% to 40% of the patients were
excluded.
First, the tests used are described and discussed with regard to
findings in previous studies with aphasics and with control groups.
Then, the structure of the memory functions underlying the performance on the tests is analyzed with the method of factor analysis. Last,
the relation of memory performance to aphasia groups is studied.
The format adopted for analysis is the cube model described in Chapter 2, with ANOVA performed an the three main dimensions (type,
severity, and chronicity) and testing for two- and three-way interactions (SPSS; Nie et al., 1975).
5.5.1. Tests
A summary of the tests used and the scores and measures derived
from them is given in Table 5.1.
5.5.1.1. Verbal Immediate Memory
5.5.1.1.1. Digit Span. The digit span test is the repetition in the
same order of a series of digits presented orally. The procedure for
administration and scoring is as in the WAlS-test (Wechsler, 1958).
The results were not taken as an indication of memory, unless the
patient could repeat one digit correctly.
The normal performance an span forward and backward was,
an the average, 6 + 5 according to the Norwegian standardization
(Engvik, Hjerkin, & Seim, 1980). Costa (1975) found the mean span
forward and backwardtobe 5.2 and 3.3 for left-brain-damaged patients
and 5.5 and 3.4 for right-brain-damaged patients. In left-lesion nonaphasic patients, de Renzi and Nichelli (1975) reported a mean span
of 5.7 forward. Black and Strub (1978) divided their left-lesion group
into patients with frontal and posterior lesions and found 5.8 forward
and 4.2 backward for frontals and 4. 9 and 4.1 for posteriors. The
group included 8% aphasics. The question of whether digit span
forward and digit span backward measure the same underlying function has been discussed. Rudel and Denckla (1974) suggested that
81
Table 5.1. Summary Table of Tests Useda

Function
Test
Derived measures (code)
Range
Verbal
immediate
memory
Digitspan
Digit span forward (DF)
1-8
Pointing span
Digit span backward (OB)

(PS)
1-8
1-18
I rials (OS-T)
1-17
0-1
Verbal
learning
Digit serial
learning
Error type: Perseveration

(DSE-P)
Sequence (DSE-S)
Intrusion (DSE-I)
Refusal (DSE-R)
Position score: Initial (DSP-I)
Middle (DSP-M)
Last (DSP-L)
Verbal
association
Easy-item score (WPA-A)

Difficult-item score (WPA-B)
0-1
0-1
0-1
0-2
0-2
0-2
0-2
0-18
0-9
Nonverbal
immediate
memory
Block pointing
Block span forward (BF)

Block span backward (BB)
1-10
1-10
Nonverbal
learning
Block serial
learning
Trials (BS-T)
Error type: Perservation (BSE-P)
Sequence (BSE-S)
Intrusion (BSE-1)
Refusal (BSE-R)
Position score: Initial (BSP-1)
Middle (BSP-M)
Last (BSP-L)
Trials (SA-T)
1-17
0-1
0-1
0-1
0-1
0-2
0-2
0-2
0-20
Shape
association
"For description, sec text.
span backward is partly determined by visuospatial abilities. Costa

(1975) found some confirmation of this, but Richardson (1977) and
Black and Strub (1978) did not.
5.5.1.1.2. Pointing Span. Pointing spanwas constructed after
the procedure used in a study by Goodglass et al. (1970). The test
materials were two displays consisting of cardboard plates with six
objects pictured on each. The subject was instructed to point to the
pictures as they were named by the examiner. lf she or he failed more
82
CHAPTER 5
than once on each display, the testwas discontinued, and the score
was not used as an indication of memory function. If the criterion
was met, the test was continued by asking the patient to point to
series of pictures of up to five items in the order named by the examiner. The test items were prerecorded on tape and were played back
through a loudspeaker from a Tandberg 3000 X tape recorder. The
display with pictures was covered while the patient listened to the
tape. Removing the coverwas the signal for the patient to start pointing. One point was awarded for the correct performance of a task,
and the results for two displays were summed foratotal score.
5.5.1.2. VerbalSerial Learning. Afterdigitspan forward had been
established, a new series of digits was presented containing two digits
more than the estimated span. The order of presentation remained
constant. Repetition of this series continued until a criterion of perfect
recall on two consecutive trials was reached. Responses after each
trial were noted. The score was the number of trials needed to reach
criterion. Hamsher, Benton, and Digre (1980) found that normal subjects leamed sequences of eight digits without problems even in groups
of high age. For patients with a span of five digits or less (up to seven
digits in the leaming task), a scoring of performance on the first,
middle, and last digitwas performed on Trials 1 and 2. They were
scored as correct when present regardless of ordering. For patients
who had not mastered the task in seven trials (the mean score), the
error type most prevalent was scored. The predetermined alternatives
were perseveration, intrusion of erroneous items, faulty sequence,
and refusal to continue. The responses on each trial were recorded,
and judgment on error type was performed by the author on the basis
of the protocols. A patient was permitted to score on, at most, two
error types. Ratings were given as presence or absence of the given
type.
5.5.1.3. Verbal Associative Learning. A Wechsler paired-associate
learning test, easy and difficult items, was taken from the Wechsler
memory scales (Wechsler, 1945). No Norwegian standardization is
available, so the author's translation was used. A series of 10 word
pairs were read three times. After each presentation, recall was tested
by the examiner's saying the stimulus word and the subject's attempting to recall the response. The order of presentation was changed
between reading and recall of the Iist, and between each reading of
83
the list. The 10 word pairs contained 6 easy pairs (up-down, northsouth, metal-iron, baby-cries, fruit-apple, rose-flower) and 4 difficult pairs
(cabbage-pen, obey-inch, crush-dark, in-also). In the standard procedure
for scoring the test, the results for easy and hard items are summed.
In the present study, they were treated as separate scores. For the
results to be used in the analyses, the patient had to be able to repeat
single words.
Previous studies of aphasics with this test are not known to the
author. The differences between aphasic subgroups in semantic associations have been described by Howes. He found defects in what he
called "group B" aphasics (Wenicke) as opposed to "group A" aphasics (Broca) (Howes & Geschwind, 1964).
In her large study of head injuries acquired in World War II,
Newcombe (1969) included a test of association learning in which the
subjects learned three unrelated pairs of items. She found no greater
deficit in left- than in right-hemisphere injuries, but patients with
parietal injury did poorly. Milner (1962) found deficit in patients with
left temporallesions for the paired-associate task from the Wechsler
memory scales (the same as those used here), but only patients with
temporallesions (left or right) were tested.
5.5.1.4. Nonverbal Immediate Memory. The block-pointing-span
testwas similar to one used by Corsi (1972) for the study of memory
functions. On a square board (20 X 20 cm), 12 blocks were mounted
in a random arrangement. Their dimension was 2 x 2 x 2 cm (see
Figure 5.1). On the side facing the examiner, the blocks were numbered 1 through 12.
The patient was instructed to point to the blocks shown by the
examiner in the same order. The examiner pointed to the blocks one
by one at the rate of one block per second. The number of blocks
pointed to by the examiner was increased by one until the patient
failed two consecutive trials. The procedure was repeated with
instructions to point to the blocks in opposite sequence to that shown
by the examiner. The score was the number of items in the Iongest
series of blocks pointed to correctly forward and the number of items
in the Iongest series of blocks pointed to correctly backward.
The original block-tapping test by Corsi (1972) has nine blocks.
The normal span for a control group with a mean age of 28 years was
given as 4. 6 by Corsi. The same test has been used by other researchers
84
CHAPTER 5
[]
[]
Figure 5.1. Block-pointing test.
who have mostly followed the procedure given by de Renzi and

Nichelli (1975), in which two trials at every length are given and a
half point is awarded for correct performance of the second trial.
According to these authors, the presence or type of aphasia is not
associated with a deficit on this test.
5.5.1.5. Nonverbal Serial Learning. After block-pointing span forward had been established, a new pointing sequence was constructed
containing two items more than the span. The sequence was demonstrated repeatedly until a criterion of two consecutive perfect reproductions had been reached. Recall was attempted after each trial, and
the number of correct items recalled was noted. The score was the
number of trials to reach criterion.
In the patients scored for position effects on the verbal serial
learning task, the same scoring was performed on the first two trials
of the block-pointing sequence. The error types were evaluated in the
same manner as with the digit learning task.
The same task (span + 2) was used by de Renzi, Faglioni, and
Previdi (1977). This task was surprisingly difficult for their subjects,
the controls needing 12.8 trials to reach the criterion. One suspects
that their measure of span was the cause of the difficulty. Assurne
that a patient repeats all sequences of two, three, and four blocks
85
correctly, but none of the Ionger sequences. The scoring system would
give the patient a score of 6 points, and he or she would be required
to learn a series of eight blocks. In the present study, the span would
be scored as 4 points, and the learning task would be to repeat a
sequence of six.
5.5.1.6. Shapc Association Learning. Six figures from the material
of nonsense shapes constructed by Vanderplas and Garvin (1959)
were selected. They were divided into three pairs. In Presentation 1,
the pairs were presented to the patient simultaneously, each column
forming a pair. The upper row was designated stimuli and the lower
responscs. The instructions were that the examiner had made an arbitrary decision that certain figures belonged together. The subject was
to inspect the array for 30 seconds and then try to remernher which
shapes went together. The response shapes were then removed, and
the stimulus shapes were rearranged in the left-to-right order (Presentation 2). The subject was handed a test shape (T) and was asked
to match it to the corresponding stimulus. The responsewas placed
with the selected stimulus, and a new response item was presented.
When three choices had been made, the examiner rearranged the
pairs into the correct combinations. This terminated the trial. The
subject was allowed to examine the correct arrangement briefly before
the next trial. The procedure was repeated until two consecutive trials
had been performed correctly. The steps of the procedure are summarized in Figure 5.2. The scorewas the number of trials to criterion.
The Vanderplas and Garvin stimulus material has been used in
previous studies with aphasics, but not administered in the same
way. Cermak and Tarlow (1978) used the material in a continuous
recognition-memory paradigm and found it too difficult tobe informative. Oe Renzi, Faglioni, and Villa (1977) used it in a study asking
patients to sort eight patterns in a prescribed sequence. The lefthemisphere group, of which half were aphasic, performed only marginally worse than normal controls.
5.5.2. The Structure of Memory in Aphasia

As noted, the relation of memory tasks to memory functions is
complex, and the question of which memory functions are measured
by the tests and the derived scores in the aphasic group is highly
86
CHAPTER 5
Presentati.on 1
s
R
ProsentaUon 2
s
R
T
Figure 5.2. Shape association test.
pertinent. Therefore, the author, in collaboration with K. Sundet,

performed factor analyses, first on the set of verbal and nonverbal
measures separately and then on the combined set of measures. The
method used was principal-component analysis to identify factors,
an examination of eigenvalues or scree test (Cattell, 1978) to decide
how many factors to include, and varimax rotation to determine factor
loadings.
For the verbal scores, four factors with eigenvalues higher than
1.0 were found, and they accounted for 63% of the variance. The
contributions of different tests to the factors are shown in Table 5.2.
In general, a test is listed only on the factor to which it gives the
highest contribution. When the values are close, the same test is listed
under two factors.
The following interpretation may be suggested of the factors:
Factor 1 is an immediate memory factor. In addition, the ward-
MEMORY ANO LEARNING OEFICITS
87
Table 5.2. Factor Composition of Verbal Tests

Factor 1
Factor 2
Factor 3
Factor 4
OF (.81)
OB (.78)
PS (.78)
WPA-A (.75)
WPA-B (.56)
OSP-F (.62)
OSE-R (.84)
OS-T (.71)
OSE-P (.64)
OSE-S (.64)
OSP-M (.79)
OSP-L (.39)
OSE-1 (.60)
association tests load on this factor. Thesetests were performed poorly,

and the results may indicate that the main basis of performancewas
rote immediate memory.
Factor 2 is a seriallearning factor with high loading on the error
type of refusal and on performance on the first item. These features
are shown elsewhere to be associated in nonfluent aphasics. Factor
3 is another serial learning factor with trials, perseveration, and
sequence errors loading highly. The two latter features are shown
elsewhere to be characteristic of fluent aphasia. Factor 4 is represented
by serial learning measures of performance on the middle and last
items and errors of intrusion. These features may be associated with
a general failure to learn over trials and, more specifically, with a
recency effect, although the latter is weak.
The factors account for all verbal measures, least satisfactorily
for performance on the last item in digit seriallearning, which loads
only .39 on Factor 4.
The results for nonverbal tests are shown in Table 5.3. The criterion of eigenvalues above 1.0 results in five factors accounting for
68% of the variance.
The first factor is a nonverbal immediate memory factor with
high loading on blocks forward and backward. Factor 2 is a serial
Table 5.3. Factor Composition of Nonverbal Tests
Factor 1
Factor 2
Factor 3
Factor 4
Factor 5
BF (.84)
BB (.76)
SA-T (.41)
BS-T (.67)
BSE-P (.68)
BSE-S (.80)
BSE-1 (.58)
BSE-R (.81)
BSP-M (.78)
BSP-L (.60)
BSP-F (.86)
SA-T (.43)
88
CHAPTER 5
learning factor with trials associated with perseveration and sequential errors. It shows an interesting parallel to Factor 3 in the structure
for verbal memory. Nonverbal Factor 3 associates two error types:
intrusion and refusal. Factor 4 associates performance on the middle
and last items on the block seriallearning. Factor 5 is best represented
by one single measure, the performance on the first item of the block
serial learning. The factor solution accounts for all nonverbal measures, but most poorly for paired-shape association, which Ioads only
.41 on Factor 1 and .43 on Factor 5.
In the design of the tasks, an attempt was made to construct
parallel tests for measuring memory for verbaland nonverbal material. The results show that the intention was fulfilled by revealing a
parallel structure of memory performances of aphasics.
The combined analysis of verbal and nonverbal tests yielded, in
all, 10 factors with eigenvalues above 1.0. The scree test (Cattell, 1978)
was used to Iimit the number of factors studied with varimax rotation,
and five factors were included, accounting for 49% of the variance.
Tests and factor loadings are shown in Table 5.4.
Factor 1 incorporates Factor 1 of the verbal factor analysis. In
addition, intrusion errors on digit serial learning and paired-shape
association are included. It is reasonable to maintain the interpretation
of this factor as mainly a verbal immediate-memory factor. Factor 2
shows an interesting coupling of Factars 1 and 2 from the nonverbal
analysis with Factor 3 from the verbal analysis. Immediate memory
for block sequences shows the highest loadings, but the factor is also
represented by other measures relating to learning and reproducing
Table 5.4. Factor Composition of Combined Verbaland Nonverbal Tests

Factor 1
Factor 2
Factor 3
Factor 4
Factor 5
DF (.58)
OB (.54)
PS (.66)
WPA-A (.78)
WPA-B (.66)
OSE-I (.62)
SA-T (.53)
BF (.68)
BB (.62)
BS-T (.54)
BSE-P (.59)
BSE-S (.61)
OS-T (.47)
OSE-P (.40)
OSE-S (.40)
OF (.58)
OB (.46)
OSE-R (.77)
OSE-S (.40)
OSP-F (.61)
BS-T (.64)
BSE-R (.54)
BSP-F (.38)
BSP-M (.75)
BSP-L (.51)
OSP-M (.34)
OSP-L (.38)
89
sequences of verbal and nonverbal material. The finding of a sequencing factor is interesting in relation to the claim that sequencing is a
distinctive error category in aphasia (Tzortzis & Albert, 1974). In studies of normal seriallearning, it has been shown that a schema of the
sequential structure of the Iist is learned independently of verbal
associative relations (Ebenholz, 1972).
Factor 3 is a verbal serial learning factor similar to Factor 2 of
the verbal analysis, except that some oftheimmediate memory measures appear again. Factars 4 and 5 are mainly nonverbal learning
factors relating to error types and position effects. Under these factors,
BSP-F, DSP-M, and DSP-L have been listed because this is where they
show their highest factor loadings within the present solution.
It is tobe expected that more of the specific factors of the verbal
and nonverbal solutions will appear if a greater number of factors are
analyzed. The present analysis is, however, sufficient to bring out
the important point that material specificity is only partly preserved.
Same factors are relatively purely material-specific, whereas some,
notably Factor 2, combine measures by a different principle. The
complexity is also brought out by the fact that tests may weil show
moderate factor loadings on several factors, some material-specific
and some not. An example is DSE-S, which Ioads moderately on
both Factor 2 and Factor 3 (a sequential factor and a verballearning
factor).
5.5.3. Relations of Memory to Aphasia Group

In these analyses, tests are grouped according to the functional
domains suggested in Table 5.1, whereas the factor structure is taken
into account in the discussion. In the analyses, the age bias has been
corrected for statistically. The results are summarized in Table 5.5 and
5.6, and a brief discussion of each functional area follows. In Table
5.5, the nonverbal functional areas are not shown because no significant relationships were found.
5.5.3.1. Verbal Immediate Memory. On all types of tests, there
were significant main effects of type and severity of aphasia (Tables
5.5 and 5.6). In general, nonfluent aphasics performed worse than
fluents on all tests. Significant interachans were found only on DF.
The interaction of type and severity (F = 2.81, p < .05) was caused
90
CHAPTER 5
Table 5.5. Summary Table of Tests Showing Relation to Type of Aphasia

Test with relative deficit
Function
Nonfluent
Verbal immediate memory
DF (p < .01)
DB (p < .05)
PS (p < .05)
Verbal seriallearning
DSP-F (p < .05)

DSE-R (p < .01)
Fluent
DSE-P (p < .05)

DSE-S (p < .05)
by the relatively poor performance of mild nonfluents as compared

with mild fluents. The observed interaction confirms the finding of
Goodglass, Gleason, and Hyde (1970) that Broca aphasics have strikingly poor immediate memory span, except that their study used a
test closely similar to PS. The other significant interactionwas severity
and chronicity (F = 7.18, p < .01). The main source of the interaction
appeared tobethat the mildly aphasic acute cases performed disproportionally better than the mildly aphasic chronic cases.
Table 5.6. Summary Table of Tests Showing Relation to Severity of Aphasia

Function
Tests with deficit

in severe aphasia
(p < .001)
(p < .001)
(p < .001)
Verbal immediate memory
DF
DB
PS
Verbal seriallearning
DSP-F (p < .01)

DSE-1 (p < .01)
DSE-R (p < .01)
Verbal associative learning
WPA-A (p < .001)

WPA-B (p < .001)
Nonverbal immediate memory
BF
BB
(p < .03)
(p < .001)
Nonverbal serial learning
BS-T
(p < .01)
Nonverbal associative learning
SA-T
(p < .01)
91
5.5.3.2. VerbalSerial Learning. The number of trials to criterion

yielded no significant main effects or interactions. Interesting group
differences appeared, however, both in error types and in position
curves for the first two trials. The position curve data show a superiority of fluent over nonfluent aphasics in recalling the first item, but
not the middle or the last.
No significant main effect or interactionwas shownon the final
item, but the visual form of the curves gives the impression of an
absence of recency effect in the fluent group, especially the severe
fluents (Figure 5.3). A direct test of the significance of the difference
between the first and the middle and between the final and the middle
positionswas therefore performed within all groups (Table 5.7). An
analysis of variance was performed to test for an overall effect of serial
position on memory score. lf the resulting F score was statistically
significant (p < .05), then the differences between the initial versus
the middle and the last versus the middle positions were tested for
significance in order to reveal the specific locus of the position effect.
The results show that all groups had a significant primacy effect, but
only the nonfluent groups had a significant recency effect.
Despite an apparent advantage on the first two trials in coding
information in long-term storage, the fluent patients performed no
better than the nonfluents. The explanation may be their tendency
to make errors of perseveration and sequential ordering, errors related
exclusively to type and not to the severity of the aphasia.
On intrusion errors, in addition to the significant main effect of
severity there was an interaction of type and severity. The difference
between mild and severe was particularly striking for fluent patients
(F = 2.66, p < .05).
Refusal to continue was most characteristic of nonfluent patients.
In addition to being significantly more common than in fluent patients,
it was also the most frequent error type in nonfluents. The result is
consistent with the finding that a so-called catastrophic reaction
(depression, rejection, withdrawal) is a reaction type found more
frequently in this group (Robinson & Benson, 1981). This type of
reaction is, however, also strongly related to severity of aphasia, and
the two effects are independent.
5.5.3.3. Verbal Associative Learning. The results (Table 5.7)
show a strong relation to the severity of the aphasia and no significant
interactions.
CHAPTER 5
92
RecLL probabilLty
RecaLL probabili.ty
1.0
NF -Se.ero (N-36)
.9
1.0
.9
-8
.8
.7
.7
.6
.s
.4
,J
\/
.
.6
.5
-~
:;.
.1
.1
miLidle
.3
.2
[II'St
NF-Hiki !N'IJ
fvst
last
Position
mitti.le
Positi.ort
Recai.L probabillty
Recall probability
to
1.0
F-SrNere ( 11=19)
.9
.9
.8
.?
.7
.6
.6
.5
.5
.4
.3
.3
.2
.z
.1
.1
fir3t
mutdie
Posititm.
last
F- Mild
last
(/J70)
\~
prst
mutdiE
last.
Posilton
Figure 5.3. Position effects in seriallearning. Key: F, fluent; NF, nonfluent.
5.5.3.4. Nonverbal Immediate Memory. The results show a strong

relation to severity but no significant relation to or interaction with
other dimensions.
5.5.3.5. Nonverbal Learning. The score on trials on block serial
learning was related to the severity of the aphasia. There were no
significant differences between groups in performance on the first,
93
Table 5.7. Within-Group Position Effects

Digits
Blocks
Position
effect
F
Severe nonfluent
Mild nonfluent
Severe fluent
Mild fluent
Note. I
initial item; M
3.29
7.03
5.86
22.18
=
Position
effect
p
I>M L>M
.04
.01
.01
.001
.01
.001
.01
.001
4.50
2.36
1.90
5.42
middle item; L
.02
.05
n.s.
n.s.
.01
n.s.
n.s.
.01
I>M L>M
.02
n.s.
.01
n.s.
last item.
the middle or the last items. As shown in Table 5.7, two of the groups
(severe nonfluent and mild fluent) showed a significant position effect,
and the other two did not. In both cases, the significant effect was a
primacy effect.
The error types on block serial learning showed no significant
relation to either the type, the severity, or the chronicity of the aphasia.
The lack of a significant relation does not mean that errors were not
made. The data show sequential error to have been present in 40%
of the cases, perseveration in 25%, intrusion in 26%, and refusal in
5%.
The results on paired-shape association show significant relation
to severity but not to other dimensions of aphasia.
5.5.4. Discussion
For a task to show a material-specific deficit, it should show no
difference from a control group when the noncritical material is tested,
and the performance on noncritical material should be uncorrelated
with performance on critical material.
In the study by de Renzi and Nichelli (1975) of the block-pointing
test, both conditions were satisfied. In the present study, only the
latter hypothesiswas tested, and the condition for material specificity
was not satisfied. Both the block-pointing test and all other nonverbal
tests showed a highly significant relationship to severity of aphasia.
The explanation of the discrepancy is probably connected with the
94
CHAPTER 5
likelihood of detecting weak statistical effects. The highly significant

F ratios relating severity of aphasia to nonverbal memory and learning
defects correspond to correlations between .27 and .36, and the high
number of subjects tested, including many severe cases, contributed
to the significant result.
The complexity of the material-specificity issue is further clarified
by the factor analyses, which show that no extreme position can be
taken for or against specificity.
When subgroups are examined the material specific nature of
the deficit is more sharply focused. Whereas the performance with
verbal material shows several features related distinctly to type of
aphasia, this is not true of nonverbal material. In no case is the nonverbal deficit related to type of aphasia. In the direct test of relations
to groups, there is no indication that specific error types in serial
learning or differences in the initial, the middle, and the last position
on the serial position curve found for verbal material are also found
for nonverbal material. The factor analyses, however, revealed parallel features of organization across material types.
With verbal material, the pervasive effect of severity of aphasia
is striking on all tests except digit seriallearning, both trials and error
types. This exception may be taken to indicate that the procedure of
adjusting the length of the digit series to the memory span rather
than using a series of fixed length for all subjects is successful in
eliminating nonspecific severity effects. When adjusted in this way,
the tasks yield valuable differential information.
The evidence for subtypes of deficits in verbal memory comes,
first, from tasks of immediate memory, which were performed more
poorly by nonfluent than by fluent aphasics. The fact that mild nonfluents had a strikingly poor performancewas noted above. From an
examination of the serial position curves, it appears further that nonfluents were inferior to fluents on the first item of a digit series, but
not on the middle or final items. The error type most characteristic
of nonfluent patients was refusal to continue.
Fluent aphasics had a better immediate memory span than nonfluents and better retention of the initial item of a series of digits.
Still, they failed to show an overall advantage in verballearning. In
the serial position curve, they failed to show an advantage of the final
compared to the middle item in recall, thus supporting the idea that
95
they had a weak acoustic memory (precategorical acoustic storage).

