NEURAL CORRELATES OF WORKING MEMORY IN PRIMARY INSOMNIA

http://dx.doi.org/10.5665/sleep.2952

Neural Correlates of Working Memory Performance in Primary Insomnia

Sean P. A. Drummond, PhD1,2,3,4; Matthew Walker, PhD6; Erin Almklov, PhD2; Manuel Campos2; Dane E. Anderson2,5; Laura D. Straus2,4
1
Psychology Service, VA San Diego Healthcare System, San Diego, CA; 2Research Service, VA San Diego Healthcare System, San Diego, CA;
3
University of California San Diego, Department of Psychiatry, San Diego, CA; 4San Diego State University-University of California San Diego Joint
Doctoral Program in Clinical Psychology, San Diego, CA; 5San Diego State University Interdisciplinary Masters Program, San Diego, CA; 6University
of California, Berkeley, Department of Psychology, Berkeley, CA

Study Objectives: To examine neural correlates of working memory performance in patients with primary insomnia (PIs) compared with well-
matched good sleepers (GSs).
Design: Twenty-five PIs and 25 GSs underwent functional MRI while performing an N-back working memory task.
Setting: VA hospital sleep laboratory and University-based functional imaging center.

Downloaded from https://academic.oup.com/sleep/article/36/9/1307/2577899 by guest on 17 May 2022

Patients or Participants: 25 PIs, 25 GSs.
Interventions: N/A.
Measurements and Results: Although PIs did not differ from GSs in cognitive performance, PIs showed the expected differences from GSs in both
self-reported and objective sleep measures. PIs, relative to GSs, showed reduced activation of task-related working memory regions. This manifested
both as an overall reduction in activation of task-related regions and specifically as reduced modulation of right dorsolateral prefrontal cortex with
increasing task difficulty. Similarly, PIs showed reduced modulation (i.e., reduced deactivation) of default mode regions with increasing task difficulty,
relative to GSs. However, PIs showed intact performance.
Conclusions: These data establish a profile of abnormal neural function in primary insomnia, reflected both in reduced engagement of task-
appropriate brain regions and an inability to modulate task-irrelevant (i.e., default mode) brain areas during working memory performance. These
data have implications for better understanding the neuropathophysiology of the well established, yet little understood, discrepancy between
ubiquitous subjective cognitive complaints in primary insomnia and the rarely found objective deficits during testing.
Keywords: Default mode, functional MRI, insomnia, N-back, sleep, working memory
Citation: Drummond SPA; Walker M; Almklov E; Campos M; Anderson DE; Straus LD. Neural correlates of working memory performance in
primary insomnia. SLEEP 2013;36(9):1307-1316.