They had a tendency to make certain error types, specifically perseveration and sequential errors. Assuming that fluent aphasics have
posterior lesions, one cannot take the result to confirm the notion frequently expressed that perseveration is a sign of frontallobe pathology
(Luria, 1966). The findings suggest that it is not so much perception
of sequence but response organization and reorganization that fails
in the fluent aphasics. The nonavailability of the final item as an
anchoring point for response organization may be critical. This
hypothesis is supported by the fact that the relation of these error
types to fluent aphasia disappeared in the block-pointing seriallearning task, in which there was also no difference between aphasia types
in performance on different serial positions.
The verbal association tests yielded no differential information,
and the expected semantic impairment in fluent aphasics could not
be demonstrated. Chronicity did not come out as a significant main
effect in any analysis. There is, thus, no type of memory deficit that
is more characteristic of acute aphasics than of chronic.
A search of the aphasia registry for patients with mild aphasia,
low verbal immediate memory, and high nonverbal immediate memory yielded no cases among the 249 patients analogaus to those
reported by Warrington and Shallice (1969) or by Basso et al. (1982).
5.5.5. Conclusion
I suggest in conclusion that severe aphasia is associated with
reduced encoding of information. This is more true of verbal than of
nonverbal material, but only relatively so. The view that verbal memory problems are just secondary to a general verbal encoding deficit
(severity of aphasia) must be rejected because there are clear indications of subtypes, as described above.
The indications of specificity found in the material show that
the view implied in the material-specificity thinking-namely, that
tasks can be decomposed into independent components-does not
hold. The factor analyses give the impression of multiple, overlapping
organization of function, so that a given aspect of a task may be
represented both in a purely verbal factor and in factors more related
to the structural properties of tasks (e.g., sequential organizations).
96
CHAPTER 5
When the severity dimension of aphasia is analyzed, the results

seem to reflect the deficit in multiple dimensions of organization of
memory. The memory deficits in severely aphasic patients are thus
probably complex, reflecting a verbal deficit, a deficit of spatiotemporal encoding, and a deficit of response organization.
DEFECTS OF VISUAL
NONVERBAL ABILITIES
The functional domain covered by the above title is not weil defined.
Excluding tests of memory it is not clear how theories of normal
mental abilities would structure the domain of remaining nonverbal
functions.
Taking a logical approach, we may classify tasks as varying in
three dimensions, complexity of stimuli (spatial configuration), complexity of response (coordinated movement), and complexity of intervening functions (logical principle for linking stimulus and response).
We may then hypothesize that deficits will reflect this dimensional
structure, and this hypothesis can serve to structure the discussion
of deficits in aphasia.
6 .1. Visual Nonverbal Functions in Aphasia

It is doubtful if perceptual complexity in itself accounts for a
dimension of the deficit in aphasia. Clinical studies have found deficits

in tests combining perceptual complexity and either manipulative
responses or symbolic complexity. In the first case, the deficits are
classified as visuoconstructive and measured by performances in
drawing or puzzle-type tasks. There is an extensive Iiterature on the
performance of aphasics on such tasks (see de Renzi, 1982, Chapter
9). Although it is acknowledged that patients with right-hemisphere
97
98
CHAPTER 6
mJury perform such tasks poorly, the discussion has been about
whether there is a difference in performance in favor of left-hemisphere-injured patients. Some studies (Arena & Gainotti, 1978) indicate that this is not the case. The question of qualitative differences
between left- and right-hemisphere injuries is complex. Hecaen and
Assal (1970) hypothesized that the deficit in left-hemisphere injuries
was on the executive rather than the perceptual side and found that
guidemarks aided the copying of a cube in left- but not in righthemisphere injuries. Other qualitative differences noted in the Iiterature are lack of detail with preserved spatial organization in lefthemisphere injuries and distorted spatial organization and neglect of
the left side in right-hemisphere injuries. The studies summarized by
de Renzi (1982) were unable to consistently quantify or reproduce
these clinical observations.
Moreover, the idea that the deficit is executive and not perceptual is Contradieted by the failure to reproduce the findings of Hecaen
and Assal (1970), as well as by low scores on spatial perceptual tests
with multiple-choice alternatives in both left- and right-hemisphereinjured patients with constructional apraxia (Arena & Gainotti, 1978).
Further evidence for a deficit of perceptual function comes from the
studies of the Gottschaldt hidden figures test (Teuber & Weinstein,
1956; Russo & Vignolo, 1967), which is performed more poorly by
aphasics than by any other group with localized injuries. The test
undoubtedly also makes intellectual demands and was interpreted as
showing an intellectual deficit in aphasics by Teuber and Weinstein
(1956). Tests suitable for demonstrating an intellectual deficit should
not require elaborate instruction, the stimuli should not be complex,
and the responses should be simple. Weinstein (1964) summarized a
series of studies showing nonverbal deficits, including visual conditional reaction, a paradigm taken from animallearning. Carson et al.
(1968) found anormal acquisition curve for simple identification learning, but inferior performance on an alternative reaction task. These
findings are consistent with the observation that even severe aphasics
(globals) can learn to attach meaning to stimulus cards with abstract
shapes (Gardner, Zurif, Berry, & Baker, 1976). The failure to perform
more complex tasks (conditional or alternating reaction) may indicate
an intellectual disturbance. Further evidence comes from a study by
DEFECTS OF VISUAL NONVERBAL ABILITIES
99
Basso, de Renzi, Faglioni, Scotti, and Spinnler (1973), who, after Controlling for confounding influences with a covariance technique, concluded that there is a specific deficit of visual-intellectual abilities with
posterior left injuries. A newer study by Basso et al. (1981) did not
clarify the issue further. Other studies with the Raven test (Kertesz
& McCabe, 1975) found a moderate correlation between test score
and degree of aphasia.
Studies with the Wechsler intelligence scales (WAlS; Wechsler,
1958) reported poor performance in aphasics (Orgass, Hartje, Kerschensteiner, & Poeck, 1972). For some scales, the results were worse
for aphasics than for right-hemisphere-injured patients. The problern
with interpreting these results is that many of the patients performing
poorly showed constructional apraxia. Indeed, some of the subtests
of the WAIS (Block Design) have been used as instruments for measuring constructional apraxia (e.g., Black & Strub, 1976). Some authors
have therefore corrected for constructional apraxia before assessing
the relationship of aphasia to "intelligence" (Borod, Carper, & Goodglass, 1982). If this correction is made, the effect of aphasia tends to
vanish. In this context, I have tried to avoid a discussion of "intelligence" and have used the moreneutral term visual nonverbal abilities.
I agree with the position of Harnsher (1982), who stated "it is not
clear how one can separate constructional praxis from nonverbal intelligence" (p. 344). The point may be not to control for constructional
praxis, but to pointout that aphasics show neither a general perceptual nor a general intellectual deficit. A more specific association of
these factors is required to bring out a deficit.
6.1.1. Apraxia
Although the visual-intellectual deficits are multidimensional and
difficult to disentangle, there is good evidence of a specific deficit in
the execution of motor activities with left-hemisphere injuries. The
term apraxia was coined by Liepmann (1900), who distinguished
between ideational and ideomotor apraxia. The concepts are intimately connected with a hierachical concept of control centers in
execution of motor acts. The disconnection of language areas from
100
CHAPTER 6
motor control pathways can account for some cases of apraxia (Geschwind, 1967a), but in addition, it is necessary to infer higher order
mechanisms for the control of skilled acts in the left hemisphere
(Kimura, 1979). Aphasia and apraxia are freguently associated (Kertesz & Hooper, 1982; Poeck & Lehmkuhl, 1980).
There are two forms of explanation for the association of aphasia
and apraxia, and for the association of aphasia with a nonverbal deficit
in general. The first is the assumption of a common-core deficit underlying both verbal and nonverbal deficits. Kimura (1979), following
Liepmann (1900), argued for this conclusion in discussing the relation
between aphasia and motor disorders.
To the apparent Counterargument that not all apraxics are aphasics and vice versa, the reply would be that clinical tests are not
sufficiently sensitive to reveal mild deficits. A more telling argument
has been given by Poeck and Huber (1977), who said that the linguistic
aspects of aphasia are left unexplained by the hypothesis.
Another version of the common-core hypothesis sees aphasia
as a defect of symbolic activity, asymbolia (Finkelnburg, translated
by Duffy & Liles, 1979). Of particular interest is the defective use and
understanding of gesture and pantomimein aphasics (Duffy & Duffy,
1981; Varney, 1978). Cicone, Wapner, Foldi, Zurif, and Gardner (1979)
found a similarity in the qualitative aspects of gesturing and speaking.
This was not found by Lehmkuhl, Poeck, and Willmes (1983) when
they analyzed error types in tests of apraxia.
Opponents of the common-core hypothesis point out that it
makes the strong assertion that phenomena are invariably associated,
and that if exceptions exist, then the hypothesis can be rejected. They
further point out that, if the phenomena are both present but are not
correlated (Goodglass & Kaplan, 1963) or show different recovery
rates (Poeck & Lehmkuhl, 1980), then they are not manifestations of
a single function. Both Goodglass and Kaplan (1963) and Poeck and
Lehmkuhl (1980) advocated the so-called anatomical hypothesis, saying that the proximity of the neural substrate makes for clinical association because of the typically large lesions occurring in common
forms of pathology. The second hypothesis is the more conservative
and should be adopted. Even so, it is legitimate to speculate that
the proximity of brain representation may give a clue to functional
relatedness or evolutionary association. Although assertions about
101
evolution are necessarily speculative, it seems likely that verbal (auditory-vocal) language evolved in a context in which visual-gestural
communicative and symbolic abilities already existed.

A number of tests representative of traditional operationalizations of constructional deficits, apraxia, and nonverbal reasoning were
performed. It was hypothesized that this three-dimensional structure
is reflected in an aphasic population. The hypothesiswas tested with
factor analysis and an ensuing analysis of variance relating the test
results to type, severity, and chronicity of aphasia. Other researchers
have reached diverging conclusions on the relation of nonverbal deficits to aphasia classification. Some have found no relation to severity
and some a strong association (see above). If the type of aphasia is
considered, the authors finding a relationship to type generally ascribe
the most severe deficits to patients with posterior lesions and fluent
speech, but the results have not been consistent. The attempt to avoid
confounding the type and the severity of the aphasia, which is an
important feature of the present system, may clarify the relationships.
6.2.1. Tests of Nonverbal Abilities
6.2.1.1. Raven Coloured Progressive Matrices (Raven, 1960). This
test (RCPM) consists of a test booklet with 36 tasks. Each task contains
a pattern with an empty slot and six alternative patterns that all fit
into the slot, but with different designs. For each task, only one of
the alternatives is a correct response, and the subject is asked to
choose. The tasks are divided into three series of increasing difficulty
(A, Ab, and B), starting with perceptual matehing problems and progressing toward more abstract reasoning problems. In some studies,
it has been useful to distinguish between the series as measures of
different underlying functions (Denes, Semenza, Stoppa, & Gradenigo, 1978; Costa, 1976).
The test is likely to show a greater deficit in patients with posterior than in patients with anterior lesions (Basso et al., 1973). It does
102
CHAPTER 6
not discriminate well between patients with right- and left-hemisphere lesions (Arrigoni & de Renzi, 1964). A moderate but significant
correlation with aphasia was found by Kertesz and McCabe (1975),
but not by de Renzi and Faglioni (1965). All in all, it is, however,
generally regarded as a test that may often be performed very well
even by patients with a severe aphasia (Zangwill, 1964). Among aphasic
subgroups, the ones with the most severe aphasia also show the
highest incidence of subnormal RCPM scores (Kertesz & McCabe,
1975).
6.2.1.2. Wechsler Adult Intelligence Scale, Performance Test. This
test (WAlS) is a commonly used psychometric test for adults, and it
has also found application in clinical neuropsychology (McFie, 1975).
The performance tests consist of five scales.
Digit Symbol (DS) involves filling empty boxes with written symbols according to a digit-symbol code. The test demands writing with
the left hand in aphasics with right-side hemiparesis and therefore
introduces undesirable possibilities of contaminating factors. In the
present study, this test was omitted and each patient was assigned
a score corresponding to the average of the other performance scales.
Picture Campletion (PC) consists of 20 pictures, each with a missing detail. The missing detail must be named or pointed out. According to McFie (1975), this is the performance test showing least
alterations after cerebral injury.
Block Design (BD) uses nine blocks with red, white, and halfwhite and half-red sides. Patterns must be constructed by joining the
blocks. Nine patterns of increasing difficulty must be reproduced.
According to McFie (1975), this test is maximally sensitive to parietooccipitallesions of the right hemisphere.
Picture Arrangement (PA) consists of cartoonlike picture series
where the subject must arrange the pictures in proper sequence. There
are 12 series. This test is maximally sensitive to frontotemporal injury
of the right hemisphere, again according to McFie (1975).
Object Assembly (OA) consists of five puzzles of naturalistic
designs. The results are often parallel to those for Block Design.
There are many studies showing a pattern of verbal tests' being
performed more poorly than performance tests in left-hemisphere
injuries, and the reverse pattern in right-hemisphere injuries. From
103
this pattern, it does not follow in general that patients with lefthemisphere injuries perform normally on performance scales. Orgass
et al. (1972) showed that, on the average, aphasics perform as poorly
on performance tests as patients with right-hemisphere injury, whereas
patients with left-hemisphere injury without aphasia perform better
than bothother groups. There are differences in pattern of performance on subtests, however, aphasics performing most poorly in relation to other groups on picture completion.
6.2.2. Tests of Motor Function
Finger Tapping (FT-L) is from the Halstead-Reitan neuropsychological battery (Reitan & Davison, 1974) and is a telegraph-type
key connected to a counter. The subject rests his or her hand on the
table and with the index finger depresses the key as many times as
he or she can in 10 seconds. The average of five trials is recorded.
Only the scores for the left (unimpaired) hand are used.
Grooved Peg-Board (GP-L) is a test from the Halstead-Reitan neuropsychological battery. lt consists of a board covered with a metal
plate containing a 5 x 5 matrix of holes. An adjoining cup contains
30 pegs. Each hole has a groove oriented in different directions, and
each peg has a tag. The peg must be appropriately oriented with
respect to the grooved holes in ordertobe inserted. The time taken
to fill all the holes is recorded. If the subject looses a peg, an error is
counted. In this study, only results with the hand ipsilateral to the
injured hemisphere were used in the analyses.
6.2.3. Apraxia
The set of disturbances known as apraxia is believed to encompass several subtypes. The tests employed vary accordingly. Tests of
constructional ability (constructional apraxia) are
Copy-a-cross (COPY) from the Halstead-Reitan battery, range 0

to 5 (Reitan & Davison, 1974).
Frostig Copying (FROS) from Frostig (1966), range 0 to 100.
104
CHAPTER 6
Tests of imitative movements (ideomotor apraxia) were designed

with reference to Luria (see Christensen, 1975) and Goodglass and
Kaplan 1963, 1972):
Imitative Finger Position (FING) with the left hand, range 0 to 20.
Hand Movements Imitation (MOV-I) with the left hand, range 0
to 34.
Tests of ideational apraxia consist of
Manipulating Real Objects (OBJ), range 0 to 16.

Picture Sequencing (PICT), in which photographs of daily activities
must be ordered correctly, range 0 to 17.
6.2.4. Results
A factor analysis with principal component analysis and varimax

rotation was performed to clarify the structure of performances. The
first factor accounts for 49% of the variance in the unrotated solution.
The second, but not the third, factor has an eigenvalue above 1.0.
Because a three-factor solution was looked for, the loadings of the
different tests on the three first factors after rotation are still shown
in Table 6.1. The factors account for 64% of the variance.
All the nonverbal intelligence tests come out as the first factor.
Apraxia can be clearly distinguished from these in the second factor,
and Factor 3 reflects simple speed and copying performances. Salutions for four and five factors were calculated to see which of the
three factors would remain intact. The first two factors are unaffected
by the more comprehensive solutions.
Table 6.1. Three-Factor Solution of Performance Structurc
Factor 1
Factor 2
Factor 3
RAV-A (.76)
RA V-Ab (.84)
RA V-B (.86)
PC (.53)
FING (.71)
MOV-1 (.77)
FT-L (.71)
OB] (.88)
PICT (.64)
FROS (.46)
COPY (.61)
BD (.68)
GP-L (.59)
PA (.59)
OA (.66)
105
Table 6.2. Relation of Visual Nonverbal Ability to Severity of

Aphasia
Function
Test related to severity
Nonverbal intelligence
RCPM (p < .001)

Wechsler PIQ (p < .001)
Apraxia
GP-L (p < .01)

MOV-I (p < .001)
FING (p < .001)
OBJ (p < .001)
PICT (p < .01)
Others
Copy (p < .01)
6.2 .4.1. Relation to Aphasia Classification. Because all the subscales of the Raven and Wechsler (PIQ) load highly on the first factor,
it is reasonable to analyze the sum scores only in relation to the aphasia
group. In the following tables, the results have been corrected for the
correlation of aphasia severity with age. The results for the functional
areas suggested by the factor analyses are summarized in Tables 6.2
and 6.3. For nonverbal intelligence tests, the relationship to severity
of aphasia is significant and corresponds to a correlation of .38 and
.27 for the Raven and Wechsler measures, respectively. A tendency
of the fluent aphasics to score somewhat better on the Raven test
than the nonfluent is not significant. There are no significant interactions between the classification variables.
For the tests of apraxia, the relationship to severity of aphasia
was consistently present, corresponding to correlations from .27 (PGL) to .51 (FING). For tests of imitating hand or finger movements,
Table 6.3. Relation of Visual Nonverbal Ability

to Type of Aphasia
Test with deficit
Function
Nonfluents
Apraxia
MOV-I (p < .01)
Fluents
106
CHAPTER 6
there was an association with type of aphasia, and the tendency was
for nonfluent patients to perform worse than fluent patients. These
are the tests most closely reflecting the concept of ideomotor apraxia.
The tests measuring ideational apraxia (OBJ and PICT) showed no
relation to type of aphasia. There were no interactions of the classification variables with respect to any of these tests.
Factor 3 was taken to reflect severe constructional difficulty coupled with slowness of repetitive finger movements.
For tests encompassed by this factor, the relationship to any
dimension of classifying aphasia was weak or absent.
6.2.5. Discussion
The three factors isolated are interpretable in relation to previous
discussions of nonverbal deficits as measuring nonverbal ability,
apraxia, and constructive deficits. The latter can be defined separately
from the ability factor when gross deficits on simple copying tasks
are used as a criterion, but within the ability factor, a split between
logical reasoning and more executive aspects of ability was not found,
even when four and five factor solutions were analyzed.
For the tests in the first factor, a correlation with severity of
aphasia of the same magnitude as previously reported by Kertesz and
McCabe (1975) was found. From the study of Basso et al. (1973), one
might have expected more severe deficits in fluent than in nonfluent
patients, but these were not found.
The tests of apraxia showed a clear correlation with severity of
aphasia, confirming the results of Kertesz and Hooper (1982) and
contradicting those of Goodglass and Kaplan (1963) and of Lehmkuhl
et al. (1983). Imitation tasks of moderate complexity (finger and hand
positions and intransitive movements) were especially difficult for
nonfluent aphasics. This type of task was not tested by Kertesz and
Hooper (1982). Lehmkuhlet al. (1983) found no differences between
aphasia types.
The disturbances of simple executive left-hand functions were
not related to the aphasic disturbance. They may have reflected functioning of the right hemisphere. It was impossible without routine
CT-scans, to exclude patients who may have had right hemisphere
lesions, and they may have accounted for the findings. So far, the
107
results are consistent with what has been termed the anatomical hypothesis, that language areas and areas underlying nonverbal functions
have proximallocalization. I shall, however, discuss the issue again
when the recovery data have been presented, considering the possibility that neither the common-core nor the anatomical hypothesis
is correct. What we are observing may be a multidimensional response
of the preserved brain to an injury in which a coupling of nonverbal
and verbal factors is apparent in severe aphasia. This coupling may
reflect the functional state of the preserved brain.
LOCALIZATION OF
LESION IN APHASIA
7.1. Status of the Localization Model

Our current aphasia classifications have developed in a context of
neurological diagnosis. The possibility of drawing firm conclusions
on the localization and type of pathology was a central concern in a
classification system. With the introduction of refined neuroradiological methods for localizing lesions, there has been renewed interest
in localization studies. The major concern of such studies today is to
answer whether accepted views on the neurological basis of language
functions are essentialy correct.
In this context three possible conclusions may be envisaged:
1. The Wernicke-Lichtheim model is correct and sufficient to
explain the phenomena of aphasia. The specific assumptions of the
model were stated in Chapter 2.
2. The Wernicke-Lichtheim model is incomplete. There are
additional areas that deserve to be called language areas in the sense
that lesions of these areas give rise to aphasia. The resulting types of
aphasia are different from the classical syndromes.
3. The Wernicke-Lichtheim model is correct in stating that the
likelihood of aphasia is high with lesions of the classical language
areas. 1t is wrong in ascribing the variations in aphasia types to lesions
of differently localized modules within the language areas. lt is the
pattern and volume of lesions combining areas within the classical
109
110
CHAPTER 7
language areas with neighboring areasthat determine symptom formation. The latter areas thus interact intimately with the language
areas but do not, in most cases, give rise to aphasia when lesioned
in isolation.
Briefly summarized, the first conclusion says that all is well with
classical aphasiology. Position 2 says that the conceptual schema of
the classical model is valid and can be extended to new areas and
new forms of aphasia. Position 3 says that localization must, to some
degree, be supplemented by a concept of interactive processing.
Early studies with CT-scan or isotope localization stressed the
essential correctness of the Wernicke-Lichtheim model (Naeser &
Hayward, 1978; Kertesz, Lesk, & McCabe, 1977; Kertesz, Harlock, &
Coates, 1979). During the last few years, reports on aphasia with
lesions outside the classical language areas have appeared with discussions of the possible mechanisms (see below).
7.2. New Candidates for Status as Language Areas

7.2.1. The "Limbic System"
The term limbic system is used to include the limbic lobe and the
associated subcortical nuclei. The limbic lobe includes the subcallosal,
cingulate, and parahippocampal gyri, as well as the underlying hippocampal formation and dentate gyrus. The main subcortical nuclei
associated with the limbic lobe are the septal nuclei, the amygdaloid
complex, the hypothalamus, the epithalamus, and various thalamic
nuclei.
Indications that the limbic system has a role in language functions were reviewed by Lamendella (1977). The cingulate gyrus may
be important in disturbances of language activation, with lesions
sometimes giving rise to mutism (Robinson, 1976). The hippocampal
area is considered of some importance in human recent memory. The
work of Corsi (1972) indicates that lesions of the left hippocampus
cause selective disturbances of verbal memory function. The possibility may therefore be considered that some of the variability in verbal
memory in aphasics is caused by varying involvement or disconnection of the hippocampal region. There are no indications that the
LOCALIZATION OF LESION IN APHASIA
111
subcortical nuclei and fiber tracts of the limbic system are important
to language.
7.2.2. The "Lenticular Zone"
The lenticular zonewas defined by P. Marie (1906) and comprises
a quadrangle of tissue extending from the anterior and posterior borders of the insular cortex to the midline of the brain.
Proceeding in the lateral-to-medial direction, the lenticular zone
contains the following structures:
7.2.2.1. Insula. The insularcortexlies in the depth of the Sylvian
fissure and is a conventionally defined neuroanatomical structure. It
is roughly triangularly shaped and its outer limits are defined by the
sulcus circularis. Wernicke (1874) attributed great importance to the
insula as an association area for fibers from the frontal and temporal
cortex. The view that lesions of the insula cause aphasia was, however, contradicted by Henschen (1922) on the basis of his review of
1,200 published autopsy reports. Penfield and Roberts (1959), as well
as Rasmussen and Milner (1975), confirmed that no interference with
language functions is found with electrical Stimulation of the insula.
The possibility remains that the insula interacts with other structures.
Mohr (1976) stressed the importance of insular involvement in conjunction with lesions of the Broca area in producing the symptom
complex of Broca aphasia.
7.2.2.2. Capsula Extrema. A fiber bundle believed to carry frontalinsular-temporal association fibers.
7.2.2.3. Claustrum. A sheet of gray matter. It has recently been
shown to have extensive reciprocal connections with sensory cortical
areas and particularly with visual areas.
7.2.2.4. Capsula Externa. A fiber bundle carrying mainly association fibers.
7.2.2.5. Basal Ganglia. The nucleus lentiformis is a prominent
structure located laterally to the internal capsule. Medially, the head
of the caudate nucleus can be seen bordering on the frontal horn of
the ventricles. Traditional interpretations have stressed the primary
motor functions of the basal ganglia, but Teuber (1976) urged a wider
interpretation of their function. A role in the language function was
considered unlikely by classical authors (e.g., Liepmann, 1915).
112
CHAPTER 7
In recent studies, Damasio, Damasio, Rizzo, Varney, and Gersh

(1982) and Wallesch, Kornhuber, Brunner, Kunz, Hollerbach, and
Sugar (1983), among others, have reported on aphasia with basal
ganglia lesions. Some of these cases have had Iasting deficits.
7.2.2.6. Interna! Capsule. A massive and compact layer of white
matter carrying all fiber projections-afferent and efferent-between
the cerebral cortex and the subcortical structures. These include fibers
linking cortical areas by a cortex-basal-ganglia-thalamus-cortex loop.
7.2.2.7. Thalamus. Medial to the internal capsule lies the thalamus, which borders on the third ventricle. The pulvinar is a !arge
thalamic nucleus located at the posterior end of the thalamus bardering on the posterior horn of the lateral ventricle. The pulvinar is
of great interest in view of its massive connection with the posterior
language area. The functional significance of these connections are,
however, uncertain. Electrophysiological studies have shown interference with language functions after stimulation of the left, but not
of the right, pulvinar (Fedio & van Buren, 1975). The stimulation
effects are not generally different from those seen after cortical stimulation. The parallel results of stimulation studies may be taken to
support a conception of the posterior language area and the pulvinar
as one functional unit. lt is possible that the stimulation effects are
caused by indirect interference with cortical functions. However, a
well-documented series of cases with subcorticallesions centering on
the basal ganglia or the thalamus was presented by Alexander and
LoVerme (1980). Penfield and Roberts (1959) interpreted the pulvinar
as an important relay station between the posterior and the anterior
language cortex. They assumed that the pathway is from the pulvinar
to the dorsomedial nucleus of the thalamus and from that to the Broca
area. Cases of thalamic aphasia have been published by, among others,
McFarling, Rothi, and Heilman (1982). Luria (1977) suggested that
the thalamus has important functions in connections with attentional
control of the corticallanguage mechanism.
7.2.3. Medial Structures
The precise localization of the supplementary motor cortex in humans
is not known. Electrophysiological work points to several discretely
located sensorimotor representations of the body surface in animals
113
(Woolsey, 1958). In humans, a supplementary motor area was

described by Penfield and Roberts (1959) on the basis of electrical
stimulation. It is located just anterior to the motor strip, occupying
part of the medial surface and extending onto the convexity of the
frontal lobe. The limitations of the area rest on electrophysiological
criteria, and anatomical extensions cannot be precisely defined.
The area is not considered in the classicalliterature on aphasia,
not even in Henschen's review (1922) of all published cases with
autopsies. The first hint of functional significance to language comes
from the work of Penfield and Roberts (1959), who found interference
with language after electrical stimulation in local anesthesia. Since
then, some cases of aphasia with infarction of the supply area of the
anterior cerebral artery have been published (Rubens, 1976). For review,
see Razy, Janotta, and Lehner (1979) and Alexander and Schmitt (1980).
Work with regional cerebral blood flow (Lassen, Ingvar, & Skinhoj,
1978) has also attributed a significant function in speech to the supplementary motor area. After lesions in this area, the characteristic
defects most frequently pointed out are preserved repetition with a
tendency toward echolalia, but paucity of spontaneaus speech.
7.2.3.1. Corpus Callosum. This isamassive bundle of commismal
fibers connecting the two hemispheres.
Lesions of the corpus callosum or of callosal fibers may produce
disconnection syndromes in the form of inability to perform verbal
commands with the left hand (a form of apraxia; Geschwind, 1967a)
or inability to read material in the left visual field (a form of alexia;
Bensan & Geschwind, 1969). Modality-specific naming defects may
also occur with callosal lesions. Hemispheric integration may be
important in memory functions and in certain motor skills.

Analysesofa subsample of the present group of patients have
been presented by Reinvang and Dugstad (1981) and by Reinvang
(1983). In those studies, the relationship of aphasia types to localization was described in terms of the conditional probabilities of the
results on one variable, given information on the other. Statistkai
114
CHAPTER 7
analyses of differences in aphasia test parameters as a function of

lesion localization were also reported.
Most analyses of lesion localization in aphasia have limited themselves to descriptive accounts in the form of case descriptions or
composite lesion diagrams (Kertesz, Harlock, & Coates, 1979). Apart
from the conditional probabilities used by Reinvang and Dugstad
(1981), Blunk, Oe Bleser, Willmes, and Zeumer (1981) have presented
a quantitative approach to analysis based on a subtractive method of
lesion comparison in major syndromes. The present study uses a more
refined quantitative method, that of canonical discriminant function
analysis, in order to relate classification and lesion, and considers the
merits of alternative classifications.
The further question asked is about the pathological basis of the
performance on different parameters of the aphasia test. Classification
systems rest on the assumption that the underlying parameters of
classification are related to lesion localization.
Finally, the difficult question of the relative independence of a
lesion effect is asked. Few of our patients had lesions restricted to
only one anatomical region. The question is whether lesions of a given
(lrea have independent effects on performance irrespective of the context of other lesions. The Wernicke-Lichtheim model asserts that this
is the case. An example is the explanation of global aphasia, which
is seen as a mixed syndrome. The deficits characteristic of Broca and
Wernicke aphasia (anterior and posterior lesions) are added tagether
to obtain the mixed syndrome.
The example also serves as a starting point for criticizing the
assumption of independent effects. Global aphasia has distinctive
featuresnot found in either Broca or Wernicke aphasia, notably stereotypy or recurring utterances. Fluency is generally poorer in a mixed
anterior-posterior lesion than in a pure anterior lesion, although a
posterior lesion in itself does not reduce fluency (Reinvang, 1983).
The analysis of interactions between lesions in the present study
used the approach of comparing correlations and performances in
groups with relatively discrete or composite lesions. A pragmatic criterion for division into lesion groups was used, and the lesions were,
of course, not truly discrete. A statistical method was therefore used
for partialing out the correlations between lesion sites within a defined
115
range of lesions. The problern selected for this analysiswas the question of the role of the insula and the basal ganglia in the context of
varying lesions of the Broca and Wernicke areas. This problern was
previously addressed by Brunner, Kornhuber, Seemuller, Suger, and
Wallesch (1982), who found that a basal ganglia lesion, in combination
with a cortical lesion, results in a more severe aphasia than a pure
basal ganglia lesion. The studies by Mohr (1976), already cited, concluded that the addition of an insular lesion to a Broca area lesion is
critical for producing a Broca aphasia. This conclusion was also confirmed in our own analysis (Reinvang, 1983).
7.3.1. Method
CT-scan was performed on clinical indications at various hospitals in Norway. For interpretation of the scans, a qualified neurologist or neuroradiologist (Drs. P. Barenstein and G. Dugstad) rated
the presence of lesions on a checklist prepared by the author (Table
7.1). On the basis of the checklist, a more limited set of lesion categories or lesion groups was derived (Table 7.2). If the scans were
suited for detailed interpretation, then diagrams of the lesions were
made on standardized slice diagrams. The systemwas developed in
the Department of Radiology, Neuroradiology, and Neurology at
Aachen, West Germany (Schmachtemberg, Hundgen, & Zeumer, 1983;
Blunk et al., 1981). Diagrams representing 16 brain slices with 5-mm
separation were prepared, based on corresponding anatomical cuts
shown in Matsui and Hirano (1978, pp. 143-157). A grid coordinate
systemwas superimposed on each slice. The grid system had 58 x 42
points and corresponded, according to the authors, to the degree of
resolution that can realistically be achieved in CT-scans.
In the fully implemented system, the lesions were processed
automatically after transfer to the grid model. In the present study,
the grid model was used only as a standardized mapping system and
as a means for estimating totallesion volume. The estimate was clone
by counting the grid squares encompassed by a lesion, added over
slices. The mappings were performed without knowledge of the
aphasia test results by a neurologist experienced in clinical aphasia
research (Dr. P. Borenstein).
116
CHAPTER 7
Table 7.1. Lesion Checklist and Grouping

Area checklist
Lesion category (L category)
Broca area
Broca
Insula
Capsula externa
Insula
Wernicke area
Wernicke
Supramarginal gyrus
Angular gyrus
Posterior
Frontomedial
Temporamedial
Medial
Cingulate gyrus
Hippocampus
Limbic
Basal ganglia
Pulvinar
Other thalamic
Subcortical nuclei
Arcuate fasciculus
Arcuate
7.3.1.1. Subjects. The population of 249 patients contained 125
individuals with CT-scans. Some of these patients were excluded

because the scans were not suitable for interpretation because of technical problems or because they had been performed too soon after
the onset of the illness to permit evaluation of lesion localization. We
found 89 patients with interpretable scans. Six different hospitals
made CT-scans available for this study.
Table 7.2. Derived Lesion Groups
Lesion in
Broca area
Group
Group
Group
Group
1
2
3
4
+
0
+
0
Wernicke or posterior
0
+
+
0
117
7.3.2. Results
7.3 .2 .1. Predictive Value of Classification Systems. The program for
canonical discriminant analysis of the SPSS (Nie et al., 1975) was used
to predict a fourfold classification (severe nonfluent, mild nonfluent,
severe fluent, and mild fluent) from the Iist of lesion variables (LlL8), with totallesion volume added as an extra variable.
The program first factor-analyzed the correlation matrix and then
used the derived factors in a linear equation for prediction. Two procedures may be used, one in which all of the terms in the variable
Iist are used, and one in which the program determines empirically
how many variables aretobe used. Both procedures were used, with
little variation in results. The results reported are those found with
the direct procedure (alllesion variables used).
In all, 66% of the patients were correctly classified (see Table
7.3). The program identified three groups of predictive lesion variables: The first function refers to Broca area lesions; the second to a
combination of arcuate fasciculus, medial, and insular lesions; and
the third is represented by Wernicke lesions and lesion volume.
The discriminative predictive factors pointed out by the procedure make sense in that the classicallanguage areas clearly came out
as discriminative variables.
In predicting aphasia type, the same statistical procedure was
used, but only classifiable cases were included. It was regarded as
appropriate to reduce the number of categories somewhat by combining transcortical motor, transcortical sensory, and isolation syndrome into one group. The classical model predicts that they will
share the property of having lesions outside the classical language
Table 7.3. Prediction of Aphasia Group

Predicted group
2
1.
2.
3.
4.
Severe nonfluent
Severe fluent
Mild nonfluent
Mild fluent
15
0
2
1
6
2
5
3
0
0
2
12
n
n
n
n
=
=
=
=
19
8
12
22
118
CHAPTER 7
areas. Jargon aphasia and Wernicke aphasia were also combined in

a common category, as the motivation for separating them was not
the expectation of differentiallesion locus.
In all, 36 classifiable cases were included in the analysis.
In all, 66% were correctly classified (see Table 7.4). The program
identified five functions used in prediction. The first function is represented by arcuate fasciculus, the second by insula and Broca area,
the third by Wernicke area and posterior areas, the fourth by lesion
volume and subcortical nuclei, and the fifth by mediallesions.
Again, the lesion variables singled out as discriminatory make
sense in terms of classical models of language pathology.
Although strict classification criteria were used in order to improve
on predictive validity at the cost of reduction in numbers, the percentage of correct classifications predicted by the lesion data is not
impressive.
The discriminant analysis technique was also used to predict
deviations in reading and writing from auditory-vocal functions. The
previous analysis of these conditions (Chapter 4) shows a variety of
combinations, of which several were not represented in the present
sample. Same combinations of categories, therefore, had tobe constructed. It was decided to run an analysis for predicting a threefold
classification: relative deficit of reading or writing (alexia or agraphia),
no deviation in reading or writing, and relative preservation of reading
or writing ("hyperlexia" or hypergraphia). We analyzed 15 cases with
relative deficit or preservation.
Of the cases, 53% are correctly classified (see Table 7.5). Because
Table 7.4. Prediction of Aphasia Type

Predicted type
1.
2.
3.
4.
5.
6.
Global
Broca
Wernicke
Transcortical
Conduction
Anomic
1
0
0
0
0
0
1
0
2
1
0
0
0
1
1
5
0
0
0
2
0
0
1
0
0
1
0
3
1
0
ll
ll
ll
ll
11
ll
=
=
=
=
=
=
7
10
6
4
7
2
119
Table 7.5. Deviation in Reading or Writing

Predicted group
1
1. Alexia or agraphia
2. No deviation
3. Hyperlexia or hypergraphia
3
6
2
3
0
23
16
n = 45
n = 11
of the relatively poor discrimination between groups, the content of

the predictive functions will not be further discussed.
7.3.2.2. Discussion. In the previous analysis by Reinvang and
Dugstad (1981) and Reinvang (1983), the conditional probabilities
showed a good correspondence between some aphasia types and
lesion groups. Global and Wernicke aphasia predicted lesion group
well, other types less well. The present attempt to improve correspondence by limiting the analysis to classifiable cases and using a
more high-powered statistical procedure to optimize predictive relationships did not succeed very well. Beyond confirming that the variables singled out in the classical clinicopathological models are of
discriminatory value, the results seem to confirm the essentially probabilistic nature of the relationship between pathology and function.
7.3.3. Analysis of Test Parameters
In examining the different parameters of the aphasia test in
relation to the locus of injury, two types of analyses are performed.
In the first (Table 7.6), a t testwas performed on the difference between
the groups having or not having a lesion of any designated area.
The result of this analysis is that, with respect to fluency, the
Broca area, the insular region, and the subcortical nuclei are significantly involved. In other parameters, all classicallanguage areas plus
the insula are involved. There are two modifications to this statement.
One is that the insula is not involved in auditory comprehension. The
other is that, with regard to repetition, the arcuate fasciculus has a
special function.
The second analysis is correlational, showing the extent of covariation between the degree of lesion in any given area and the aphasia
7
7
Insular
Wernicke
Posterior
Medial
Limbic
Subcortical
= Significant deficit in groups having the designated lesion (p < .05). Empty areas indicate no significant difference.
Words per minute

Repetition
Naming
Reading aloud
Writing
Aphasia coefficient
Broca
Table 7.6. Results ofT Tests Comparing Patients Having or Not Having a Given Lesion"
Arcuate
"
Al
tT1
:r:
>-
N
0
121
test parameters. The total lesion volume was added to the Iist of
anatomical categories. The pattern of results indicated by the t tests
are, on the whole, confirmed. It should be noted that, in addition to
correlation of lesion in any given area with different parameters, total
lesion volume has a significant correlation with all test parameters.
The arcuate fasciculus stands out more clearly in this analysis as
intimately connected to the classical language areas in functioning
(Table 7.7).
7.3.4. Lesions and Their Context
The above is a descriptive account of relationships in the present

pool of data. Our problern in interpreting these data is that lesion
sites are not independent of each other. The distribution of the blood
supply of the brain determines the likelihood of particular combinations of lesions in vascular cases. Ideally one would like to make a
parametric study of the effect of, for example, a Broca area lesion in
isolationandin all possible permutahans with lesions in related areas.
This ideal can be reached only in an experimentally controlled design,
andin clinical research, we are forced to use statistical means of trying
to correct for biases (in this case, correlations between lesion sites) in
the independent variables.
The first approach to this problern is to try to disentangle the
seemingly uniformly high contribution of all classicallanguage areas
by excluding from the analysis all patients with combined lesions of
Broca and Wernicke areas and all patients with no lesions of the
classicallanguage areas.
The t tests and correlation matrix for this remaining group are
shown in Tables 7.8 and 7.9.
It is not reasonable to analyze medial or limbic lesions because
these occur primarily in a context without lesions of classicallanguage
areas, and this lesion group was excluded from the analyses.
It is worth noticing that only a few significant correlations remain.
The influence of lesions in the Broca and posterior areas, as well as
the arcuate fasciculus, is no Ionger significant. The general lesion
volume affected only auditory comprehension and reading comprehension. Patients with lesions of the Wernicke area showed a higher
rate of speech output and lower comprehension than patients without
a Wernicke lesion. Reduced speechrate was associated with lesions
+.34
+.20
7.23
+.29
+.21
+.38
+ .28
+ .31
+.34
+.42
7.38
7.42
+.47
7.42
"Empty areas have insignificant correlations.
Words per minute

Repetition
Naming
Reading aloud
Writing
Aphasia coefficient
Insular
Broca
+.20
+.18
7.35
~.20
+.37
~ .32
+ .39
Wernicke
+ .36
+ .32
+.36
+.32
+.32
+.25
+.38
Posterior
Medial
7.20
Limbic
Table 7.7. Correlations of Lesion Size in Given Area and Performance (N
61)"
+ .33
Subcortical
Valurne
+.34
+ .49
+ .39
+.42
+.55
+.40
+.45
+.52
Arcuate
+.28
+.30
+.36
+ .32
+.29
+.31
+.32
+.37
'-l
"'
tTl
>-
::r::
(J
N
N
,_..
Insular
Wernicke
"Empty areas mean no significant difference between groups with and without lesion.
"Means that !arger lesion is associated with higher score.
'Means that !arger lesion is associated with lower score.
Words per minute

Auditory comprehcnsion
Repetition
Naming
Reading aloud
Writing
Aphasia coefficient
Broca
Posterior
Subcortical
Table 7.8. Results ofT Tests Comparing Patients with Restricted Lesions of a Given Areaa
Arcuate
.....
:;
C/l
;:t>
::r:
;:t>
'"0
6
z
C/l
trl
r-
.."
6
z
2:i
r-
;:t>
()
r-
124
CHAPTER 7
of the insular region and of the subcortical central nuclei (basal ganglia
or thalamus).
The hypothesis immediately suggests itself that the richer set of
correlations in the total group is a spurious effect of correlations
between the independent variables. This is certainly true, but not
sufficient to explain the findings.
Consider a task like naming. According to the analysis of restricted
lesions, naming performance is unpredictable from either lesion locus
or lesion volume. In the total group, both Broca, Wernicke, posterior,
and arcuate lesions, as well as lesion volume, predicted naming quite
well. If naming is related to a limited locus, then it should have shown
up in the analysis of cases with restricted lesions. If naming is related
only to lesion volume, then the volume effect should have shown up
even in the restricted-lesion cases, among whom many had large
lesions (mean lesion volume 1,469 units, range 376-5,573 against 1,474
units as the mean for the whole sample).
It may thus be that the relation between function and localization
is different in restricted and composite lesions, but to test this possibility, a more rigorous analysis must be performed.
The problern of analyzing the contribution of insular or basal
ganglia lesions in the context of lesions of the Broca area, the Wernicke
area, or both was addressed. Three groups were formed. The first
group (GA) had Broca area lesions or lesions outside the classical
language areas. The second group (GB) had Wernicke area lesions or
lesions outside the classicallanguage areas. The third group (GC) had
combined Broca and Wernicke area lesions or lesions outside the
classicallanguage areas.
The program "regression" from SPSS (Nie et al., 1975) was performed in the 3 groups separately. The previous tables indicated that
words per minute and auditory comprehension had a relationship to
some of the areas considered in this analysis. The multiple regression
took the correlation of a Broca area lesion with the variable in question.
It added the partial correlation of a Wernicke area lesion to the same
variable and proceeded with the insular region and the basal ganglia.
Figure 7.1 shows the cumulative curve of percentage variance
explained for fluency by each lesion in this procedure. The significance
of each partial correlation was tested with an F test.
LOCALIZATION OF LESJON IN APHASIA
GuruJative so
%varianc.e
explained
GA
'lo
.----
21J
10
./
roca
Jnsu.la.
Cumui/.ive so
%v8riance 'lo
20
Cpsulii
GB
expi!Jined
JO
125
----
88.Sl1L gtJflglia
----
* stgni.(Lcant ilurease
10
------+-------
l.'errtiLke
------+--
Jnsula.
Cumuiative 50
(o Yariance
explained 'Ia
_____-
10
{apsui11 BasaL gangli.8.
GC
30
20
signi fic8ni llllrea.se
roca
Wernicke
.-
Jnsu.La
(apsuLa
BasaL ganglift
Figure 7.1. Predicting fluency from lesion combinations.
The equivalent function for comprehension is shown as Figure

7.2. The figure demonstrates that an insular lesion partially determines fluency when there is a restricted Broca area lesion. It does not
have the same influence in the presence of a Wernicke area lesion or
a combined Broca and Wernicke lesion.
CHAPTER 7
126
[IJmuJiiive
%variance
ex{iiJ.ined..
GI!
:;1
:L--30
roca
~-~- -~ - ~--+----
CapsliiEl
Jnsul;;.
a.sal gangli
Cumulative
%variance so
expl.ined
GB
40
30
20
10
.---
'v/f'rnl.().e Jnsu.!a
CumulatiYe
%varince 5o
ex-plttined 'lo
30
20
CapsuLa
BasaL go:ng&l
-
-----------
GC
10
Broe
werni&e Jnsula.
Capsuia.
Basal ga.nglia
Figure 7.2. Predicting auditory comprehension from lesion combinations.
7.37
Insular
"Empty areas have nonsignificant correlations.
Words per minute

Auditory
comprehension
Repetition
Naming
Reading aloud
Writing
Aphasia coefficient
Broca
7.31
Wernicke
Posterior
7.35
Subcortical
Arcuate
Table 7.9. Correlations of Lesion Size in Given Area and Performance in Patients with Restricted Lesions (N = 28)"
7.45
7.36
Volume
l""'
.....
~s;:
>
gJ
l""'
'Tl
~
z0
()
128
CHAPTER 7
Auditory comprehension can be explained only in a combined

Broca and Wernicke lesion. The apparent specific effect of a Wernicke
area lesion in Table 7.9 seems to be a spurious effect, and the true
effect is a volume effect in the context of classicallanguage area lesions.
7.4. Conclusion
It is difficult to predict the type of aphasia from combinations
of lesion variables. The parameters singled out by classical clinicopathological models as distinctive also emerged in these statistical
analyses, but not with predictive relationships of highly significant
strength. A possible explanation of these probabilistic relationships
can be derived from the analyses of restricted and composite lesions.
At least in some cases, it can be shown that the effect of a given lesion
is different according to the context of additional lesions. This interactionwill obscure any pattern of results pursued with methods looking only for independent effects.
The results are thus in favor of the third position outlined initially
in this chapter. A concept of interaction- or context-dependent lesion
effects is needed to supplement a classicallocalizationist form of analysis. The results are in accord with reports by neurosurgeons that
limited cortical removals give mild deficits (Penfield & Roberts, 1959)
and that combined cortical and subcortical involvement predicts a
more severe deficit (Hecaen & Consoli, 1973).
RECOVERY AND
PROGNOSIS
8 .1. The Recovery Process