INTRODUCTION date for overt cognitive impairment in insomnia patients is

Approximately 5-35% of the population report sleep distur- equivocal.15 Patients with insomnia exhibit slightly slower
bance problems at some point in their lives1-3 and approximately reaction times, may show more attention lapses, often show
10-15% suffer from persistent insomnia.4 Insomnia is associated impaired declarative memory, typically (though not always)
with diminished quality of life,5,6 reduced work productivity,7 show intact working memory, and may show greater deficits
and increased absenteeism8; it increases vulnerability for medical on more complex versions of tasks.15-19 These findings suggest
disorders,8,9 psychiatric disorders,10,11 and accidents and injuries12; patients with insomnia may show differential vulnerability
and is associated with increased health care utilization13,14 and depending on the specific cognitive process examined and/
decreased subjective physical and mental health.2 In addition, or the difficulty level of the task. However, objective cogni-
patients with insomnia often complain of impaired concentration, tive performance deficits in insomnia seem to be subtle and
memory, and ability to accomplish daily tasks. These complaints remain in stark contrast to the level of subjective complaint
worsen as their sleep problems increase in severity.4 about cognition in primary insomnia (PI).
Given the ubiquitous nature of daytime complaints in this The failure to find clear patterns of cognitive impairment may
population, one would expect substantial literature supporting be due to the following factors: (1) patients have an exaggerated
the notion that patients with insomnia experience significant sense of their impairment; (2) typical cognitive tests are too
objectively measured cognitive deficits. However, data to insensitive to find differences; and/or (3) there are alterations in
support this concept are surprisingly sparse. Studies inves- the neural substrates of cognitive performance in insomnia, and
tigating daytime performance in patients with insomnia these alterations relate to the discrepancy between subjective
have focused on three main areas of performance: atten- and objective performance.15 This latter hypothesis is supported
tion, learning, and working memory. Overall, evidence to by the single functional magnetic resonance imaging (MRI)
study to assess cognition in insomnia,20 as well as indirectly
supported by other functional21 or structural22-24 MRI studies
A commentary on this article appears in this issue on page 1273. published dealing with insomnia, each of which reports abnor-
malities in insomnia relative to healthy sleepers. For example,
Submitted for publication October, 2012 in the one study to date using functional MRI to examine
Submitted in final revised form January, 2013 cognitive performance in insomnia, Altena et al.20 studied 21
Accepted for publication January, 2013 patients with chronic insomnia and 12 normal control patients
Address correspondence to: Sean P. A. Drummond, PhD, VA San Diego during performance of two verbal fluency tasks. On each task,
Healthcare System, 3350 La Jolla Village Dr., MC 151B, San Diego, CA patients with insomnia showed decreased cerebral responses in
92161; Tel: (858) 642-1274; E-mail: drummond@ucsd.edu left inferior frontal regions typically associated with language
SLEEP, Vol. 36, No. 9, 2013 1307 Neural Correlates of Working Memory in Insomnia—Drummond et al
 be 25 to 50 y old; (4) not meet criteria for any Axis I
Table 1—Demographics
disorder other than insomnia or a specific phobia and
Cohen have no personal history of any Axis I disorder other
GS PI P d
than a single major depressive episode or a specific
Age, years, mean ± SD 32.4 ± 7.1 32.3 ± 7.2 0.967 -0.011
phobia; and (5) not be taking any over-the-counter
Education, years, mean ± SD 15.6 ± 1.2 15.8 ± 1.8 0.684 0.119 sleep medication or psychotropic medication. Handed-
Sex, number of females (%) 12 (48) 12 (48) ness was controlled due to potential hemispheric differ-
Ethnicity, n (%) ences on imaging measures. Age range was restricted
Hispanic or Latino(a) 5 (20) 3 (12) to minimize circadian rhythm influences on the diag-
Race, n (%) noses of PI, increase our ability to recruit a medically
White 17 (68) 20 (80) healthy sample, and reduce aging effects in the func-
American Indian or Alaskan Native 0 (0) 0 (0) tional MRI measures. Patients with insomnia had to
Asian 4 (16) 0 (0) meet criteria for PI as assessed with Duke Structured
Native Hawaiian or Pacific Islander 1 (4) 0 (0) Interview for Sleep Disorders, including sleep difficul-
Black or African American 2 (8) 4 (16) ties ≥ 3 nights/week for ≥ 3 months. They also had

Downloaded from https://academic.oup.com/sleep/article/36/9/1307/2577899 by guest on 17 May 2022