As argued by Sarno (1976), knowledge of the natural process of recovery is an important prerequisite for judging the efficacy of therapeutic
efforts. In Chapter 1, it was also suggested that such knowledge has
theoretical importance because, by comparing the immediate and Iangterm adjustment of the brain to a structural injury, the mechanism
of localization of function in the brain (systemic or nonsystemic) may
be assessed. The tendency to ignore change because it introduces
noise into a system for predicting the locus of the lesion from aphasia
type is unfortunate.
Just what the conditions are under which a natural process of
evolution can be assumed to take place is unclear. It has been pointed
out that any confounding influence of therapy should be eliminated
if spontaneaus recovery is tobe assessed (Sarno, 1976). Reinvang and
Engvik (1980a) and others have pointed out that there is an equally
large risk of confounding influence by negative factors of deprivation
and isolation. The more slowly developing recovery processes at work
beyond the first few weeks after stroke or trauma must operate in a
context of intrapsychic and social influences. Wehave suggested that
"neutral treatment" is an appropriate control condition, and we define
it as
129
130
CHAPTER 8
and he is encouraged to make maximum use of intact verbal capabilities
for communication. (Reinvang & Engvik, 1980a, p. 79)
Previous studies describing general recovery trends have suggested that, in untreated patients, spontaneaus recovery is seen only
for the first 2 to 3 months (Vignolo, 1964; Culton, 1969). This suggestion was Contradieted by Kertesz and McCabe (1977), who found
continued improvement, but at a decelerating rate, for the first year
in unrehabilitated aphasics.
In patients having received some form of treatment, a Ionger
duration and a greater extent of recovery have been found, perhaps
dependent on the duration of treatment and the time of starting treatment (Vignolo, 1964; Basso et al., 19~5, 1979).
The question of whether all aspects of aphasia improve to the
same degree is somewhat controversial. The most frequent opinion
is that auditory comprehension improves more than expressive performances (Vignolo, 1964; Basso et al., 1979; Lamas & Kertesz, 1978;
Prins, Snow, & Wagenaar, 1978). However, some authors have found
better recovery of repetition (Kenin & Swisher, 1972) or naming
(Kreindler & Fradis, 1968; Reinvang & Engvik, 1980a). The disagreements may be due to differential recovery rates for different functions,
or different degree of recovery in subgroups. The latter explanation
was suggested by Lamasand Kertesz (1978).
The question of "Syndromenwandel" (Leischner, 1972)-that is,
how often the aphasia type may change-has rarely been discussed.
Kertesz and McCabe (1977) found relative stability of aphasia types
over time. Reinvang and Engvik (1980a) found that aphasia types
were, on the whole, rather stable because changes with time were of
a general nature and did not alter the shape of the test profiles.
Changes of test profiles leading to redefinition of the aphasia type
did occur, however. This occurrence has also been noted by others
and has led to questioning whether all of the traditional aphasia syndromes should be regarded as independent entities. Mohr, Pessin,
Finkelstein, Finkelstein, Duncan, Davis, and Grand (1978) regarded
Broca aphasia as being the result of improvement in global aphasia.
Several other authors, among them Liepmann (1915), have recognized
this development, but it is not commonly accepted as the only context
in which Broca aphasia can occur.
Doubt has also been raised about the independent status of
RECOVERY AND PROGNOSIS
131
Doubt has also been raised about the independent status of

conduction aphasia. Same authors think that it may be related to
Wernicke aphasia, as a stage in the process of recovery (Kertesz &
Benson, 1970). On the other hand, Bensan et al. (1973) reported that
conduction aphasia occurs frequently in the first weeks after brain
injury, with subsequent recovery.
Anomic aphasia is regarded by several authors (Goodglass &
Kaplan, 1972; Kertesz & McCabe, 1977) as an end point of development for improving aphasics with various aphasia types, but it may
also appear as a primary syndrome.
The studies cited above lead one to expect that, by specifying
the time interval during which improvement is studied, the type of
task, and the type of aphasia, some of the ambiguities in the reported
findings can be resolved.
8.2. Recovery of Nonverbal Functions

The recovery of nonverbal functions has been little studied, but
some informationwas given by Kertesz (1979). In general, he found
a greater recovery of nonverbal functions than of language functions
in global aphasics, whereas in other types of aphasia, the nonverbal
recoverywas said to lag behind. The conclusions were based on visual
inspection of test-retest results, with no control for level of initial
performance.
Poeck (1983a) asserted that aphasia and ideational apraxia have
different recovery rates and, on that basis, argued for separate underlying mechanisms. Kertesz (1979), on the other hand, reported a close
parallel between recovery from apraxia and language recovery.
8.3. Prognosis
The prognosis for complete recovery from aphasia is poor in
patients in whom the symptoms persist beyond the first few weeks
(Culton, 1969; Brust et al., 1976). Prognosis in the present context is
therefore a question of predicting the amount of improvement in
individuals who will remain aphasic. Same background factors are
132
CHAPTER 8
believed to be of predictive significance. Several authors agree that

etiology (traumatic vs. vascular) is of importance and that patients
with traumatic etiology have the best prognosis (Butfield & Zangwill,
1946; Kertesz & McCabe, 1977). Age is believed tobe important, with
young patients making the best recovery. This trend, however, failed
to reach statistical significance in several studies (Kertesz & McCabe,
1977; Sarno & Levita, 1971; Smith, Champoux, Levi, London, &
Muraski, 1972).
There is no direct evidence that sex or education is prognostically
significant, although McGlone (1980) speculated that the lower incidence of women reported in many samples of aphasics may be due
to better prognosis as a result of the more diffuse cerebral representation of language in women.
Kertesz and Sheppard (1981) explained the reported differences
in sex distribution as being due to the differential risk of stroke in the
age groups from which candidates for study are recruited. They also
failed to show differential recovery rates in the two sexes. Kimura
(1983) found that differences between the sexes appeared with lesions
of the anterior language area. These were also found by Fredriksen
and LernCEs (1984) in our group.
Left-handedness has been stated to be associated with a better
prognosis than right-handedness (Subirana, 1969), but the evidence
is equivocal.
With respect to subgroups, Kertesz and McCabe (1977) found
that Broca aphasics showed the best improvement and globals the
poorest among all types of aphasics. These were untreated patients.
Sarno, Silverman, and Sands (1970) found no improvement in globals
even with treatment. However, Sarno and Levita (1979) reported some
improvement in globals in the latter part of the first poststroke year.
This finding was confirmed by Sarno and Levita (1981). They stressed
the fact that these patients were "alert" globals. Mohr et al. (1973)
found continued improvement long after injury in a case study of
global aphasics. Wernicke (1874) regarded sensory aphasia (Wernicke
aphasia) as having a good prognosis. Basso et al. (1975, 1979) found
the improvement in fluent aphasics tobe similar tothat in nonfluent
aphasics. Lomas and Kertesz (1978) regarded the initiallevel of auditory comprehension as a good predictor of improvement in other
language performances. Brust et al. (1976) found that fluency was a
133
good predictor of spontaneaus recovery, with fluent patients improving most, whereas Vignolo (1964) reported a poor prognosis for patients
with severe oral expressive (apraxic) problems.
8.4. Mechanisms of Recovery

Recovery has both a functional and a physiological aspect, and
a satisfactory theory accounts for both. A predominantly neuropsychological theory need not specify the nature of the physiological
mechanisms involved in recovery. A neuropsychological theory should
specify the areas of functioning involved in recovery, the influence
of premorbid factors, the environmental influences in the recovery
period and the form of organized interplay of functional processes
taking place. The findings of cellular mechanisms of recovery, both
anatomical (Raisman & Field, 1973; Schneider, 1973) and biochemical
(Glick & Zimmerberg, 1978), are valuable, but they do not constitute
an alternative to neuropsychological models, nor do they exclude any
specific form of neuropsychological model, with the exception of
models that explicitly state that all recovery is an epiphenomenon
based on compensatory mechanisms for covering up the deficit.
In the following, some theoretical alternatives for neuropsychological recovery models are outlined. They are ordered along a
dimension introduced in Chapter 1, the dimension of degree of systemic relationship between brain regions. The simplest theory states
that recovery is dependent on internal processes in preserved parts
of the language areas. The most complex theory sees recovery as one
aspect of readjustment of the whole brain to localized injury. Two inbetween positions are outlined.
8.4 .1. Relearning or Facilitation

Simple relearning by selective practice of language content material (words and sentences) was found by Wiegel-Crump and Koenigsknecht (1973) and by Wiegel-Crump (1976). Learning generalized to
language material of similar content or structure. Schuell (1974) has
been a strong advocate of using learning principles in therapy.
134
CHAPTER 8
Facilitation is invoked by authors favoring a psychosocial Stimulation treatment of aphasia (Wepman, 1953). The aphasic suffers
mental blocks and frustrations. If they are removed by appropriate
therapeutic attitudes, the way lies open for easier access to previously
learned language material.
Sometimes, aphasics experience a sudden breakthrough or rapid
improvement of speech as a result of strong emotional stimuli. Lifting
of inhibition has been invoked to explain such phenomena, which
mainly occur in the early weeks of the recovery process (Luria, 1970).
It is reasonable to view both simple relearning and facilitation
as a function of increased efficiency in the injured structures normally
responsible for the function. Authors favoring a learning approach
are typically not concerned with the neural basis of performance.
8.4.2. Reorganization of Function

Reorganization of function (Luria, 1966) is a complex notion
based on the idea that performances rest on functional systems. Injury
interferes with functional systems, but by reorganizing the remaining
components or adding new ones, a new foundation for adequate
performance is created. Dressed in terms of cognitive theory, one
may say that the patient learns or spontaneously invents new cognitive strategies. Some studies (Weinberg, Diller, Gordon, Gerstman,
Lieberman, Lakin, Hodges, & Ezrachi, 1977) with nonaphasic strake
patients have shown that it is possible to teach patients strategies for
directing attention by making an intact resource (verbalization) the
key element in the strategy. Attempts to teach memory strategies to
aphasics by encouraging visualization have been only moderately
successful. These approaches differ from relearning approaches mostly
in terms of focus. Relearning focuses on concrete content material
(word and sentences), whereas reorganization focuses on general
strategies. It is also explicitly stated that the neurological basis for the
reorganized function is partly different from the normal basis.
8.4.3. Release of Vicarious Neural Structures and Functional

Relocalization
It seems obvious that the recovery process must be based on
activity in preserved brain structures. Either the right hemisphere
takes over, or preserved areas of the left hemisphere are responsible
135
for recovery of function, or both. Smith et al. (1972), finding significant

correlations between improvement and a wide range of phenomena,
suggested that wide areas of the brain, even those traditionally termed
sensorimotor, take part in recovery. Milner, Branch, and Rasmussen
(1964) concluded, on the basis of studies of cortical ablations in epileptics with injury dating back to childhood, that dominance for the
language function does not shift unless there is a massive left-hemisphere injury. Rasmussen and Milner (1975), however, speculated
that dominance for the anterior or posterior speech center may s,hift
independently, indicating the possibility of a partial shift of dominance with less severe injuries. These hypotheses indicate that different recovery mechanisms are at work in severe and in less severe
aphasics (right and left hemisphere, respectively). When relocalization takes place, arelease from inhibition permits dormant functional
competence tobe utilized.
The hypothesis that right-hemisphere mechanisms are involved
in recovery has been tested on aphasics with the dichotic listening
technique (Johnson, Sommers, & Weidner, 1977; Pettit & Noll, 1979;
Castro-Caldas & Botelho, 1980). All authors found some evidence for
increased right-hemisphere participation in recovery, but Castro-Caldas and Botelho (1980) found the evidence only in the case of fluent
aphasics. The differences on which these conclusions are based are
small, and the results vary with the nature of the test.
8.4.4. Complementary Redifferentiation of Function

The hypothesis of complementary specialization has been applied
to the cerebral hemispheres and implies that the specialization of one
hemisphere is linked to the specialization of the other. This hypothesis
was reviewed in Chapter 1. lt may be applied to recovery by viewing
recovery as partly a product of redifferentiation of function in a preserved neurological substrate.
A paradigm that may be particularly useful in shedding light on
the sort of complex interactive process considered here is the paradigm of the single versus the seriallesion. A lesion of an anatomical
structure generally has more serious consequences when it occurs in
one stage than when it occurs in several stages with time intervals
between them. Even when both groups of animals recover, lesions
in additional structures have different effects in the two groups. The
CHAPTER 8
136
available evidence was reviewed by Finger (1978). Although several

explanatory principles must still be considered, it seems that simpler
mechanisms like neural shock effects cannot account for the data. A
model of functional reorganization taking into account the systemic
properlies of interconnected neural structures is needed (Finger, 1978).

The present study asks three questions. They are answered and
discussed in separate sections:
1. What is the recovery pattern for verbal and nonverbal performances? This question has both quantitative and qualitative aspects. The question of stability of clinical aphasia
syndromeswill be addressed.
2. Which are the important prognostic indicators? In addition
to the language and neuropsychological variables, backgmund vanables of age, sex, education, etiology, and lesion
volume are considered.
3. What is the structure of relations between functions in
recovered and unrecovered patients?
The answers to these three questions, taken together, are pertinent to deciding between the neuropsychological models of recovery
mechanisms. Although they are not precisely formulated, Table 8.1
is an attempt to summarize which findings to expect on the basis of
different hypothetical recovery mechanisms.
A full discussion of the results with respect to this table will not
be attempted here. How we view mechanisms of recovery depends
on how we view cerebrallocalization of function, and this discussion,
built on an integration of material from the different chapters, is
reserved for the final chapter.
8.5.1. Recovery Pattern
In all, 134 patients were retested. The likelihood of being retested

was compared in the different groups of the 2 x 2 x 2 "cube" model
(Chapter 3) and was not found tobe significantly biased.
No nonverbal deficit.
Global recovery of all
impaired functions at
decelerating rate.
Quantitative and
qualitative jumps.
No change in nonverbal
abilitics.
Nonverbal function may

suffer if it competes with
verbal function. Best
recovery of
comprehension and
single-word utterances.
Partially injured
Janguage areas
Whole brain
Right-hemisphere
homologue
Whole brain
Effectivization
Facilitation
Simple relearning
Reorganiza tion
New strategies
Vicarious function
Complementary
redifferentition
Impairment of nonverbal
function. Verbaland
nonverbal recovery are
coordinated.
Recovery pattern
Neurological substrate
Behavior mechanism
Verbaland nonverbal
function become more
clearly differentiated in
recovery.
Relations of verbal and

nonverbal performances
unaffected for simpler
tasks but may compete in
more resource-demanding
tasks.
Change toward closer

integration of verbal and
nonverbal function in
recovered patient.
Performance structure
unaltered by time or
recovery.
Functional structure
Table 8 .1. Hypothetical Recovery Mechanisms and Findings Consistent with Them
lesion volume.
Positive effect of Jesion
momentum.
Negative influence of
Better recovery with !arge

than with medium-sized
lesions.
Appropriate training.
Preserved
neuropsychological
functions.
Negative effect of lesion
volume.
Age.
Initial state of language

function.
Time after injury.
Lesions volume.
Age.
Prognostic signs
:;<:)
1'!1
......
VJ
'.J
ti'l
0(fJ
Cl
:;<:)
"V
CJ
>z
-<
:;<:)
n
0
<
1'!1
138
CHAPTER8
Table 8.2. Stability of Classification

Nonfluent
Test 2
5evere
Fluent
Mild
5evere
Mild
10
8
37
Test 1
F-5
F-M
0
1
NF-5
NF-M
36
0
12
17
2
0
3
7
37
30
12
55
Total
The likelihood of retaining the same classification at retest is

shown both for the 4-group classification and the 11-group classification (Tables 8.2 and 8.3). Stability is present in both systems in the
sense that 64% to 75% of the patients retained the same classification
at retest.
When one studies recovery as a function of time, the possible
confounding influence of other variables must be considered. The
Table 8.3. Stability of Aphasia Types

Test 1/Test 2
1. Global
2. Isolation
3. Transcortical
motor
4. Broca
5. Jargon
6. Wernicke
7. Transcortical
sensory
8. Anomic
9. Conduction
10. Mixed nonfluent
11. Others
13
Total
10
11
3
1
1
1
0
10
3
2
3
1
3
2
2
1
11
13
22
10
2
2
28
4
3
4
17
17
41
31
137
139
hypothesis that recovery is a function of the initial state was discussed

with respect to a subset of the present study group by Reinvang (1983).
There were high test-retest correlations for the subscores of the
aphasia test, but if the difference score (Test 2-Test 1) is taken as a
dependent measure, then the correlation with initial score (Test 1)
was not significant.
Another surprising finding (Reinvang, 1983) is that the testretest interval was not correlated with the recovery score. The computations were rechecked on the present sample, with the same results.
Based on these findings, it is justified to show recovery as a
function of time between the onset of aphasia and the initial test.
Time was divided into 1-month intervals. Thus, the numbers in some
groups were low, and the appearance of the curve in some sectors
may be an unreliable indication of trends. One-way analyses of variance were performed across time groups, and the presence of linear
or nonlinear trends was tested for.
As can be seen (Table 8.4), the trends were similar for different
parts of the aphasia test, and all tests showed a significant time-related
trend. The recovery of the aphasia coefficient is shown (Figure 8.1)
to illustrate the findings.
8.5.1.1. Recovery of Nonverbal Functions. Memory functions were
discussed in Chapter 5, and the factorial structure of the tests was
discussed. A multifactorial structure was found. Although several of
these factors were related to type or severity of aphasia, they did not
show the time-related trends shown by the aphasia test variables. A
Table 8.4. Recovery Trend with Time for Subtest of the NGA
Between-group
Words per minute

Repetition
Naming
Reading aloud
Writing
Aphasia coefficient
Linear trend
Nonlinear trend
2.44
1.92
1.96
3.09
2.28
2.79
2.66
4.19
.05
.05
.05
.01
.05
.01
.01
.0001
1.25
4.99
6.23
8.44
7.56
15.09
9.45
15.09
n.s.
.05
.05
.01
.01
.01
.01
.001
2.58
1.61
1.53
2.55
1.75
1.56
1.98
3.10
.05
n.s.
n.s .
0.1
n.s.
n.s.
.04
.01
140
CHAPTER 8
Rt.covery score
100
90
80
10
50
4o
~r------"'---~~------------~m-~~n__re~~~~ry~~~re~
2.0
10
2
&
fO
II
12
15
months since onset
Figure 8.1. Recovery as a function of time since onset.
weak statistical trendwas shown by two tests (DSE-I and BSE-I), but
they gave too few clues to offer an interpretation.
The results support the position that the varieties of memory
defects are not secondary consequences of the aphasic syndromes,
but independent deficits.
Nonverbal abilities were discussed in Chapter 6, and the factor
analysis showed two types of functions called visual nonverbal abilities
and apraxia. The results on apraxia are clear-cut. The recovery did not
parallel that of aphasia, and the results support the position of Poeck
(1983a) rather than that of Kertesz (1979).
The factor analysis yielding a visual nonverbal factor could not
distinguish functions measured by the Raven CPM from those measured by the Wechsler PIQ. The two tests behaved differently in recovery, and it was the Wechsler PIQ that seemed to show the same timedependent phenomena as the aphasia coefficient. Other tests representing measures of apraxia, constructional deficit, or left-hand functioning showed no time-related recovery trend.
141
8.5.2. Prognosis
The aphasia coefficient difference score was used as measure of
recovery. Prognosis was tested first with respect to the cube-classification system, with age added as a fourth dimension. A dichotomy
was obtained by setting the cutoff point at the medium age of 53
years. The results (Table 8.5) show several independent effects.
Whereas higher order interactions could not be tested, there
were four significant two-way interactions (Figure 8.2).
The interaction of fluency and chronicity was that the fluent and
chronic group recovered less. The fluency and age interaction was
that young nonfluents recovered less than young fluents, but old
nonfluents recovered more than old fluents. This was a true crossover
interaction. The severity and chronicity interaction was that severe
patients recovered better than mild only in the acute phase. Recovery
was less dependent on chronicity in mild patients. The severity and
age interaction was that young and severe patients made a better
recovery than any other group defined with these two variables. The
finding of Sarno and Levita (1979) of better late recovery in nonfluent
than in fluent patients was confirmed.
8.5.2.1. Background Variables. Age was analyzed in conjunction
with aphasia variables because of the statistical association of age and
severity, which can thus be corrected for. The results for other backgmund variables (sex, education, and diagnosis) are shown (Table
8.6).
The significant effect of education was that the group classified
as students improved more than other groups. This was obviously a
young group, and it may be assumed that the age factor accounted
for the apparent effect of education.
Table 8.5. Recovery as Function of Subgroup
Type (nonfluent vs. fluent)

Severity (Iow aphasia coefficient
vs. high)
Chronicity (acute vs. chronic)
Age (young vs. old)
:x1
:x2
29.4
35.8
25.9
19.7
1.43
14.27
n.s.
.001
30.7
32.3
19.1
20.4
5.71
9.31
.02
.01
CHAPTER 8
142
fe:ovr:ty
score
50
40
30
20
Recwery .score
7jpe x Cli.ronicify
5o
(p(05)
NF
F~ffF
30
20
.F
10
!--
Acllie
qo
30
20
/0
ChroniL
Acu!e
Recovff'j
~NF
'F
Severiiy x llge (p<.02)
30
.zo
H'~
YoUilq
0/d
fkom.c
10
-+----
YOIJJUj
'lo
NF
---+-
score
50
Type x !lge (p <.0/)

F.
10
,f'f
f1'
-~-t---------
ftrovr:ry score
50
s~Severd:y x Chromaty (p<.OI)
~0
.s
Ottf.
Figure 8.2. Two-way interaction effects.
Diagnosis was analyzed as thromboembolic, hemorrhagic

(including subarachnoidal bleeds), head injuries, and others. Pairwise
comparisons showed that the head-injured group improved more
than the 2 cerebrovascular groups. To separate out the possible contribution of age to this finding was relevant but was not attempted.
The problern was that there was little overlap in age in the cerebrovascular and the traumatic groups.
Table 8.6. Relationship of Aphasia

Coefficient-Recovery Score to Sex,
Education, and Diagnosis
Variable
df
Sex
Education
Diagnosis
1
3
3
.53
3.18
5.38
n.s.
.05
.01
143
8.5.2.2. Neuropsychological Variables. The set of neuropsychological variables analyzed in Chapters 5 and 6, measured at the first
test, were correlated with the aphasia recovery score. No significant
correlations were found in the group as a whole.
As recovery is sensitive to several dimensions of an aphasia
classification, the analyses were repeated on subgroups suggested by
variables showing significant main effects on recovery (Table 8.5). A
summary of the results of these analyses is given (Table 8.7). The
fluent versus nonfluent distinction is omitted because this distinction
showed no main effect on recovery. The groups defined by the significant interaction terms were not analyzed because the number of
subjects in each group was small.
Correlations derived by such an explorative procedure should
be interpreted with reserve. If we bear in mind the factor analyses of
neuropsychological tests, it is reasonable to require that several tests
representative of a factor show significant correlations before this
factor is taken to have a significant prognostic value.
Using the factor analyses as guidelines for interpretation, one
Table 8.7. Results of Correlating Neuropsychological Variables with
Improvement in Aphasia Coefficient in Subgroups
Subgroup
Significant predictors of improvement
Severe
Pegboard, left hand (- .26)

Digitspan (.31); pointing span (.20)
Apraxia testsFING (.41); MOV-I (.40); OBJ (.29)
Block span (.20); block seriallearning-sequence (- .20)
Mild
Raven CPM (.24)

Frostig (.50)
Apraxia tests MOV-I (- .44) and OBJ (- .28)
Block serial learning-trials (- .32)
Block serial learning-intrusion (- .20)
Acute
Digitspan (- .27)
Digit serial learning-perseveration (- .22)
Apraxia test MOV-1 (.23)
Block seriallearning-sequence (- .19)
Chronic
Finger tapping, left hand (- .32)

Apraxia tests COPY (- .39)
144
CHAPTER 8
Using the factor analyses as guidelines for interpretation, one

may conclude that apraxia was a prognostic sign in both severe and
mild aphasics. The prognostic relation had the opposite sign in the 2
groups, and therefore the effects canceled each other out in an overall
analysis.
Verbal immediate memory (represented by two tests) was also
a prognostic sign in severe aphasia. Combinations of nonverballearning factors that could be recognized from the factor analyses showed
up as significant in both severe and mild aphasics.
The clearest prognostic implications of neuropsychological functions appeared in relation to the severe-mild distinction. The findings
are not so readily interpretable with respect to the acute versus the
chronic distinction. Low but significant correlations on a scattered
selection of tests are difficult to assess. In the following section, the
underlying functions in acute and chronic patients are analyzed with
another statistical approach.
8.5.3. Relations between Functions in Acute and Chronic Patients

The theoretical models outlined (Table 8.1) can be viewed as
making different predictions about the structure of recovered language functions. The reorganization theory says that language becomes
more closely interwoven with nonlanguage abilities because of compensatory strategies relying on intact abilities. The differentiation theory says that language and nonverbal abilities become less
interdependent in recovery because of an underlying process of ongoing differentiation.
I have used the method of factor analysis (principal factor with
varimax rotation) to assess the number of significant factors present
in acute and chronic groups and the specificity of these factors. These
notions are hard to give a satisfactory operationalization. The SPSS
program for principal component analysis Iimits the number of factors
analyzed by requiring eigenvalues at or above 1.0. The percentage of
variance explained by each factor is also calculated. As a rule of decision, I will say that Functional Domain 1 was more highly differentiated than Functional Domain 2 if the factor analysis isolated a higher
number of factors and if the proportion of variance explained by the
first factor was lower for Domain 1 than for Domain 2.
145
Bear in rnind that the recovery patterns rnay have been specific
to subgroups. The factor analyses were perforrned on the subgroups
suggested by the rnain effects in Table 8.5. According to this analysis,
severity and chronicity were significant variables in addition to age.
The variables entered into the analyses were the neuropsychological tests described in Chapters 5 and 6 and the subtests of the
aphasia test (words per rninute was taken as rneasure of fluency).
According to the criterion suggested, it was the severe aphasic
group that showed the clearest trend in the direction of greater differentiation of function with time. Six factors are pointed out in the
acute group and seven in the chronic. The percentage of variance
explained by the first factor shows a 5% decrease frorn acute to chronic
patients (see Table 8.8).
8.5.4. Conclusions
The findings of the last two sections suggest strongly that the
general recovery pattern described in the first section was not representative of every subgroup. Likewise, the general finding that neuropsychological variables show no close relationship to aphasia rnust
Table 8.8. Factars Revealed in Principal Component Analysis of Aphasia Test and
Neuropsychological Variables
Group
Number of factors
Percentage of variance
explained by first factor
All acute
45.5
All chronic
45.1
Acute: Severe
Mild
Young
Old
6
6
5
6
36.5
30.5
45.8
45.2
Chronic: Severe
Mild
Young
Old
7
7
31.4
29.8
41.2
44.0
5
5
146
CHAPTER 8
be qualified.
The finding of distinctive patterns over time and across functions
may indicate different recovery mechanisms in different subgroups.
Because the differences between groups were not dramatic, it is more
reasonable to conclude that several recovery mechanisms may have
operated simultaneously, but with different weights in different
groups.
The group yielding the clearest signs of a comprehensive adjustment of the whole brain to theinjurywas the severe aphasics. This
was a group that made a greater-than-average recovery, especially if,
in addition to being severe, they werein the acute stage and young.
Several neuropsychological variables representing factors of apraxia,
verbal memory, and nonverbal memory were predictive of recovery.
Finally, the severe aphasics showed signs that a process of functional
differentiation was taking place between the acute and the chronic
stages of the illness.
THE ORGANIZED
RESPONSE OF THE
BRAIN TO INJURY
The assertion that the organization of the brain is complex is trivial,

as is the statement that models reflecting this complexity must have
more sophisticated options for explaining integrative and selective
action than those of postulating the addition or the subtraction of
independently localized functional components.
The clinical models of localizationist thinking seem to attribute
to the brain less complexity of organization than general physiology
would attribute to the normal human and animal brain. Although we
suspect that the clinical models are too simple, we must know exactly
in what respect they fail in order to introduce more complex models
constrained by facts. This does not mean that all simple explanatory
models must be abandoned. lt may be that a model is applicable in
a narrow domain although it is not applicable in a wider domain. It
would constitute a scientific advance if a more generally applicable
model could be proposed, in which the predictions of the first model
are retained as special cases rather than general explanatory principles. The work of Wood (1978) is a good example. In his theory, the
results of a Iocalizationist model are incorporated as special cases in
a more general associate-network model.
147
148
CHAPTER 9
9.1. Evidence for Organized Complexity

A simple prerequisite for finding the evidence of complexity that
motivates revisions of models is to look for it. The present study has
looked for evidence of complexity by quantifying variables and using
appropriate methods of experimental control and statistical analysis.
Several examples of theoretically significant results can be given, of
which three are the following:
Example 1. The procedure of using analogaus tasks with verbal
and nonverbal content was used in the analysis of memory. This
experimental control, tagether with multivariate statistical methods,
showed a complex pattern of organization that did not easily fit the
verbal-nonverbal dichotomy. This dichotomy was not rejected. lt was
found to be a dominant but not exclusive mode of organizing the
underlying functions.
Example 2. The degree of quantification of both lesion and aphasia
variables permitted the search for an empirically derived regression
function for predicting function from combinations of lesion variables.
The procedure confirmed the discriminative value of lesions in the
classicallanguage areas, but not with impressive predictive success.
Rather than stopping at this stage and reporting the results as a partial
confirmation of the localizationist model, the further analysis of
subgroups revealed the principle that the effects of lesions are contextsensitive (i.e., variable as a function of the presence of other lesions).
There may even be areas (e.g., the basal ganglia) that arenot language
areas in the classical sense that an isolated lesion gives rise to aphasia.
In the context of a lesion in classicallanguage areas, these other areas
may still contribute to symptom formation.
Example 3. The recovery curve shows a time-dependent decelerating shape and is relatively independent of associated neuropsychological findings. When multivariate methods are used to isolate
subgroups with possible distinctive recovery patterns, the general
view gives way to several contrasting pattems. The distinction between
severe and mild aphasia seems essential to the study of the interdependence between aphasic recovery and associated neuropsychological function. In severe aphasia, this coupling seems most clear.
I hold that these three examples are sufficient to show organized
complexity in clinical phenomena and their relation to injury with
BRAIN'S ORGANIZED RESPONSE TO INJURY
149
evidence of multiple interaction between factors. Thus, sufficient

motivation, aside from plausibility outside the clinical context, is present for considering a systems theory type of explanation of the findings. In order to do so, I will offer a more general sketch of a model
before applying it to explanations of clinical phenomena.
9.2. A Proposed Systemic Model

9.2.1. Abstract Model
Consider a set of tasks (1 ... 6). Hypothetical Devices A and B
are necessary and sufficient to perform the tasks, and when working
conjointly as the system AB, they account for all the variance observed
in the tasks.
Assurne that we are able to test the functioning of A independently of Band vice versa. We find that A performs certain tasks better
than B and performs some worse. Express this finding in terms of
the percentage of total variance in a task accounted for by A and by
B. The resulting curve shows the response characteristics of A and B
and might Iook as in Figure 9.1.
%variflnce
expla.in.ed
%variatl.ce
explaLned
100
\\....-2
50
100
_/
50
10
10
2.
"
Figure 9 .1. Performance of two hypothetical devices.
150
CHAPTER 9
The devices have some degree of complementary specialization,

in that Ais relatively more proficient at solving 1, 2, and 3, and B is
more proficient at solving 4, 5, and 6.
The variances of the two components R2 (A) and R2 (B) are related
to the variance of the system R2 (AB) by the formula for multiple
correlation (Snedecor & Cochran, 1967, p. 402):
R2 (AB) = R~ + R~ - 2R\ X R 8
1 -
r AB
TAB
In this equation, the term r AB stands for the correlation of A with

B. If the two devices are working independently, the correlation is
zero and the variance of the output is the sum of individual variances,
in this case about 85%.
If we want to improve on this performance, we might be able
to improve on A by redesigning the components involved in solving
Tasks 4, 5, and 6 so that they are specially designed to solve only
Tasks 1, 2, and 3. That would be a good solution if Device A or its
designer "knew" that Device B would always be at hand to take care
of tasks 4, 5, and 6.
The above equation indicates another possibility. The correlation
r AB, if different from zero, will change the variance of the output
without requiring a change in the individual components of A and
B. lt was postulated that the system (AB) could account for 100% of
the variance in the tasks. lt can easily be seen that a negative correlation would increase the total variance explained because it increases
the numerator of the ratio. This means that activity in A reduces
activity in B and vice versa. This relationship makes some sense if
Devices A and B are coupled in such a way that they inform each
other of what tasks they are best able to solve. As long as Device A
is kept informed that another device with superior efficiency is operative, it does not do anything when Tasks 4, 5, and 6 are presented.
By being freed of this responsibility, it is able to improve on its performance of Tasks 1, 2, and 3. If information comes that Device Bis
not working, then Device A reverts to the mode of operation shown
in Figure 9.1. This mode of cooperation would be a truly systemic
interaction, in which specialization is partly determined by the builtin structure of the device, but in addition, a dynamic component is
151
added by way of a negative feedback coupling. One might say that

the dynamic component serves to sharpen and enhance the performance profile of the device. Note that a positive correlation, in which
activity in A stimulates activity in B, has the effect of making the
output of (AB) less than the sum of the individual components.
9.2.2. Neural Model
It is suggested that, in the organization of higher nervaus activity, principles of both structural specialization and dynamic sharpening of specialization are employed. Anatomically determined
differences in response characteristics are referred to as specialization,
and differences determined by dynamic interplay as differentiation.
Anatomkai specialization means that cells with a special structure
perform special functions. This mode of organization is present in the
sensory cortex; the sensorimotor cortex, in which cells with closely
similar response characteristics are localized in columns (Mountcastle,
1957) in the cortico-cortical connectivity, consists typicallyof short connections to neighboring cells (Jones & Powell, 1970).
Dynamically based differentiation of the response characteristic
of a cell or a cell assembly is a common phenomenon at lower Ievels
of the nervaus system. In perceptual systems, lateral inhibition is a
well-documented mechanism accounting for contrast enhancement
and perceptual illusions like the Mach-band effect (Ratliffe, 1961).
Note that, at higher Ievels of the nervaus system, differential
functions may evolve in areas with similar cell structure. The most
striking case is that of hemispheric specialization, in which homologaus areas (i.e., areas with essentially the same structural characteristics) have different functions. In general, the so-called "association
cortex" is present in several cerebrallobes, the cell groups are anatomically similar within this cortex, and differently located association
cortices are connected by long fiber tracts. The functioning of these
lang fiber Connections is crucial to the hypothesis that the presence
of lang connecting fibers allows distant cortical areas to develop complementary differentiation. Presumably, some structural asymmetry, in
the form of either mild difference of cellular specialization or simple
size difference between the connected areas, is necessary toset off a
process of further differentiation.
152
CHAPTER 9
The classical interpretationisthat the long fiber tracts allow the

formation of complex concepts based on association. In commenting
on the functioning of callosal fibers, Pribram (1971) proposed a different interpretation:
What kind of connectivity is it that rends asunder functions it supposedly
associates? This question has not been asked until now. My own answer
isthat perhaps the connections, rather than functioning to associate, tend
to separate through suppression the various parts of cerebral mantle.
(p. 362)
The neural model is a further extension of this proposal. First,

the evidence for the neural model will be examined, in relation to
hemispheric specialization and differentiation, and then an application of the same model to anterior-posterior specialization and differentiation will be suggested. This is, to my knowledge, a novel
application.
9.2.3. Clinical Evidence an Hemispheric Relationships

Evidence relating the functioning of the corpus callosum to the
development of hemispheric lateralization was cited in Chapter 1
(Seines, 1974; Denenberg, 1981). Some clinical evidence is relevant to
further evaluation of the model.
A large brain injury changes the structure of the remaining tissue, as well as damaging fiber connections. Thus, the basis for both
structural specialization and dynamic differentiation is altered. With
large hemispheric lesions at an early age, the evidence suggests that
lesions of the left hemisphere are followed by some deficits in both
language and nonlanguage performances, but by no means by severe
aphasias. This finding indicates that the right hemisphere is involved
in restitution of the language functions in these cases, but that the
loss of structurally specialized tissue cannot be fully compensated for.
Injuries to the right hemisphere are followed by normallanguage
development and severe visuospatial deficits (Kohn & Dennis, 1974;
Woods & Teuber, 1973). The higher recovery of hemisphere-specific
functions in left- than in right-hemisphere lesions has led to notions
that language development takes priority in development (Woods &
Teuber, 1973). This notion is difficult to relate to neural mechanisms.
How does a neurological structure "decide" to give up performing
153
the functions that it does best? The present neural model offers an
explanation.
Rather than saying that the preserved areas take over new functions, I hypothesize that neural tissue becomes functionally dedifferentiated when information about specialized activity in connected
tissue is missing. It is interesting in this connection to note that the
hypothesis can explain the asymmetry between hemispheres in the
ability to compensate for early injury. Assurne that the hypothetical
Devices A and B correspond to the left and right cerebral hemispheres,
and that Tasks 1 to 3 areverbaland 4 to 6 are nonverbal. As it stands
in Figure 9.1, the model fails to predict the results because the hemispheres are depicted as having the same degree of structurally based
specialization. The classical students of higher cortical function
regarded the left hemisphere as "leading," or as specialized in relation
to the right (Head, 1915), whereas modern students have tended to
stress complementarity in the degree of specialization between hemispheres (Milner, 1974). Gazzaniga and Ledoux (1978) took up the
thought that the only basic (structural?) specialization in higher cortical function isthat of the left hemisphere for language, and that the
apparent specialization of the right is a secondary effect of this
(dynamically based differentiation, according to our terminology).
Assurne that the relation between A and Binterms of structural
specialization is as in Figure 9.2. The fact that A performs Tasks 1, 2,
and 3 better and 4, 5, and 6 worse than B should lead to a complementary, but dynamically based, differentiation of response in B, so
%variance
%vatW!ce
expflJJned
expJai.ned
100
100
fl
50
50
10
10
;2.
.q
'I
Figure 9.2. Performance of two devices, one of which is specialized.
154
CHAPTER 9
that in a nervaus system with matured connections between A and

B, they would, in conjunction, operate as in Figure 9.1. If A received
a massive injury, however, the basis for upholding the differentiation
of B would be removed. B would revert to the mode of response
determined in its inherent structure, that is, no advantage in responding to any specific type of stimuli.
A massive injury of A would give an even performance profile
with somewhat subnormal performance on all tasks. A massive lesion
of B would give an uneven profile with relative preservation of language performances. If A is taken as the left hemisphere and B as
the right hemisphere, then this is exactly what is found in studies of
early lesions to the hemispheres.
The adult lesion data concern mainly evidence for right-hemisphere function in recovery of language function. In some recovered
aphasics, language function is interfered with when the right hemisphere is temporarily inactivated (Kinsbourne, 1971). The lack of
knowledge of the preinjury degree of hemispheric specialization makes
such studies somewhat difficult to interpret.
9.2.4. Within-Hemispheric Specialization and Differentiation in
Humans
Consider the hypothesis that the anterior and posterior language
areas function in analogy with Devices A and B in the conceptual
model, and that the dense cortico-cortical connections between them
(the arcuate fasciculus and possibly the uncinate fasciculus) are important in developing and to some degree maintaining differentiation of
function. This hypothesis may be an aswer to the puzzlement expressed
by Galaburda (1982) and cited in Chapter 2: How can richly interconnected and architectonically similar areas have different functions?
There is indication of incomplete anteroposterior differentiation
of language areas in normal development, as reduction in fluency of
speech follows injury in any part of the language areas in children,
whereas it is related to anterior injuries in adults (Alajouanine &
Lhermitte, 1965).
It may be discussed whether the model in Figure 9.1 or Figure
9.2 gives the best explanation of the data. The original WernickeLichtheim hypothesis assumes that Figure 9.1 depicts the functional
155
specialization of the language areas, with Device A specialized for

articulatory programming and Device B for auditory language
perception.
In the present study, the findings led to the acceptance of a less
localized representation for most language performances, including
auditory comprehension. The evidence for localization (specialization)
is strongest with respect to fluency. Maintaining fluency depends not
only on the Broca area but on the insula and possibly on the basal
ganglia. If the term anterior language area is taken to refer to this anatomical subsystem, then I suggest that the relation between language
areas is as in Figure 9.2, where A now is the anterior and B the
posterior language area. As a result of further dynamic sharpening
of specialization, the posterior language area will also perform in a
differential mode, but depending to a greater extent on input from
the anterior language area to uphold this differentiation than vice
versa.
9.2.5. The Effect of Lesions and the Systemic Basis of Recovery

Wehave so far considered the Situation where two components
or areas of the brain are systemically related and serve to sharpen
each other' s response profiles by reciprocal inhibition. If one of the
components is injured, then the other suffers some lass of ability to
respond differentially. The lass may be moderate, as in Figure 9.1,
or severe, as in Figure 9.2, depending on the design features of the
preserved component. Lass of differentiation will be termed Stage 1
in the response to injury.
Although one is apt to think in dichotomies of brain structures,
as left-right or anterior-posterior, it is not likely the case that systemic
organization is built only on structures connected in pairwise fashion.
Figure 9.3 shows a more complex situation with three systemically
related components.
Component A does best at Task 1, Component Bat Task 2, and
Component C at Task 3. The performance of these tasks is therefore
"boosted" in the respective components, whereas the performance of
other tasks is inhibited. The arrows indicate the direction of the systemic influence on the ability of a given component to perform a given
task.
156
CHAPTER 9
"
100
~
Cl..
~
<lJ
<..J
t::
<
'I::
~
50
10
\I
2
100
50
50
!/
10
10
2
TllSk
too
Ta.sk
Task
Figure 9.3. System with three components.
Injury to Component A causes some loss of differentiation in

both Components Band C (Stage 1; see Figure 9.4). It now appears,
however, that Component Bis relatively betterat Task 1 than Component C. This potential capability of Component B to solve Task 1
is "unmasked" by the lesion of A. The term unmasking has been used
to describe the situation where neural pathways that are present but
inefficient may assume functional importance when more efficient
pathways are disrupted (for review, see Bach-y-Rita, 1981). The case
SfJJ.ge 1
A
100
10D
10o
.s
50
./.'
!IJ
JN]URED
-~
!/)
10
10
2
Task
10
.1
7sk
.3
Task
Figure 9.4. Loss of differentiation after injury (Stage 1).
157
described above is analogaus to unmasking, but at a systems Ievel.

The same term has therefore been used, but in quotation marks.
With time, the reciprocal influence of B and C willlead to some
redifferentiation of function in both components. This redifferentiation is termed Stage 2 in response to injury. The most important
consequence of redifferentiation in the hypothetical example is that,
in Component B, its natural capacity for performing Task 1 is boosted
rather than inhibited, as was the case before injury to A (see Figure
9.5).
An interesting question is whether Tasks 2 and 3 will return to
a normal functional Ievel. My guess is that the loss of one major
systemic component cannot be fully compensated for. In any case, it
is important to note that Tasks 1, 2, and 3 all improve in paralleland
as a result of the same underlying process of differentiation. Although
recovery takes place in parallel, there will, of course, be differences
in the Ievel from which recovery starts and in the Ievel ultimately
attained. The idea that some functions have to deteriorate or be "sacrificed" so that others may improve is foreign to this model.
The immediate effects of a lesion followed by dedifferentiation
are encompassing deficits and high correlations between a wide range
of performances. The physiological responses of the brain in the acute
Stage 2
B
100
1j
e
'ij
~
Q.j
50
QJ
:::>.
100
50
10
10
<f-
Task
Task
Figure 9.5. Redifferentiation of function after injury (Stage 2).
158
CHAPTER 9
stage of illness tend to add to this picture, but the effects of dedifferentiation can be distinguished from those of acute physiological
responses like edema by being in effect for a much Ionger time period,
usually months after the insult.
With limited lesions, some specialized neurological structures
are usually preserved, tagether with cortico-cortical pathways. This
means that the basis of recovery may be a system in which little
relocalization has taken place, if the preserved areas are able to uphold
the previous pattern of localization.
9.3. Testing the Model

As stated in Chapter 1, general systems theory is not a testable
theory but a framework for formulating theories to account for phenomena showing evidence of organized complexity. Adding specific
assumptions makes a theory testable. Within a systems theory framework, the present theory has made several assertions that are in
principle testable. These assertions may be summarized as follows:
Fully developed hemispheric specialization depends on a differentiation process. A necessary condition for such a process is an
anatomical bias or asymmetry, which in the case of the hemispheres
is the advantage in size of the language areas of the left side over the
homologaus areas of the right (Geschwind & Levitsky, 1968; Wada,
Clark, & Hamm, 1975). Within the left hemisphere, anterior-posterior
differentiation depends on a more rigid specification of functional
characteristics of the anterior language area than of the posterior. The
evidence for this conclusion comes from clinical data; however, there
is as yet no anatomical evidence for asserting a higher degree of
specialization of the anterior than of the posterior language area.
Areasthat are connected via bidirectional fiber tracts, that show
homological structure, and that show some degree of anatomical
asymmetry are candidates for developing functional differentiation.
The nature of the input to such a system may serve to speed or
retard the differentiation process. High demands for functional competence speed the differentiation process, andin general a more differentiated pattern of functionallocalization leads to higher functional
efficiency.
159
The result of the differentiation process is not a conglomerate

of single-channel processing devices, but a set of multichannel processing units differing from each other in functional profile rather than
in the nature of the functions represented. Functional localization,
even when resulting from dynamic differentiation, becomes more
rigid and entrenched ("mechanized" is the term used in general systems theory) with advancing age. Structural darnage to the system
Ieads to some loss of differentiation but with some possibilities for
new differentiation to develop.
9.3 .1. Application of the Model to the Present Findings

In the introduction to this chapter, three examples are given of
findings from the present study requiring a complex explanation.
In Example 1, the multiple overlapping organization of the functions underlying verbal and nonverbal memory is cited. The model
described above is suited to explain this type of organization but does
not contain assumptions that allow a further discussion of this example.
In Example 2, it is pointed out that the general features of the
classical clinico-pathological model were confirmed by the results.
These features are accounted for in the model by making the anteriorposterior difference in function a consequence of hard-wired specialization of the anterior language area. Strong indications were found
of context-sensitive lesion effects, as would be predicted from the sort
of interplay of dynamic forces postulated by the model. Because auditory comprehension is highly sensitive to an interplay between lesion
loci, no anatomically specialized substrate for that function can be
asserted.
If the present model is essentially correct, then some serious
methodological consequences follow for clinico-pathological research.
Todetermine the specific functional role of any anatomical structure
is a monumental task requiring extensive parametric studies. Patients
with "pure" lesions arenot a sufficient basis for drawing any inferences.
In Example 3, several contrasting pattems of recovery are pointed
out. This example is weil suited to bring tagether several features of
the model. Severe aphasia (large lesions) Ieads to more extensive
involvement of nonverbal functions than mild aphasia (smalllesions),
because the comerstones of functional differentiation are to a greater
160
CHAPTER 9
extent destroyed. In severe aphasics, the conditions favoring better

recovery are those favoring a more radicalloss of differentiation with
ensuing redifferentiation and relocalization of function. If the lesion
is anterior, then the posterior language area has a greater degree of
freedom to enter into new relationships. If the lesion is posterior,
then the anterior language area has less freedom to adjust because
its mode of functioning is to a greater extent specialized. Add the
higher degree of mechanization of function in advanced age, and the
prediction follows that patients who are severe, fluent, and old have
the worst prognosis. This three-way interaction could not be tested
directly, but, of the possible two-way interactions (Severity x Type,
Severity x Age, and Type x Age), the latter two were found tobe
significant. The Severity x Chronicity interaction showed that it was
only the severe (not the mild) aphasics who made a better recovery
in the acute than in the chronic phase. This interaction supports the
hypothesis that there are different recovery mechanisms at work in
severe and mild aphasics. In mild aphasics, the process underlying
recovery is probably to a greater extent dependent on effectivization
of processing in preserved parts of the language areas and development of effective cueing strategies. This process is not strongly
dependent on time after injury.
9.3.2. General Applications of the Model

The model has the virtue of not postulating special mechanisms
that only have application in a limited context. In addition to the
model' s applicability to research on hemispheric specialization, two
examples of related models used to explain normal phenomena are
mentioned here.
The phenomena of selective attention have been addressed by
Kinsbourne & Hicks (1978), whose views were cited in Chapter 1. A
paradigm for studying selective attention is the dual-task interaction
paradigm, in which the person has to perform two tasks or processes
at the same time. The results of such experiments bear on the nature
of the cognitive processing system as a serial or parallel processing
device, an issue of central concern in cognitive theory. There have
been recent attempts to specify differentiated resource pools underlying cognitive performance and to tie them to a neurological substrate, that is, the cerebral hemispheres (Friedman & Polson, 1981;
161
Navon & Gopher, 1979). If it is hypothesized that cerebrally localized

functions are resource pools, then the present model can explain both
relatively constant, independenthuman resource pools and the ability
of subjects to learn with extensive practice to perform some types of
dual tasks by invoking the mechanisms of specialization and differentiation. By specifying necessary conditions for differentiation, the
model avoids the emptiness of allowing unlimited and arbitrary postulation of resource pools. Walley and Weiden (1973) addressed the
sameproblern area, taking the analogue of lateral inhibition in sensory
systems (Ratliffe, 1961) as an explicit model for interaction and differentiation between cognitive functions.
The second example of a related model is found in the theory
of Witkin, Goodenough, and Oltmann (1977), in which cognitive and
emotional development is characterized as starting from a relatively
global, diffuse, undifferentiated state and developing toward a more
differentiated, context-independent mode of cognition. Witkin et al.
(1977) cited cerebral asymmetry as a neural correlate of differentiation.
The concept of cognitive style implies the ability of individuals to vary
along a continuum of degrees of differentiation according to experiential background, sex, and situational demands. The concept of
differentiation is also central in the theory of personality presented
by Royce and Powell (1983). They presented evidence derived from
factor analytic studies of a continuing developmental process of differentiation and hierarchization of functions.
The present model thus seems to incorporate concepts and processes central to the psychological theory of cognitive processes. Its
novelty is in applying these concepts to the analysis of the consequences of cerebral injury. In doing so, it has prompted some viewpoints on the neurological substrate to emerge, which are of interest
to cognitive theory.
9.4. Concluding Remarks

One of the starting points of the present essay was a plea for a
research strategy that might yield results applicable to a wide range
of aphasics, not only to selected and rare cases. It is appropriate,
therefore, to devote the final discussion to this point.
162
CHAPTER 9
The study has presented a Iot of descriptive information that is

generally of interest to clinicians and users of similar methods. More
important, however, is the general model derived from these results.
The goal of cognitive and neurolinguistic analysis is to bring to
light processing units that fail to function. The basis for inferring such
units is a model based on the study of normal performance (Marshall
1982; see Chapter 1, p. 3). The underlying assumption is that the
normal system can be used as a model for the impaired system. This
assumption is of course essentially correct, but in the present view it
is only strictly valid in the case of mildly aphasic individuals with
smalllesions. Only in these cases can the whole retraining program
be based on the strategy of identifying, and then strengthening or
bypassing, weak links in the chain of processing.
Our patients give us many clues that there is something more
amiss with them than the failure to execute certain subroutines of
information processing. They show various degrees of disability in
conforming to task requirements; they are distractible, perseverating,
or unable to refrain from unrelated activity (e.g., talking when a nonverbal response is called for). Some neuropsychological theories have
tried to come to grips with such problems by stressing the preprocessing or "micro-genetic" aspect of any cognitive task. Before the highly
differentiated information processing routine can be executed, the
processing system has to go through stages of preparation to meet
the task requirements.
The neuropsychological theory of Brown (1979) was cited earlier
(Chapter 1, p. 15). He used the term "micro-genesis" and stressed
the analogy between phylogenetic and ontogenetic development of
a process and its unfolding in the present.
The neuropsychological theory advanced here differs from that
of Brown in emphasis. The present theory does not stress hierarchical
aspects of processing but emphasizes differentiation between types
of processes dependent on cortical areas that are presumably closely
related in both phylogenesis and ontogenesis.
An undifferentiated brain can only process integrated or "fused"
information (e.g., by verbal-contextual-emotional integration). If the
presented information is not fused but demands parallel or selective
processing at an early (preprocessing) stage, then interference may
result. Likewise, the emission of fused responses (e.g., by verbalgestural-autonomic integration) may be feasible, but selective or
163
sequentially organized activation is hampered by interference. These

phenomena are weil known to experienced clinicians, who can sometimes exploit them to make patients perform surprising feats of comprehension or expression. These severely aphasic patients, however,
do poorly in formal training programs and on formal testing.
The fault with our approach to severe aphasics, then, may be
that we think of them as only quantitatively different from the mild
aphasics. The conclusion of this essay is that impaired brains differ
in the degree of functional differentiation present after injury. If this
conclusion is correct, both the content of therapy and the mode of
task presentation in therapy must depend on assumptions about the
capacity for differentiated processing present in the preserved system.
A goal of therapy is not only to exploit the abnormal interachans of
an undifferentiated system but to further the process of differentiation. A first step toward this goal is to understand the more general
features of the differentiation process. Cognizant of the difficulties
and of the many steps that will have to follow if practical benefits are
to be attained, I offer the model presented here as a first step.
REFERENCES
Alajouanine, T. H., & Lhermitte, F. (1965). Acquired aphasia in children.

Brain, 88, 653-662.
Alexander, M. P., & LoVerme, S. R. (1980). Aphasia after left hemispheric
intracerebral hemorrhage. Neurology, 30, 1193-1202.
Alexander, M. P., & Schmitt, M. A. (1980). The aphasia syndrome of stroke
in the left anterior cerebral artery territory. Archives of Neurology, 37, 97100.
Arbib, M., & Caplan, D. (1979). Neurolinguistics must be computational. The

Behavioral and Brain Sciences, 2, 449-483.
Arena, R., & Gainotti, G. (1978). Constructional apraxia and visuoperceptive
disabilities in relation to laterality of cerebrallesions. Cortex, 14, 463-473.
Arrigoni, G., & de Renzi, E. (1964). Constructional apraxia and hemispheric
locus of lesion. Cortex, 1, 170--197.
Atkinson, R. C., & Shiffrin, R. M. (1968). Human memory: A proposed system
and its control processes. In K. W. Spence & J. T. Spence (Eds.), The
psychology of learning and motivation (Vol. 2). New York: Academic Press.
Bach-y-Rita, P. (1981). Brain plasticity as a basis of the development of rehabilitation procedures for hemiplegia. Scandinavian Journal Rehabilitation
Medicine, 13, 73-84.
Baddeley, A. D., & Hitch, G. (1_974). Warking memory. In G. H. Bower (Ed.),
The psychology of learning and motivation (Vol. 8). New York: Academic
Press.
Bailey, P., & von Bonin, G. (1951). The isocortex of man. Urbana: University
of Illinois.
Basso, A., de Renzi, E., Faglioni, P., Scotti, G., & Spinnler, H. (1973). Neuropsychological evidence for the existence of cerebral areas critical to
performance of intelligence tasks. Brain, 96, 715-728.
Basso, A., Faglioni, P., & Vignolo, L. A. (1975). Etude controlee de la reeducation du Iangage dans l'aphasie: Comparaison entre aphasiques traites
et non-traite. Revue Neurologique, 131, 607-614.
165
166
REFERENCES
Basso, A., Capitani, E., & Vignolo, L. A. (1979). Influence of rehabilitation

on language skills in aphasic patients: A controlled study. Archives of
Neurology, 36, 190-196.
Basso, A., Capitani, E., Luzzati, C., & Spinnler, H. (1981). Intelligence and
left hemisphere disease. Brain, 104, 721-734.
Basso, A., Spinnler, H., Vallar, G., & Zanobio, M. E. (1982). Left hemisphere
darnage and selective impairment of auditory verbal short term memory.
A case study. Neuropsychologia, 20, 263-274.
Bay, E. (1962). Aphasia and non-verbal disorders of language. Brain, 3, 412426.
Bay, E. (1966). The classification of disorders of speech. Cortex, 3, 26-31.
Beauvois, M. F., & Lhermitte, F. (1975). Repetition immediate et apprentissage d'une serie de mots chez 30 sujets aphasiques. Neuropsychologia, 13,
247-251.
Benson, D. F. (1977). The third alexia. Archives of Neurology, 34, 327-331.
Benson, D. F., & Geschwind, N. (1969). The alexias. In P. }. Vinken & G. W.
Bruyn (Eds.), Handbook of clinical neurology. Amsterdam: North Holland
Publishing Company.
Benson, D. F., & Geschwind, N. (1977). The aphasias and related disturbances. In A. B. Baker & L. H. Baker (Eds.), Handbook of Clinical Neurology
(Vol. 1). New York: Harper and Row.
Benson, D. F., Sheremata, W. A., Bouchard, R., Segarra, }. M., Price, D., &
Geschwind, N. (1973). Conduction aphasia: A clinicopathological study.
Archives of Neurology, 28, 339-346.
Benton, A. L. (1961). The fiction of the "Gerstmann syndrome." Journal of
Neurology, Neurosurgery and Psychiatry, 24, 176-181.
Benton, A. L. (1967). Problems of test construction in the field of aphasia.
Cortex, 3, 32-53.
Black, F. W., & Strub, R. L. (1976). Constructional apraxia in patients with
discrete missile wounds of the brain. Cortex, 12, 212-220.
Black, F. W., & Strub, R. L. (1978). Digitrepetitionperformance in patients
with focal brain damage. Cortex, 14, 12-21.
Blunk, R., DeBleser, R., Willmes, K., & Zeumer, H. (1981). A refined method
to relate morphological and functional aspects of aphasia. European Neurology, 20, 69-79.
Bogen, J. E., & Bogen, G. M. (1976). Wernicke's region-Where is it? InS. R.
Harnad et al. (Eds. ), Origins and evolution of language and speech. New York:
New York Academy of Sciences.
Borod, J. C., Carper, M., & Goodglass, H. (1982). WAlSperformance IQ in
aphasia as a function of auditory comprehension and constructional
apraxia. Cortex, 18, 199-210.
Bradshaw, J. L., & Nettleton, N. C. (1981). The nature of hemispheric specialization in man. The Behavioral and Brain Sciences, 4, 51-92.
Broadbent, D. E. (1984). The maltese cross: A new simplistic model for memory. The Behavioral and Brain Sciences, 7, 55-94.
REFERENCES
167
Broca, P. (1861). Remarques sur Ia siege de Ia faculte du Iangage articule;

suivies d'une observation d'aphemie (perte de Ia parole). Bulletin de Ia
Societe Anatomique, 6, 330-357.
Broman, T., Lindholm, A., & Melin, B. (1967). Rehabilitering av afasipasienter. Lkartidningen, 64, 4595-4600.
Brown, J. W. (1979). Language representation in the brain. In H. Steklis &
M. Raleigh (Eds.), Neurobiology of social communication in primates. New
York: Academic Press.
Brown, J. W., & Jaffe, J. (1975). Hypothesis on cerebral dominance. Neuropsychologia, 13, 107-110.
Brunner, R. ]., Kornhuber, H. H., Seemuller, E., Suger, G., & Wallesch,
C-W. (1982). Basal ganglia participation in language pathology. Brain
and Language, 16, 281-299.
Brust, J. C M., Schafer, S., Richter, R., & Bruun, B. (1976). Aphasia in acute
stroke. Strake, 7, 167-174.
Bub, 0., & Kertesz, A. (1982). Deep agraphia. Brain and Language, 17, 146165.
Butfield, E., & Zangwill, 0. L. (1946). Re-education in aphasia: A review of
70 cases. Journal of Neurology, Neurosurgery and Psychiatry, 7, 75-79.
Caplan, D. (1982). Representation in neurology and linguistics. In M. A.
Arbib, D. Caplan, & J. C Marshall (Eds.), Neuralmodels of language processes. New York: Academic Press.
Caramazza, A., & Berndt, R. S. (1978). Semantic and syntactic processes in
aphasia: A review of the literature. Psychological Bulletin, 85, 898-918.
Carson, D. H., Carson, F. E., & Tikofsky, R. S. (1968). On learning characteristics of the adult aphasic. Cortex, 4, 92-112.
Castro-Caldas, A., & Botelho, M. A. S. (1980). Dichotic listening in the recovery of aphasia after stroke. Brain and Language, 10, 145-151.
Cattell, R. B. (1978). The scientific use of factor analysis in behavioural and life
sciences. New York: Plenum Press.
Cermak, L. S., & Moreines, J. (1976). Verbalretention deficits in aphasic and
amnestic patients. Brain and Language, 3, 16-27.
Cermak, L. S., & Tarlow, S. (1978). Aphasie and amnesic patients' verbal vs.
nonverbal retentive abilities. Cortex, 14, 32-40.
Chomsky, N. (1965). Aspects of the theory of syntax. Cambridge, Mass.: MIT
Press.
Christensen, A.-L. (1975). Luria's neuropsychological investigation. Copenhagen:
Munksgaard.
Cicone, M., Wapner, W., Foldi, N., Zurif, E., & Gardner, H. (1979). The
relation between gesture and language in aphasic communication. Brain
and Language, 8, 32~349.
Coltheart, M., Patterson, K., & Marshall, J. C (Eds.). (1980). Deep dyslexia.
London: Routledge & Kegan Paul.
Conrad, R. (1964). Acoustic confusion in immediate memory. British Journal
of Psychology, 55, 75-84.
168
REFERENCES
Corsi, P. M. (1972). Human memory and the medial temporal region of the brain.
Ph.D. dissertation, McGill University, Montreal.
Costa, L. D. (1975). The relation of visuospatial dysfunction to digit span
performance in patients with cerebrallesions. Cortex, 11, 31-36.
Costa, L. D. (1976). Interset variability on the Raven Coloured Progressive
Matrices as an indicator of specific ability deficit in brain-lesioned patients.
Cortex, 12, 31-40.
Crowder, R. G. (1976). Principles of learning and memory. Hillsdale, N.J.: Lawrence Erlbaum.
Crowder, R. G., & Morton, J. (1969). Precategorical acoustic storage (PAS).
Perception and Psychophysics, 5, 365-373.
Culton, G. L. (1969). Spontaneous recovery from aphasia. Journal of Speech
and Hearing Research, 12, 825-832.
Damasio, A. R., Damasio, H., Rizzo, M., Varney, N., & Gersh, F. (1982).
Aphasia with nonhemorrhagic lesions in the basal ganglion and internal
capsule. Archives of Neurology, 39, 15-20.
Dejerine, J. (1892). Contribution a l'etude anatomo-pathologique et clinique
des differentes varietes de cecite verbal. Memoires de la Societe de Biologie,
4, 61-90.
Dejerine, J. (1914). Semieologie des affections du systeme nerveux. Paris: Masson.
Denenberg, V. H. (1978). Dilemmasanddesigns for developmental research.
In C. L. Ludlow & E. Doran-Quine (Eds. ), The neurological bases of language
disorders in children: Methods and directions for research. Bethesda, Maryland:
U.S. Department of Health, Education and Welfare.
Denenberg, V. H. (1981). Hemispheric laterality in animals and the effects of
early experience. The Behavioral and Brain Sciences, 4, 1-50.
Denes, F., Semenza, C., Stoppa, E., & Gradenigo, G. (1978). Selective
improvement by unilateral brain-damaged patients on Raven Coloured
Progressive Matrices. Neuropsychologia, 16, 749-752.
Dennis, M. (1976). Dissociated naming and locating of body parts after left
anterior temporallobe resection: An experimental case study. Brain and
Language, 3, 147-163.
Dennis, M., & Whitaker, H. A. (1977). Hemispheric equipotentiality and
language acquisition. InS. J. Segalowitz & F. A. Gruber (Eds.), Language
development and neurological theory. New York: Academic Press.
de Renzi, E. (1982). Disorders of space exploration and cognition. New York: Wiley.
de Renzi, E., & Faglioni, P. (1965). The comparative effect of intelligence and
vigilance tests in detecting hemispheric cerebral damage. Cortex, 1, 410433.
de Renzi, E., Faglioni, P., & Previdi, P. (1977). Spatial memory and hemispheric locus of lesion. Cortex, 13, 424-433.
de Renzi, E., Faglioni, P., & Villa, P. (1977). Sequential memory for figures
in brain-damaged patients. Neuropsychologia, 15, 42-50.
de Renzi, E., & Nichelli, P. (1975). Verbaland non-verbal short term memory
impairment following hemispheric damage. Cortex, 11, 341-354.
REFERENCES
169
de Renzi, E., & Spinnler, H. (1967). Impaired performance on color tasks in

patients with hemispheric damage. Cortex, 3, 194-216.
Duffy, R. J., & Duffy, J. R. (1981). Three studies of deficit in pantomime
expression and pantomime recognition in aphasia. Journal of Speech and
Hearing Research, 24, 70-84.
Duffy, R. J., & Liles, B. Z. (1979). A translation of Finkelnburg' s (1870) lecture
on aphasia as "asymbolia" with commentary. Journal of Speech and Hearing
Disorders, 44, 156-168.
Ebenholz, S. M. (1972). Seriallearning and dimensional organization. In G. H.
Bower (Ed.), The psychology of learning and motivation (Vol. 5). New York:
Academic Press.
Efron, R. (1963). Temporal perception, aphasia and deja vu. Brain, 86, 403424.
Engvik, H., Hjerkinn, 0., & Seim, S. (1980). WAlS: Norsk utgave. Oslo: Norsk
Psykologforening.
Exner, S. (1881). Untersuchungen ber die Localisation der Functionen in der Grasshirnrinde der Menschen. Wien: Braunmueller.
Fedio, P., & van Buren, J. M. (1975). Memory and perceptual deficits during
electrical stimulation in the left and right thalamus and parietal subcortex.
Brain and Language, 2, 78-100.
Finger, S. (1978). Environmental attenuation of brain-lesion symptoms. InS.
Finger (Ed.), Recovery from brain damage. New York: Plenum Press.
Fodor, J., & Bever, T. G. (1965). The psychological reality of linguistic segments. Journal of Verbal Learning and Verbal Behaviour, 4, 414-420.
Fredriksen, A. L., & Lern~s, L. K. (1984). Sex differences in cerebral organization
of language. (In Norwegian). Unpublished thesis, Department of Psychology, University of Oslo.
Friederici, A. D., Schoenle, P. W., & Goodglass, H. (1981). Mechanisms
underlying writing and speech in aphasia. Brain and Language, 13, 212222.
Friedman, A., & Polson, M. C. (1981). Hemispheres as independent resource
systems: Limitecl-capacity processing and cerebral specialization. Journal
of Experimental Psychology: Human Perception and Performance, 7, 1031-1058.
Frostig, M. (1966). Developmental test of visual perception. Palo Alto, Calif.:
Consulting Psychologists Press.
Galaburda, A. M. (1982). Histology, architectonics, and asymmetry of language areas. In M. A. Arbib, D. Caplan, & J. C. Marshall (Eds.), Neural
models of language processes. New York: Academic Press.
Galaburda, A. M., LeMay, M., Kemper, T. L., & Geschwind, N. (1978). Rightleft asymmetries in the brain. Science, 199, 852-856.
Gardner, H., Zurif, E., Berry, T., & Baker, E. (1976). Visual communication
therapy in aphasia. Neuropsychologia, 14, 275-299.
Gazzaniga, M., & Ledoux, J. (1978). The integrated mind. New York: Plenum
Press.
Gazzaniga, M. S., Velletri Glass, A., Sarno, M. T., & Posner, J. B. (1973).
170
REFERENCES
Pure word deafness and hemispheric dynamics: A case history. Cortex,

9, 136-143.
Geschwind, N. (1965). Disconnection syndromes in animals and man. Brain,
88, 234-294, 585-644.
Geschwind, N. (1967a). The apraxias. In E. W. Strauss & R. M. Griffith
(Eds.), Phenomenology of will and action. Pittsburgh: Duquesne University
Press.
Geschwind, N. (1967b). The variety of naming errors. Cortex, 3, 97-112.
Geschwind, N. (1979). Specializations of the human brain. Scientific American,
241, 180-199.
Geschwind, N., & Fusillo, M. (1966). Color-naming defects in association
with alexia. Archives of Neurology, 15, 137-146.
Geschwind, N., & Levitsky, W. (1968). Human brain: Left-right asymmetry
in temporal speech region. Science, 161, 186-187.
Geschwind, N., Quadfasel, F., & Segarra, J. M. (1968). Isolation of the speech
area. Neuropsychologia, 6, 327-340.
Glanzer, M. (1972). Storage mechanisms in recall. In G. H. Bower (Ed.), The
psychology of learning and motivation (Vol. 5). New York: Academic Press.
Glick, S. 0., & Zimmerberg, B. (1978). Pharmacological modification of brain
lesion syndromes. InS. Finger (Ed.), Recovery from brain damage: Research
and theory. New York: Plenum Press.
Goldstein, K. (1948). Language and language disturbances. New York: Grune
and Stratton.
Goodglass, H., Denes, G., & Calderon, M. (1974). The absence of covert
verbal mediation in aphasia. Cortex, 10, 264-269.
Goodglass, H., Gleason, J. B., & Hyde, M. (1970). Some dimensions of auditory language comprehension in aphasia. Journal of Speech and Hearing
Research, 13, 595-606.
Goodglass, H., & Kaplan, E. (1963). Disturbance of gesture and pantomime
in aphasia. Brain, 86, 703-720.
Goodglass, H., & Kaplan, E. (1972). The assessment of aphasia. Philadelphia:
Lea and Febiger.
Goodglass, H., Klein, B., Carey, P. W., & Jones, K. J. (1966). Specific semantic
word categories in aphasia. Cortex, 2, 74-89.
Gordon, W. P. (1983). Memory disorders in aphasia-1. Auditory immediate
recall. Neuropsychologia, 21, 325-339.
Green, E., & Howes, D. (1977). The nature of conduction aphasia. In H.
Whitaker & H. A. Whitaker (Eds.), Studies in neurolinguistics (Vol. 3).
New York: Academic Press.
Hamsher, K. (1982). Intelligence and aphasia. In M. T. Sarno (Ed.), Acquired
aphasia. New York: Academic Press.
Hamsher, K., Benton, A. L., & Digre, K. (1980). Serial digit learning: Normative and clinical aspects. Journal of Clinical Neuropsychology, 2, 39-50.
Hasher, L., & Zacks, R. T. (1979). Automatie and effortful processes in memory. Journal of Experimental Psychology: General, 108, 356-388.
REFERENCES
171
Head, H. (1915). Hughling Jackson on aphasia and kindred affections of

speech. Brain, 38, 1-27.
Head, H. (1926). Aphasia and kindred disorders of speech. Cambridge: Cambridge
University Press.
Hecaen, H., & Assal, G. A. (1970). Comparison of constructive deficits following right and left hemispheric lesions. Neuropsychologia, 8, 289-303.
Hecaen, H., & Consoli, S. (1973). Analyse des troubles du langage au cours
des lesions de l'aire de Broca. Neuropsychologia, 11, 371-388.
Hecaen, H., & Dubois, J. (1971). La neurolinguistique. In G. E. Perren &
J. L. M. Trim (Eds.), Applications of linguistics. Cambridge: Cambridge
University Press, 1971.
Hecaen, H., & Kremin, H. (1976). Neurolinguistic research on reading disorders resulting from left hemisphere lesions: Aphasie and "pure" alexia.
In H. Whitaker & H. A. Whitaker (Eds.), Studies in neurolinguistics (Vol.
2). New York: Academic Press.
Heilman, K. M., Safran, A., & Geschwind, N. (1971). Closed head trauma
and aphasia. Journal of Neurology, Neurosurgery and Psychiatry, 34, 265-269.
Heilman, K. M., Schales, R., & Watson, R. T. (1976). Defects of immediate

memory in Broca's and conduction aphasia. Brain and Language, 3, 201208.
Henschen, S. E. (1922). Klinische und anatomische Beitraege zur Pathologie des

Gehirns (Vols. 5--7). Stockholm: Nordiska Bokhandel.
Hirst, W. (1982). The amnesic syndrome: Descriptions and explanations. Psychological Bulletin, 91, 435-460.
Howes, D., & Geschwind, N. (1964). Quantitative studies of aphasic language. In D. Rioch & E. Weinstein (Eds.), Disorders of communication.
Baltimore: Williams and Wilkins.
Huber, W., Stachowiack, F.-J., Poeck, K., & Kerschensteiner, M. (1975). Die
Wernicke Aphasie. Journal of Neurology, 210, 77-97.
Huber, W., Poeck, K., Weniger, D., & Willmes, K. (1983). Der Aachener Aphasie
Test (AAT). Gttingen: Hogrefe.
Jackson, J. H. (1878). On affections of speech from disease of the brain. Brain,
1, 304-330.
Jakobson, R. (1971). Aphasia as a linguistic topic. Reprinted in Roman Jakobsan, Selected Writings (Vol. 2). The Hague: Mouton.
Johnson, J. P., Sommers, R. K., & Weidener, W. E. (1977). Dichotic ear
preference in aphasia. Journal of Speech and Hearing Research, 20, 116--129.
Jones, E. G., & Powell, T. P. S. (1970). An anatomical study of converging
sensory pathways within the cerebral cortex of the monkey. Brain, 93,
793-820.
Jones, L. V., & Wepman, J. M. (1961). Dimensions of language performance

in aphasia. Journal of Speech and Hearing Research, 4, 220--232.
Kenin, M., & Swisher, L. (1972). A study of pattern of recovery in aphasia.
Cortex, 8, 56--68.
172
REFERENCES
Kerschensteiner, M., Poeck, K., & Brunner, E. (1972). The fluency-nonfluency

dimension in the classification of aphasic speech. Cortex, 8, 233-247.
Kerschensteiner, M., Poeck, K., Huber, W., Stachowiack, F.-J., & Weniger,
D. (1975). Die Broca Aphasie. Journal of Neurology, 217, 223-242.
Kertesz, A. (1979). Aphasia and associated disorders. New York: Grune and
Stratton.
Kertesz, A., & Benson, D. F. (1970). Neologistic jargon: A clinico-pathological
study. Cortex, 6, 362-386.
Kertesz, A., & Hooper, P. (1982). Praxis and language: The extent and variety
of apraxia in aphasia. Neuropsychologia, 20, 275-286.
Kertesz, A., & McCabe, P. (1975). Intelligence and aphasia: Performance of
aphasics on Raven's Coloured Progressive Matrices (RCPM). Brain and
Language, 2, 387-395.
Kertesz, A., & McCabe, P. (1977). Recoverypatternsand prognosis in aphasia.
Brain, 100, 1-18.
Kertesz, A., & Phipps, J. B. (1977). Numerical taxonomy of aphasia. Brain
and Language, 4, 1-10.
Kertesz, A., & Phipps, J. (1980). The numerical taxonomy of acute and chronic
aphasic syndromes. Psychological Research, 41, 179-198.
Kertesz, A., & Poole, E. (1974). The aphasia quotient: The taxonomic approach
to measurement of aphasic disability. Canadian Journal of Neurological Sciences, 1, 7-16.
Kertesz, A., & Sheppard, A. (1981). The epidemiology of aphasic and cognitive impairment in strake. Brain, 104, 117-128.
Kertesz, A., Harlock, W., & Coates, R. (1979). Computer tomographic localization, lesion size, and prognosis in aphasia and nonverbal impairment.
Kertesz, A., Lesk, D., & McCabe, P. (1977). Isotope localization of infarcts
in aphasia. Archives of Neurology, 34, 590-601.
Kimura, D. (1979). Neuromotor mechanisms in the evolution of human communication. In H. Steklis & M. Raleigh (Eds.), Neurobiology of social communication in primates. New York: Academic Press.
Kimura, D. (1983). Sex differences in cerebral organization for speech and
praxis functions. Canadian Journal of Psychology, 37, 19-35.
Kinsbourne, M. (1971). The minor cerebral hemisphere as a source of aphasic
speech. Archives of Neurology, 25, 302-306.
Kinsbourne, M. (1982). Hemispheric specialization and the growth of human
understanding. American Psychologist, 37, 411-420.
Kinsbourne, M., & Hicks, R. E. (1978). Functional cerebral space: A model
for overflow, transfer and interference effects in human performance: A
tutorial review. In J. Requin (Ed.), Attention and performance (Vol. 7).
Hillsdale, N.J.: Lawrence Erlbaum.
Kinsbourne, M., & Warrington, E. K. (1964). Observations on colour agnosia.
Journal of Neurology, Neurosurgery and Psychiatry, 27, 296-299.
Kohn, B., & Dennis, M. (1974). Selective impairments of visuospatial abilities
REFERENCES
173
in infantile hemiplegics after right hemidecortication. Neuropsychologia,

12, 505-512.
Kolers, P. A., Palet, S. R., & Stelmach, L. B. (1980). Graphemic analysis

underlying literacy. Memory and Cognition, 8, 322-328.
Kreindler, A., & Fradis, A. (1968). Performances in aphasia. Paris: Gauthier-Villars.
Laine, T., & Marttila, R. J. (1981). Pure agraphia: A case study. Neuropsychologia, 19, 311-316.
Lamendella, J. T. (1977). The limbic systeminhuman communication. In H.
Whitaker &1 H. A. Whitaker (Eds.), Studies in neurolinguistics (Vol. 3).
New York: Academic Press.
Lassen, N. A., Ingvar, D. H., & Skinhoj, E. (1978). Brain function and blood
flow. Scientific American, 239, 50--59.
Lebrun, Y., & Hoops, R. (Eds.). (1974). Intelligence and aphasia. Amsterdam:
Swets and Zeitlinger.
Lecours, A. R., & Rouillon, F. (1976). Neurolinguistic analysis of jargon aphasia
and jargon agraphia. In H. Whitaker & H. A. Whitaker (Eds. ), Studies in
neurolinguistics (Vol. 2). New York: Academic Press.
Lehman, E. B. (1982). Memory for modality: Evidence for an automatic process. Memory and Cognition, 10, 554-564.
Lehmkuhl, G., Poeck, K., & Willmes, K. (1983). Ideomotor apraxia and aphasia:
An examination of types and manifestations of apraxic symptoms. Neuropsychologia, 21, 199-212.
Leischner, A. (1972). ber den Verlauf und die Einteilung der aphasischen
Syndrome. Archiv fr Psychiatrie und Nervenkrankheit, 216, 219-231.
Lenneberg, E. H. (1967). Biological foundations of language. New York: Wiley.
Levine, D., & Calvanio, R. (1982). The neurology of reading disorders. In
M. A. Arbib, D. Ca plan, & J. C. Marshall (Eds. ), Neural models of language
processes. New York: Academic Press.
Levy, B. A. (1983). Proofreading familiar text: Constraints on visual processing. Memory and Cognition, 11, 1-12.
Lichtheim, L. (1885). On aphasia. Brain, 7, 443-484.
Liepmann, H. (1900). Das Krankheitsbild der Apraxie ("motorischen Asymbolie"). Monatschrift fr Psychologie und Neurologie, Val. 8.
Liepmann, H. (1915). Diseases of the brain. In C. W. Barr (Ed. ), Curschmanns
textbook an nervaus diseases (Vol. 1). Philadelphia: Blakiston.
Lomas, J., & Kertesz, A. (1978). Patterns of spontaneaus recovery in aphasic
groups: A study of adult stroke patients. Brain and Language, 5, 388-401.
Luria, A. R. (1966). Higher cortical function in man. New York: Basic Books.
Luria, A. R. (1970). Traumatic aphasia (English ed.). The Hague: Mouton.
Luria, A. R. (1973). The working brain. London: Penguin.
Luria, A. R. (1977). On quasi-apha~ic speech disturbances in lesions of the
deep structures of the brain. Brain and Language, 4, 432-459.
Marie, P. (1906). Revision de Ia question de l'aphasie: La troisieme circonvolution frontale gauehe ne joue aucun role special dans Ia fonction du
Iangage. Semaine Medicale, 26, 241-247.
174
REFERENCES
Marquardsen, J. (1969). The natural history of acute cerebrovascular disease.

Acta Neurologica Scandinavica, 45, Supp. 38.
Marshall, J. C. (1982). What is a symptom-complex? In M. A. Arbib, D.
Caplan, & J. C. Marshall (Eds.), Neuralmodels of language processes. New
York: Academic Press.
Marshall, J. C., & Newcombe, F. (1973). Patterns of paralexia: A psycholinguistic approach. Psycholinguistic Research, 2, 175-199.
Matsui, T., & Hirano, A. (1978). An atlas of the human brain for computerized
tomography. New York: Fischer.
McFarling, D., Rothi, L. J., & Heilman, K. M. (1982). Transcortical aphasia
from ischemic infarcts of the thalamus: A report on two cases. Journal of
Neurology, Neurosurgery and Psychiatry, 45, 107-112.
McFie, J. (1975). Assessment of organic intellectual impairment. New York: Academic Press.
McGlone, J. (1980). Sex differences in human cerebral asymmetry: A critical
survey. The Behavioral and Brain Sciences, 3, 215-263.
Milner, B. (1962). Laterality effects in audition. In V. B. Mountcastle (Ed.),
Interhemispheric relations and cerebral dominance. Baltimore: The Johns Hopkins Press.
Milner, B. (1974). Hemispheric specialization: Scope and Iimits. In F. 0. Schmitt
& F. G. Worden (Eds.), The neurosciences: Third study program. Cambridge,
Mass: MIT Press.
Milner, B., Branch, C., & Rasmussen, T. (1964). Observations on cerebral
dominance. In A. V. S. de Reuch & M. O'Connor (Eds.), Disorders of
language. Boston: Little, Brown.
Mohr, J. P. (1976). Broca's area and Broca's aphasia. In H. Whitaker & H. A.
Whitaker (Eds.), Studies in Neurolinguistics (Vol. 1). New York: Academic
Press.
Mohr, J. P., Pessin, M., Finkelstein, S., Finkelstein, H., Duncan, G., & Grand
Davis, K. (1978). Broca aphasia: Pathological and clinical aspects. Neurology, 28, 311-324.
Mohr, J. P., Sidman, M., Stoddard, L. T., Leicester, J., & Rosenberger, P. B.
(1973). Evolution of the deficit in total aphasia. Neurology, 23, 13021312.
Morton, J. (1979). Facilitation in word recognition: Experiments causing change

in the Iogogen model. In P. A. Kolers, M. Wrolstad, & H. Bouma (Eds.),
Processing of visible language (Vol. 1). New York: Plenum Press.
Moscowitch, M. (1977). The development of lateralization of language functions and its relation to cognitive and linguistic development: A review
and some theoretical speculations. In S. J. Segalowitz & F. A. Gruber
(Eds.), Language development and neurological theory. New York: Academic
Press.
Mountcastle, V. B. (1957). Modality and topographic properties of single
neurons of cat's somatic sensory cortex. Journal of Neurophysiology, 20,
408-434.
REFERENCES
175
Naeser, A., & Hayward, R. W. (1978). Lesion localization in aphasia with

cranial computed tomography and the Boston Diagnostic Aphasia Exam.
Neurology, 28, 545-551.
Navon, D., & Gopher, D. (1979). On the economy of the human-processing
system. Psychological Review, 86, 214-255.
Newcombe, F. (1969). Missile wounds of the brain. London: Oxford University
Press.
Nie, N. H., Hadlai Hull, C., Jenkins, G. J., Steinbrenner, K., & Brent, D. H.
(1975). SPSS: Statistical Package for the Social Sciences. New York: McGrawHill.
Nunally, J. C. (1967). Psychometrie theory. New York: McGraw-Hill.
bler, L. K., Albert, M. L., Goodglass, H., & Benson, D. F. (1978). Aphasia
type and aging. Brain and Language, 6, 318-322.
Orgass, B., Hartje, W., Kerschensteiner, M., & Poeck, K. (1972). Aphasie
und nicht-sprachliche Intelligentz. Nervenartzt, 43, 623-627.
Orgass, B., Poeck, K., & Kerschensteiner, M. (1974). Das Verstandnis fr
Nomina mit spezifischer Referenz bei aphasischen Patienten. Zeitschrift
fr Neurologie, 206, 95-102.
Oxbury, J. M., Oxbury, S. M., & Humphrey, N. K. (1969). Variety of colour
anomia. Brain, 92, 847-860.
Pandya, D. N., & Galaburda, A. M. (1980). Role of architectonics and connections in the study of primate brain evolution. American Journal of
Physical Anthropology, 52, 197.
Penfield, W., & Roberts, L. (1959). Speech and brain mechanisms. Princeton,
N.J.: Princeton University Press.
Petlund, C.-F. (1970). Prevalence and invalidity from strake in Aust-Agder county
of Norway. Oslo: Universitetsforlaget.
Pettit, J. M., & Noll, J. D. (1979). Cerebra! dominance in aphasia recovery.
Poeck, K. (1983a). Ideational apraxia. Journal of Neurology, 230, 1-5.
Poeck, K. (1983b). What dowe mean by "aphasic syndromes"? A neurologist's
view. Brain and Language, 20, 79-89.
Poeck, K., & Huber, W.(1977). To what extent is language a sequential activity?
Neuropsychologia, 15, 359-363.
Poeck, K., & Lehmkuhl, G. (1980). Das Syndrom der ideatorischen Apraxie
und seine Lokalisation. Nervenartzt, 51, 217-225.
Poeck, K., & Stachowiak, F.-J. (1975). Farbbenennungsstoerungen bei aphasischen und nichtaphasischen Hirnkranken. Journal of Neurology, 209, 95102.
Pribram, K. H. (1971). Languages of the brain. Englewood Cliffs, N.J.: PrenticeHall.
Prins, R., Snow, C., & Wagenaar, E. (1978). Recovery from aphasia: Spontaneausspeech versus language comprehension. Brain and Language, 5,
192-211.
Raisman, G., & Field, P. M. (1973). A quantitative investigation of the
176
REFERENCES
development of collateral reinnervation of the septal nuclei. Brain Research,

50, 241-264.
Rasmussen, T., & Milner, B. (1975). Clinical and surgical studies of the cerebral speech areas in man. In K. S. Zlch et al. (Eds.), Cerebral/oca/ization.
New York: Springer.
Ratliffe, F. (1961). Inhibitory interaction and the detection and enhancement
ofcontours. InS. W. Rosenblith (Ed.), Sensorycommunication. Cambridge,
Mass.: MIT Press.
Raven, J. C. (1960). Coloured Progressive Matrices. London: H. K. Lewis.
Razy, A., Janotta, F. S., & Lehner, L. H. (1969). Aphasia resulting from
occlusion of the left anterior cerebral artery. Archives of Neurology, 36,
221-225.
Reinvang, I. (1981). Patterns of restitution in aphasia. Paper presented at INS
European conference, Bergen, Norway, June.
Reinvang, I. (1983). The systemically organized neural basis of language:
Aphasia syndromes and their recovery. Doctoral dissertation, University of
Oslo.
Reinvang, 1., & Dugstad, G. (1981). Aphasia typology and lesion localization
with computed axial tomography. Scandinavian Journal of Rehabilitation
Medicine, 13, 8~8.
Reinvang, 1., & Engvik, H. (1980a). Language recovery in aphasics from 3 to
6 months after stroke. In M. T. Sarno & 0. Hk (Eds.), Aphasia: Assessment and treatment. Stockholm: Almquist & Wiksell.
Reinvang, 1., & Engvik, H. (1980b). Norsk Grunntest for Afasi. Oslo: Universitetsforlaget. (Danish edition, Copenhagen: Dansk Psykologisk Forlag,
1984.)
Reitan, R. M., & Davison, L. A. (Eds.). (1974). Clinical neuropsychology: Current
status and applications. Washington, D.C.: Winston.
Richardson, J. T. E. (1977). Functional relationship between forward and
backward digit repetition and a non-verbal analogue. Cortex, 13, 317320.
Robinson, B. W. (1976). Limbic influences on human speech. In S. Harnad
et al. (Eds.), Origins and evolution of language and speech. New York: The
New York Academy of Sciences.
Robinson, R. G., & Benson, D. F. (1981). Depression in aphasic patients:
Frequency, severity and clinico-pathological correlations. Brain and Language, 14, 282-291.
Rothi, L. J., & Hutchinson, E. C. (1981). Retention of verbal information by
rehearsal in relation to the fluency of verbal output in aphasia. Brain and
Language, 12, 347-359.
Royce, J., & Powell, A. (1983). Theory of personality and individual differences:
Factors, systems, and processes. Englewood Cliffs, N.J.: Prentice Hall.
Rubens, A. B. (1976). Transcortical motor aphasia. In H. Whitaker & H. A.
Whitaker (Eds.), Studies in neurolinguistics (Vol. 1). New York: Academic
Press.
REFERENCES
177
Rudel, R. G., & Denckla, M. B. (1974). Relation of forward and backward

digit repetition to neurological impairment in children with learning disabilities. Neuropsychologia, 12, 109-118.
Russo, M., & Vignolo, L. A. (1967). Visual figure-ground discrimination in
patients with unilateral cerebral disease. Cortex, 3, 113--127.
Sarno, M. T. (1976). The status of research in recovery from aphasia. In Y.
Lebrun & R. Hoops (Eds.), Recovery in aphasics. Amsterdam: Swets and
Zeitlinger.
Sarno, M. T., & Levita, E. (1971). Naturalcourse of recovery in severe aphasia.
Archives of Physical Medicine and Rehabilitation, 52, 175-179.
Sarno, M. T., & Levita, E. (1979). Recovery in treated aphasia in the first year
post-strake. Strake, 10, 663--670.
Sarno, M. T., & Levita, E. (1981). Some observations on the nature of recovery
in global aphasia. Brain and Language, 13, 1-12.
Sarno, M. T., Silverman, M., & Sands, E. (1970). Speech therapy and language
recovery in severe aphasia. Journal of Speech and Hearing Research, 13, 607623.
Schmachtemberg, A., Hundgen, R., & Zeumer, H. (1983). Ein EDV-adaptiertes Rastermodell des Gehirns zur topographischen Analyse von Lsionen im kranialen Computertomogramm. Fortschritte Rntgenstrahlung,
139, 499-502.
Schneider, G. E. (1973). Early lesions of superior colliculus: Factars affecting

the formation of abnormal retinal projections. Brain Behavior and Evolution,
8, 73--109.
Schuell, H. (1974). Aphasia theory and therapy. New York: Macmillan.

Schuell, H., Jenkins, J. J., & Carroll, J. B. (1962). A factor analysis of the
Minnesota test for differential diagnosis of aphasia. Journal of Speech and
Hearing Research, 5, 349-369.
Schuell, H., Jenkins, J. J., & Jimenez-Pabon, E. (1965). Aphasia in adults. New
York: Harper and Row.
Schwartz, M. F. (1984). What the classical aphasia categories can't do for us
and why. Brain and Language, 21, 3--8.
Seines, 0. A. (1974). The corpus callosum: Some anatomical and functional
considerations with special reference to language. Brain and Language, 1,
111-140.
Shallice, T. (1979). Case study approach in neuropsychological research. Journal of Clinical Neuropsychology, 1, 183--211.
Shallice, T., & Warrington, E. K. (1974). The dissociation between short term
retention of meaningful soundsandverbal material. Neuropsychologia, 12,
553--555.
Smith, A., Champoux, R., Levi, J., London, R., & Muraski, A. (1972). Diagnosis, intelligence and rehabilitation of chronic aphasics. Ann Arbor: University
of Michigan.
Snedecor, G. W., & Cochran, W. G. (1967). Statistical methods. Ames: The
Iowa University Press.
178
REFERENCES
Speike, E., Hirst, W., & Neisser, U. (1976). Skills of divided attention. Cognition, 4, 215-230.
Sperling, G. (1960). The information available in brief visual presentations.
Psychological Monographs, 74, No. 498.
Stachowiak, F. ]., Huber, W., Kerschensteiner, M., Poeck, K., & Weniger,
D. (1977). Die globale Aphasie. Journal of Neurology, 214, 75-87.
Studdert-Kennedy, M., & Shankweiler, D. P. (1970). Hemispheric specialization for speech perception. Journal of the Acoustical Society of America,
48, 579-594.
Subirana, A. (1969). Handedness and cerebral dominance. In P. Vinken &

G. Bruyn (Eds. ), Handbook of clinical neurology (Vol. 4). New York: Elsevier.
Sundet, K., & Engvik, H. (1984). The validity of aphasic subtypes. Paper presented at INS-European conference, Aachen, West Germany, June.
Swisher, L. P., & Hirsh, I. J. (1972). Brain darnage and the ordering of two
temporally successive stimuli. Neuropsychologia, 10, 137-152.
Szentagothai, J., & Arbib, M. A. (1975). Conceptual models of neural organization.
Cambridge, Mass.: MIT Press.
Teuber, H. L. (1976). Complex functions of basal ganglia. In M. D. Yahr (Ed.),
The basal ganglia. New York: Raven Press.
Teuber, H. L., & Weinstein, S. (1956). Ability to discover hidden figures after
cerebrallesions. Archives of Neurology and Psychiatry, 76, 369-379.
Tulving, E. (1972). Episodic and semantic memory. In E. Tulving and W.
Donaldson (Eds.), Organization of memory. New York: Academic Press.
Tulving, E. (1982). Elements of episodic memory. New York: Oxford University
Press.
Tzortzis, C., & Albert, M. L. (1974). Impairment of memory for sequences
in conduction aphasia. Neuropsychologia, 12, 355-366.
Ulrich, G. (1978). Interhemispheric functional relationship in auditory agnosia. Brain and Language, 5, 286-300.
Vanderplas, J. M., & Garvin, E. A. (1959). The association value of random
shapes. Journal of Experimental Psychology, 37, 147-163.
Varney, N. R. (1978). Linguistic correlates of pantomime recognition in aphasic
patients. Journal of Neurology, Neurosurgery and Psychiatry, 41, 564--568.
Vignolo, L. A (1964). Evolution of aphasia and language rehabilitation: A
retrospective exploratory study. Cortex, 1, 344--367.
von Bertallanfy, L. (1948). Generalsystems theory. New York: Braziller.
Wada, J., Clark, R., & Hamm, A. (1975). Cerebra! hemispheric asymmetry
in humans. Archives of Neurology, 32, 239-246.
Wallesch, C.-W., Kornhuber, H. H., Brunner, R. ]., Kunz, T., Hollerbach,
B., & Suger, G. (1983). Lesions of the basal ganglia, thalamus and deep
white matter: Differential effects on language functions. Brain and Language, 20, 286-304.
Walley, R. E., & Weiden, T. D. (1973). Lateralinhibition and cognitive masking: A neuropsychological theory of attention. Psychologica/ Review, 80,
284--302.
REFERENCES
179
Warrington, E. K., & Shallice, T. (1969). The selective impairment of auditory

verbal short-term memory. Brain, 92, 885-896.
Warrington, E. K., Logue, V., & Pratt, R. T. C. (1971). The anatomicallocalization of selective impairment of auditory verbal short-term memory.
Neuropsychologia, 9, 377-387.
Wechsler, D. (1945). Standardized memory scale for clinical use. Journal of
Psychology, 19, 87-95.
Wechsler, D. (1958). The measurement and appraisal of adult intelligence. Baltimore: Williams and Wilkins.
Weinberg, J., Diller, L., Gordon, W. A., Gerstman, L. J., Lieberman, A.,
Lakin, P., Hodges, G., & Ezrachi, 0. (1977). Visual scanning training:
Effect on reading-related tasks in acquired right brain damage. Archives
of Physical Medicine and Rehabilitation, 58, 479-486.
Weinstein, S. (1964). Deficits concomitant with aphasia or lesions of either
hemisphere. Cortex, 1, 154--169.
Weisenburg, T., and McBride, K. (1935). Aphasia. New York: The Commonwealth Fund.
Weiss, P. (1969). The living system: determinism stratified. In A. Koestler
and J. R. Smythies (Eds.), Beyond reductionism. Boston: Beacon.
Wepman, J. M. (1953). Recovery from aphasia. New York: Ronald.
Wernicke, C. (1874). Der aphasische Symptomen-complex. Breslau: Cohn und
Weigert. Reprinted in Boston Studies on the Philosophy of Science (Vol. 4).
Dordrecht: Reidl, 1964.
Whitaker, H. (1971). Neurolinguistics. In W. 0. Dingwall (Ed.), A survey of
linguistic science. College Park: University of Maryland.
White, N., & Kinsqourne, M. (1980). Does speech output controllateralize
over time? Evidence from verbal-manual time-sharing tasks. Brain and
Language, 10, 215-223.
Wiegei-Crump, C. A. (1976). Agrammatism and aphasia. In Y. Lebrun & R.
Hoops (Eds.), Recovery in aphasics. Amsterdam: Swets and Zeitlinger.
Wiegel-Crump, C. A., & Koenigsknecht, R. (1973). Tapping the lexical store
of the adult aphasic: Analysis of the improvement made in word retrieval
skills. Cortex, 9, 410-418.
Willmes, K., Poeck, K., Weiniger, D., & Huber, W. (1980). Der Aachener
Aphasie Test: Differentielle Validitt. Nervenartzt, 51, 55~560.
Witkin, H. A., Goodenough, D. R., & Oltmann, P. K. (1977). Psychological
differentiation: Current status. Princeton, N.J.: Educational Testing Service.
Wood, C. C. (1978). Variations on a theme by Lashley: Lesion experiments
on the neural model of Anderson, Silverstein, Ritz and Jones. Psychological Review, 85, 582-591.
Woods, B. T., & Teuber, H. L. (1973). Early onset of complementary specialization of cerebral hemispheres in man. Transactions of the American
Neurological Association, 98,
11~115.
Woolsey, C. N. (1958). Organization of somatic sensory and motor areas of

the cerebral cortex. In H. F. Harlow & C. N. Woolsey (Eds.), Biological
180
REFERENCES
and biochemical bases of behavior. Madison: University of Wisconsin

Press.
Yakovlev, P. 1., & Lecours, A. R. (1967). The myelogenic cycles of regional
maturation in the brain. In A. Minkowski (Ed.), Regional development of
the brain in early life. Oxford: Blackwell.
Zangwill, 0. (1964). Intelligence in aphasia. In A. V. S. deReuch & M. O'Connor (Eds.), Disorders of language. Boston: Little, Brown.
Zurif, E. B., & Caramazza, A. (1976). Psycholinguistic structures in aphasia:
Studies in syntax and semanties. In H. Whitaker & H. A. Whitaker (Eds. ),
Studies in neurolinguistics (Vol. 1). New York: Academic Press.
Appendix
Norsk Grunntest for Afasi

(Norwegian Basic Aphasia
Assessment)
The following is a detailed description of the procedure and content

of the test. Procedures and instructions for the patient are given here
in English. The test items themselves, however, are given in Norwegian with the English translations in parentheses.
A.l. Spontaneaus Speech

The patient should, during the course of an interview, be asked
three specific questions: "What is your occupation?" "Where do you
live?" "What is your favorite TV-program?" He should also be asked
to respond to at least two more general questions. Suggestions are
"Can you tell me about your family?" "What do you usually do during
summer vacation?" The interview should be tape recorded.
A general evaluation of the speech is scored for communicative
efficiency. The ratings are
0--Normal.
1-The patientexpresses his intentions adequately without aid
from the examiner. However, his manner of expression is
deviant.
181
APPENDIX
182
2-The patient produces content words and significant clues

about his intentions, but some degree of guessing and questioning by the examiner is necessary.
3--The examiner must carry the initiative in order to arrive at a
result. The result is questionable, however, because the patient
gives too few clues (e.g., responds only yes or no) or because
he gives partly contradictory responses.
4---A functional communicative relation is not established.
The next step is to score the presence and the degree of the
deviant features that are characteristic of different types of aphasic
speech.
Literal Paraphasia
0--Normal.
1-Speech contains distorted words, but these are in most cases
interpretable and can be recognized as related to a target
word by way of phonemic substitution.
2-A more severe degree of distortion is present, so that interpretability suffers. Grammatical words and filler words are
usually not distorted, but content words may be neologistic.
3--Speech consists of neologistic jargon. This means varied combinations of syllables and phonemes without recognizable
meaning.
Camplex Paraphasia
0--Normal.
1-Speech contains semantically related word substitutions or
circumlocutory expressions. The target word can be readily
guessed.
2-Speech has a topic, but the precise meaning cannot be determined because of bizarre word substitutions and associative
leaps in which the target of communication does not seem
tobe consistently maintained.
3--Uninterpretable jargon is present but mainly with lexically
interpretable words.
APPENDIX
183
Visible Elfort
0-Normal.
1-Some groping and false starts mark the commencement of
an articulatory sequence, but a flow of articulation can then
be maintained over several words.
2-Effortful articulation is apparent on almost every word, except
on automatized phrases. Effort generalizes to facial musculature and to bodily posture.
3--There is an extensive and general mobilization of physical
effort in an attempt to initiate speech. The attempt usually
does not get beyond the first syllable or a stereotype.
Hestitation, Pauses
0-Normal.
1-There is some tendency to make unnatural pauses within a
sentence.
2-Most sentences acquire an unmelodic and broken character
because of frequent pauses.
3--Speech consists of isolated words separated by long pauses.
Stereotypy
0-Not present.
1-An exaggerated use of cliches and empty phrases is apparent.
2-Speech consists exclusively of conventional expressions, for
example, swears or polite phrases.
3--Speech consists of an individual stereotype, often a meaningless syllable. The lack of variation distinguishes stereotypy
from neologistic jargon.
Articulation (Dysarthria)
0-Not present.
1-Mild. Some slurring of speech occurs, but it does not affect
in terpreta bili ty.
APPENDIX
184
2-Moderate. Phonation or articulation is affected to a degree,

making speech difficult to interpret.
3-Severe. A condition of anathria with no or little intelligible
speech exists.
Self Correction
0--Normal reaction. The patient corrects slips of the tongue.
1-The patient reacts to paraphasias and usually produces an
improved response.
2-The patient reacts to paraphasias but does not produce an
improved response.
3-No reaction to severe paraphasia or high degree of stereotypy
is present.
The further scoring of spontaneaus speech is based on a transcription of the patient' s response to the general questions of the
interview. Number of words per minute and number of words per
utterance are calculated.
A.2. Auditory Comprehension

Instruction: "Point to the body part (object) as I name it. Where
is the ---?"
Body parts
Objects
1. Klokke (watch)
1. Tenner (teeth)
2. Ball (ball)
2. Nese (nose)
3.
Sikkerhetsnl (safety pin)
3. 0re (ear)
4. Kopp (cup)
4. Panne (forehead)
5. Nokler (keys)
5. Hand (hand)
6. Knapp (button)
6. Hake (chin)
7. Book (book)
7. Hals
8. Kne (knee)
8. Penn (pen)
9. Stein (stone)
9. Lr (thigh)
10. Albue (elbow)
10. Sokk (sock)
11. Legg (shin)
11. Mynt (coin)
Indirect description (no further instruction given):
185
APPENDIX
Body parts
1. Det du horer med.
Objects
1. Det du drikker kaffe av.
(What you hear with.)

2. Det du tramper med.
2.
(What you stamp with.)

3. Det du biter med.
3.
(What you bite with.)
4. Det du lukter med.

(What you smell with.)
5. Det sm barn sutter p.
(What babies suck on.)
4.
5.
6.
(What you drink coffee

from.)
Det du Iaser opp med.
(What you use to
unlock.)
Det barn kaster til
hverandre.
(What children throw to
one another.)
Det du Ieser i.
(What you read in.)
Det du mler tiden med.
(What is used to measure
time.)
Det du betaler med.
(What you pay with.)
Instruction: "Now I want you to do exactly as I ask."

Body parts
1. Snu deg mot dora.
(Turn toward the door.)

2. Reis deg opp og sett
deg ned.
(Stand up and sit
down.)
3. Tramp med foten.
(Stamp with your foot.)
4. Strekk ut armen og
knytt neven.
(Stretch out your arm
and make a fist.)
5. Legg hnden under
haka.
(Piace your hand under
your chin.)
Objects
1. Sl opp boka.
(Open the book.)

2. Trekk opp klokka.
(Wind the watch.)
3. Snu koppen p hodet.
(Turn the cup on its

head.)
4. Legg sikkerhetsnla i
koppen.
(Put the safety pin in
the cup.)
5. Kast ballen til meg.
(Throw me the ball.)
APPENDIX
186
Body parts
6. Dekk eynene med
hnden.
(Cover your eyes with
your hand.)
7. Kle deg p leggen.
(Scratch your shin.)
8. Left benet med hnden.

(Lift your leg with your
hand.)
9. Pek nese til meg.
(Thumb you nose at
me.)
10. Pek ferst p eret, s p
kneet og s pa albuen.
(Touch the ear, knee,
and elbow in that
order.)
Objects
6. Left neklene og la dem
falle.
(Lift the keys and drop
them.)
7. Skyv koppen vekk fra
deg.
(Push the cup away
from you.)
8. Legg boka opp klokka.
(Put the book on top of
the watch.)
9. Sla opp boka og finn
forordet.
(Open the book and
find the preface.)
10. Pek ferst p boka, s p
ballen og s p klokka.
(Touch the book, ball,
and watch in that
order.)
Instruction: "Respond yes or no to the following questions. Take

time to reflect about them."
Ideas, meaning
1. Brukes en saks til a
klippe med?
(Is a pair of scissors
used to cut with?)
2. Har kyllinger horn?
(Do chickens have
antlers?)
3. Er en hest et dyr?
(Is a horse an animal?)
Ideas, relations
1. Er en dag kortere enn
en uke?
(Is a day shorter than a
week?)
2. Er en bjern sterre enn
en mus?
(Is a bear bigger than a
mause?)
3. Er en bestefar eldre enn
en gutt?
(Is a granddad older
than a boy?)
APPENDIX
Ideas, meaning
4. Er Norge et land?
(Is Norway a nation?)
187
Ideas, relations
4. Er et r lengre enn en
mned?
(Is 1 year longer than 1
month?)
5. Er kongen en kvinne?
(Is the king a female?)
6. Har sauer ull?
(Do sheep have wool?)
7. Er en heks snill?
(Is a witch good?)
8. Brukes en klut til
vispe med?
(Is a rag used to stir
with?)
9. Er vann et metall?
(Is water a metal?)
10. Er en dverg liten?
(Is a dwarf small?)
11. Brukes en oks til
skjcere med?
(Is an ax used to cut
with?)
12. Er en jolle en bat?
(Is a dinghy a boat?)
13. Har hunder snute?
(Do dogs have a snout?)
14. Brukes en kopp til
spise av?
(Is a cup used to eat
from?)
A.3. Repetition
Instruction: "Please repeat these words (sentences) exactly as I
say them."
APPENDIX
188
Words
1. Mann (man)
2. Bord (table)
3. Ire (three or tree)
4. Femten (fifteen)
5. Tolv (twelve)
6. Pil (arrow)
7. Katt (cat)
8. Lys (light)
9. Parafin (kerosene)
10. Syvogtredve (thirtyseven)
11. Fire tusen (four

thousand)
12.
13.
14.
15.
16.
Tomater (tomatoes)
Parkere (park)
Skrekk (horror)
Omvende (convert)
Fangst (catch)
Sentences
1. Fine greier
(A fine mess)
2. Sterke saker
(Hot stuff)
3. Takk for maten
(Thanks for the meal)
4. Ryk og reis
(Son of a gun)
5. Sola skinte hele dagen.
(The sun shone all day.)
6. Du store all verden
(What in the world!)
7. Regnet trommet p
taket.
(The rain was drumming on the roof.)
8. Bten sank i hytt og
v<Er. (The ship sank as
the wind blew.)
9. Han forlangte gjelden
betalt.
(He demanded payment
of the debt.)
10. Aldri annet enn om og
men
(Never anything except
ifs and buts.)
11. Tomme hmner er bedre
enn ti p taket.
(Empty barreis are better than 10 on the roof.)
189
APPENDIX
Sentences
Words
17. Hundreogtreognitti
(hund and ninety-three)
18. Skramleorkester
(junk-instrument
orchestra)
19. Ansette (hire)
20. Janitsjarkonsert (brassband concert)
Instruction: "These words don't mean anything. Try to repeat

them exactly as I say them."
Nonsense syllables
1.
2.
3.
4.
ral
sob
tef
gyp
5.
6.
7.
8.
omlette
foniter
balfere
maloper
A.4. Naming
Instruction: "Now I point to a body part (object) and you teil
me what it is."
Body parts
1. Tenner (teeth)
2. Nese (nose)
3. 0re (ear)
4.
5.
6.
7.
8.
9.
10.
11.
Panne (forehead)
Hand (hand)
Hake (chin)
Hals (throat)
Kne (knee)
Lr (thigh)
Albue (elbow)
Legg (shin)
Objects
1. Klokke (watch)
2. Ball (ball)
3. Sikkerhetsnl (safety
pin)
4. Kopp (cup)
5. N0kler (keys)
6. Knapp (button)
7. Bok (book)
8. Penn (pen)
9. Stein (stone)
10. Sokk (sock)
11. Mynt (coin)
190
APPENDIX
Instruction: "Describe what I am doing."
Body parts
Obiects
1. Reise seg. (Stand up.)
1. Lese. (Read.)
2. Knytte neven. (Make a

fist.)
3. Trampe. (Stamp your
foot.)
4. Dekke 0ynene med hnden. (Cover your eyes
with the hand.)
5. Kl0 seg p !eggen.
2. Kaste ball. (Throw a

ball.)
3. Trekke klokka. (Wind the
watch.)
4. Snu koppen. (Turn the
cup around.)
5. Legge sikkerhetsnla i
koppen. (Put the safety
pin in the cup.)
Instruction: "Respond to the following questions. Abrief answer

is sufficient."
1. Hvilken farge har sn0? (What is the color of snow?)
2. Hvem bor p slottet? (Who lives in the royal castle?)
3. Hvor mange dager er det i en uke? (How many days are
there in a week?)
4. Hva heter den f0rste mned i ret? (Which is the first month
of the year?)
5. Hva brukes spe til? (What is soap used for?)
6. Hva brukes en saks til? (What isapair of scissors used for?)
7. Hva brukes en blyant til? (What is a pencil used for?)
8. Hvor mange kilometer er det i en mil? (How many yards
make a mile?)
9. Hva bruker man hogge ved med? (What do you use to
chop wood?)
10. Hvilken smak har sitroner? (How do lemons taste?)
A.S. Reading
Instruction: "Read what is printed on this card" (reading aloud);
"Point to the Ietter or word that I say" (comprension); "Read the card
and find the object" (comprehension); "Read the card and do what
it says" (comprehension).
APPENDIX
191
Reading aloud
1.
2.
3.
4.
5.
6.
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
B
A
R
0
y
K
Ball (ball)
Bok (book)
Kopp (cup)
Klokke (watch)
Nokler (keys)
Sikkerhetsnl (safety pin)
Tomater (tomatoes)
Jugoslavia (Yugoslavia)
Sentimental (sentimental)
Aktivitet (activity)
2.
3.
4.
5.
(Wind the watch.)

Lukk oynene.
(Close your eyes.)
Kast ballen til meg.
(Throw me the ball.)
Klo deg p leggen.
Ror forst ved ballen, s
ved boka og s ved
klokka.
(Touch the ball, the
book, and the watch in
that order.)
Comprehension
1. B
2. A
3. R
4. 0
5. y
6. K
1. Ball
2.
3.
4.
5.
6.
Bok
Kopp
Klokke
Nokler
Sikkerhetsnl
2. Lukk oynene.
3. Kast ballen til meg.
4. Klo deg p leggen.

5. Ror forst ved ballen, s
ved boka og s ved
klokka.
A.6. Syntax
Instruction: "The words on these three cards make up a sentence. Try to arrange them in the correct order. Do not try to make
questions."
192
APPENDIX
1. Festen varte/ tillangt/ p natt/

(The party Iasted long into the night.)
2. Flyet alle/ ventet p/ kom ikke/
(The plane everyone waited for did not come.)
3. Brevet jenta/ skrev forsvant/ i posten/
(The Ietter the girl wrote got lost in the mail.)
4. Store sultne/ hwer breite/ etter mat
(Big, hungry lions roared for food.)
5. Buksa damen/sydde til/sr;mnen passet/
(The pants the woman sewed for her son fitted.)
6. Fer dagen/var slutt/kom sneen
(Before the day was over the snow came.)
A.7. Writing
Instruction: "Sign your name;" "Copy these words."
1. Kopp (cup)
2. Sikkerhetsnl (safety pin)
"Write the words I say."
1. Ball (ball)
2. Klokke (watch)
"What is this? Write down the name."
1. Nekler (keys)
2. 0re (ear)
"Write the following sentence."
1. Kle deg p leggen. (Scratch your shin.)
2. Bten sank i hytt og v<Er. (The ship sank as the wind blew.)
Index
AAT (Aachener Aphasie Test), 33-34,
45-46
Abstract model, 149-150
Acoustic memory, 72, 74-76, 92-93, 95
Age
and aphasia test scores, 40-41
and classification, 57-58
and prognosis, 132, 141-142, 160
Agnosia
auditory-verbal, 59, 61, 68, 118-119
visual, 60
Agraphia
classifica tion of, 68
description of, 59-60
lesion localization in, 28, 118-119
Alexia
with agraphia, 60
classification of, 68
deep, 61
lesion localization in, 118-119
pure, 60
Angular gyrus, 24, 26, 28
Anomic aphasia, 24, 28, 49, 52-54, 131
Anterior language area, 26, 136, 154155, 158. Sec also Broca area
Insula
'
Aphemia, 59
Apraxia
and aphasia, 99-101
and articulation, 25
and prognosis, 144
recovery of, 140
tests of, 103-104
Arcuate fasciculus, 24, 29, 117-118,
119, 121
Association area, 6, 151
Association fibers, 6, 151-152
Associationism, 7-8, 62
Attention, 112, 134, 160
and prognosis, 132
recovery of, 130, 139
Auditory comprehension (Cant.)

and relation to lesion localization
'
27-29, 33, 121
Basal ganglia, 111-112, 115, 124, 155
Body parts, 32, 61, 66
Boston Diagnostic Aphasia Assessment
(BOA), 33-34, 43
Broca aphasia, 24-25, 48, 51-53, 118,
130, 132
Broca area. See also Anterior language
area
anatomy of, 23-25
function of, 111, 117-119, 121, 124
Chronicity of aphasia
and prognosis, 132, 144
and relation to memory, 95
and relation to recovery, 141-142,
144-145
Cingulate gyrus, 110
Classification of aphasia
and relation to lesion localization
117-119, 128
'
and relation to memory, 89-93
and relation to nonverbal abilities
105-106
'
and relation to prognosis, 132-133,
144
system of, 6, 47-51, 55-56
Clinicopathological model, 10, 119,
147, 159
Cluster analysis, 8, 54-55
Communication, 31
Conduction aphasia, 24, 28-29, 49, 5253, 131
Corpus callosum, 13-14, 113, 152
Diagnosis
and relation to prognosis, 141
of sample, 37, 56
193
194
Differentiation, 14, 151, 153-154, 156,
158-161, 162-163
Disconnection, 29, 60, 62, 99
Dominance, 5, 135. See also Specialization of hemispheres
Dysarthria, 25, 31, 48
Education, 40-42, 57
Effort, 31, 48
Error types, 76, 87-90
Factor analysis
of aphasia test, 38-39, 65-66
of lesion scores, 117
of memory tests, 85-89
of nonverbal tests, 104
of recovery groups, 144-145
Fluency. See also Speech
in early lesions, 154
and locus of lesion, 119, 124, 155
and prognosis, 133,. 141-142, 160
Frontal lobe, 24, 29, 113
Functional system, 17-18, 134
Gestures, 100, 162
Global aphasia, 14, 28, 47, 51-52, 69,
114, 118-119, 130, 132
Hesitation, 31, 48
Hierarchy, 15, 62, 162
Hippocampus, 110
Inhibition, 8, 13--14, 134, 151, 155, 161
Isolated-speech-area sydrome, 24, 29,
48, 51-52, 117
Insula, 25--26, 29, 111, 115, 117-118,
119, 124, 155. See also Anterior
language area
Intelligence, 9-10, 98, 104. See also
Nonverbal functions
Interaction
of classification parameters, 89-90,
106, 141-142
of functions, 4, 12, 161
of hypothetical devices, 149-150, 153
of lesion areas, 12, 111, 114, 128
Jargon aphasia, 27, 47, 49, 52-53, 118
Language area, 5, 107. See also Anterior language area; Posterior
language area
Lateralization, 13--14, See also Specialization of hemispheres
INDEX
Learning, 72-73, 76--77, 133--134
Lenticular zone, 111
Letters, 61
Limbic system, 110
Localization. Sec also WernickeLichtheim model; Clinicocopathological model
Long-term memory, 71, 76, 92-93
Material-specificity, 10, 61, 65, 69, 94
Mechanism
of perception, 11, 63
of recovery, 137
Mechanization, 12, 159
Memory
of nonverbal material, 77-78, 92-93,
143, 146, 159
of verbal material, 72, 74-76, 144,
146, 159
Modality-specificity, 59, 64-65, 69
Model
of recovery, 133--136
of selected aphasias, 62-63
Motor functions, 20, 73, 103
Naming
of colors, 61
and object category, 62
recovery of, 130
and relation to lesion localization,
28, 124
test of, 32
Neural model, 151-152
Neurolinguistics, 2, 58
Neuronal network, 15--16, 147
Neuropsychological approach, 4, 16,
133, 162
Neurosensory Center Comprehensive
Examinabon for Aphasia
(NCCEA), 32-33, 43
Nonverbal functions. See also Intelligence; Memory of nonverbal
material; Apraxia
and aphasia, 100-101
and localization of lesions, 100, 107
and recovery pattern, 159
Objects
comprehension of, 32, 66
manipulation of, 104
naming of, 33, 66
Occipital lobe, 28
Organization, 4, 95, 147-148. See also
Systems theory
195
INDEX
Paraphasia, 27, 29, 30-31, 48
Parietal lobe, 27
Perseveration, 90--91, 93, 162. See also
Error types
Posterior language area, 26, 112, 121,
135, 154--155, 158
Prognosis, 128, 141-144. See also
Recovery
Raven Colared Progressive Matrices
(RCPM), 99, 101-102
Reading, 25, 27, 29, 33, 61, 67, 121,
139. See also Alexia
Recovery
pattern of, 136--139
process of, 129
of nonverbal functions, 131, 139-140
and systemic basis, 27, 155
Redifferentiation, 135, 157
Reorganization, 134
Repetition
and Iocalization of Iesions, 27, 29,
119
and memory, 76
recovery of, 130, 139
test of, 32
Resource pool, 160-161
Self-correction, 31
Severity of aphasia
classification of, 55
measurement of, 37
and nonverbal functions, 105
and prognosis, 141, 143-144, 160
Sex
and aphasia classification, 56
and aphasia test results, 41, 43
and prognosis, 132, 141-142
Short-term memory, 71, 73--75, 92-93.
See also Memory of nonverbal
material; Memory of verbal
material
Specializa tion
of hemispheres, 13, 151-153, 158
of hypothetical devices, 150
of language areas, 5, 13
Speech. See also Fluency; Dysarthria
classifica tion of, 48
fluent type of, 27-29, 48, 51, 69
Speech (Cant.)
nonfluent type of, 25, 28, 48, 51
Spontaneous speech, 29, 31-32
Stereotypy, 28, 31, 48
Structural analysis, 2. See also Factor
analysis
Supramarginal gyrus, 26--27. See also
Wernicke area; Posterior Ianguage area
Snydrome, 2-4, 10, 20-21, 130. See also
Classification of Aphasia; Wernicke-Lichtheim model
Systemic model, 16--18
Systems theory, 11-12, 149, 155--158.
See also Organization; Interaction
Temporal lobe, 26
Tests. See also AAT; BOA; NCCEA;
RCPM; W AB; W AIS
of apraxia, 103--104
construction of, 30-33
of Iearning, 82, 83--84
of memory, 72, 80-82, 83
of motor functions, 103
of nonverbal functions, 101-103
normative data of, 40-43
standardization of, 35-37
Thalamus, 112, 124
Therapy, 129
Transcortical aphasias, 24, 30, 49, 5253, 118
Types of aphasia. See Classification of
aphasia
Visuoconstructive deficits, 97-98
Wechsler Adult Intelligence Test
(WAlS), 99, 102-103
Wernicke aphasia, 24, 27, 49, 52-53,
118--119, 131, 132
Wernicke area, 23, 24, 26, 117-118,
121, 124
Wernicke-Lichtheim model, 17, 23-24,
109-110, 154
Western Aphasia Battery (WAB), 33-34, 44, 45-46, 52-55
Writing, 25, 27, 30, 33, 67, 138. See also
Agraphia

Aphasia and Brain Organization

Uploaded by

Document Informationclick to expand document informationlibro de afasia

Document Informationclick to expand document information

Copyright:

Available Formats

Aphasia and Brain Organization

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Aphasia and Brain Organization

Uploaded by

Copyright:

Available Formats

APHASIA and

PSYCHOLOGY OF LANGVAGE AND THOUGHT

APHASIA AND BRAIN ORGANIZATION

SPRINGER SCIENCE+BUSINESS MEDIA, LLC

Library of Congress Cataloging in Publication Data

1985 Springer Science+Business Media New York

in many cases interachans between severallesion sites are important

Chapter 1. Approaches to the Study of Aphasia

Chapter 2. Operationalization of a Model

Chapter 3. Types of Aphasia

Chapter 4. Selective Aphasias

Chapter 5. Memory and Learning Deficits

Chapter 6. Defects of Visual Nonverbal Abilities

Chapter 7. Localization of Lesion in Aphasia

70201. The "Limbic System" 0 0 00 0 0 0 0 00 0 0 0 00 0

Chapter 8 Recovery and Prognosis

The Recovery Process 0 0 0 00 0 0 0 0 00 0 0 0 00 0 0 0 00 129

Chapter 9 The Organized Response of the Brain to Injury

901. Evidence for Organized Complexity 00 0 0 00 0 0 148

9.2.2. Neural Model ....................... 151

1.1. Clinical and Theoretical Approaches

fonnulated in different ways, and different methodological approaches

APPROACHES TO THE STUDY OF APHASIA

1. The Ievel at which the nature of computation is expressed. With

Clinical and theoretical motivations have been closely wedded

For the neurolinguist, the unit of analysis is language, and

and nonclinical models of it as its units of study-grossly disregards

APPROACHES TO THE STUDY OF APHASIA

1.2. Historical Approaches

1.2 .1. Localization Theory of Language-to-Brain Relation

that show the most consistent relationships to the locus of injury, a

1.2.2. Criticism of Basic Premises

APPROACHES TO THE STUDY OF APHASIA

However, if care is taken to define behavioral categories in behavior

APPROACHES TO THE STUDY OF APHASIA

historically. The inadequacy of an associationist model of the brain,

1.2.3. Criticisms of Assumptions Regarding the Nature of Aphasia

is symptomatic of a more general intellectual disturbance

The spoken and written narne of an object is not a new attribute of

APPROACHES TO THE STUDY OF APHASIA

1.3. Systems Theory Approach

Conditions for considering a systems theory type of analysis are

APPROACHES TO THE STUDY OF APHASIA

difficulty of deriving testable predictions. General systems theory is

1.3.1 Applications of Systems Theory Concepts in Neuropsychology

Brown (1979) was concerned with the mechanisms of symptom

APPROACHES TO THE STUDY OF APHASIA

This is called a micro-genetic process.

These authors explored the dimensions of functional cerebral space

The concept has been used in studies of developing lateralization

The insistence on the interactive, dynamic properties of neural

1.3.2 Localization of Function in Light of Systems Theory

Therefore, because of the known specifity and diversity of anatomical

APPROACHES TO THE STUDY OF APHASIA

which says that the brain is a collection of independently working

it may be that the language areas function in this way in relation to

APPROACHES TO THE STUDY OF APHASIA

1.4. The Present Study