Mixed race 1 (4) 1 (4) to report the following averages on 1 week of sleep
diaries (see following paragraphs): ≥ 45 min total time
P values are univariate analyses not corrected for multiple comparisons. Cohen awake after lights out (i.e., sleep latency + wake after
d effects sizes are positive when PI > GS and negative when PI < GS. GS, good sleep onset), and either < 6 h total sleep time or sleep
sleeper; PI, primary insomnia; SD, standard deviation.
efficiency < 80%. Good sleepers (GSs) had to report
(in interview and on sleep diaries) 7-9 h total sleep
tasks and left prefrontal regions associated with executive time per night, an average sleep efficiency ≥ 90%, taking fewer
control. Performance was measured only indirectly (self-report than one daytime nap per week, and having no daytime perfor-
of number of words generated) in that study, though no group mance complaints. GSs were matched one-to-one with PIs on
differences were found. age (± 3 y), sex, and education (± 3 y).
Here, we seek to extend the literature by using functional
MRI to examine another fundamental capacity of human cogni- Procedures
tion, working memory. Given the suggestion in the literature Following consent, patients were (1) screened for eligi-
that patients with insomnia may be particularly vulnerable to bility criteria listed previously; (2) completed a series of self-
more complex tasks, we chose an N-back task with multiple report measures assessing sleep, depression, and fatigue (these
levels of difficulty. Our behavioral hypothesis was patients measures allowed us to better characterize and differentiate the
with insomnia would show intact performance on N-back two samples, and provided further documentation the PIs did not
tasks. Given the dearth of insomnia studies using functional have comorbid depression); and (3) completed sleep diaries and
MRI to measure cognition, functional MRI hypotheses were wore an actigraph for 7-10 days. PI and GS status was based on
also informed by studies on obstructive sleep apnea, and to a interview and sleep diaries. Patients then underwent two consecu-
lesser extent, acute sleep deprivation. Here, we hypothesized tive nights of polysomnography (PSG). The first night served as
that patients with insomnia would show increased activation an adaptation night and screen for unreported sleep disorders.
on N-back tasks (consistent with a compensatory response to Twelve hours after waking from the second night in the laboratory,
maintain performance), and this would be most prominent on patients underwent a functional MRI scanning session where they
the hardest version of the task. completed the N-back task. Practice effects were controlled by
having patients take separate, matched versions of the task prior to
METHODS sleep on night 2. Prior pilot studies indicated patients needed two
administrations of the N-back task to reach asymptotic perfor-
Patients mance. Patients refrained from intake of stimulants and alcohol
Patients were recruited from advertisements on Craigslist, for 72 h prior to the functional MRI session (except for small
fliers posted on local college campuses and within the commu- habitual quantities of caffeine in the morning, if appropriate).
nity, and word of mouth. A total of 63 individuals signed
informed consent for the study. Of these, six never started the Cognitive Task
study (all were potential patients with primary insomnia (PIs) The N-back design stresses maintenance and updating of
who did not meet eligibility criteria during the intake interview verbal information in working memory. The task is commonly
after signing consent), and eight withdrew before completion used in cognitive neuroscience studies and activates a reliably
(reasons included anxiety during MRI, sleep study revealed replicated set of working memory brain regions.25 Although
occult sleep disorder, and positive drug test). A total of 50 it has been used in studies of sleep deprivation26,27 and sleep
patients (25 with insomnia and 25 control patients) provided apnea,28-30 this is the first report to use the task in PI. The partici-
data for this report. Demographics for each group are presented pant saw a series of letters on the screen, one at a time, and had
in Table 1. to determine if the letter currently displayed was the same as
To be eligible, all patients needed to meet the following the one presented n-letters ago. There were four levels of diffi-
criteria: (1) have a stable sleep/wake schedule with a preferred culty in the task: 0-back, 1-back, 2-back, and 3-back. Condi-
sleep phase between 22:00 and 08:00; (2) be right handed; (3) tions were presented in blocks, with each block containing only
SLEEP, Vol. 36, No. 9, 2013 1308 Neural Correlates of Working Memory in Insomnia—Drummond et al
one N-back condition (e.g., all 1-backs). The task alternated in model to the data containing parameters for the constant,
a pseudo-random order through four repetitions of each block linear drift, three-motion parameters, and the reference func-
type. Participants responded via a button press whenever they tions convolved with an idealized hemodynamic response
saw a target. For the 0-back condition, the letter “X” was the function to account for the time course of the BOLD signal.31
target. For the 1-, 2-, and 3-back conditions, the target was a Reference functions were vectors representing 1-, 2-, and
letter that is a repeat of the letter one, two, or three letters ago, 3-back blocks, respectively. 0-back blocks served as the
respectively. The task began with direction reminders lasting common baseline.
7.5 sec. Each block began with a 2.5-sec prompt to identify the
condition and then presented 11 letters, three of which were Analyses
targets. Letters were presented in the middle of a projector
screen for 1 sec followed by an asterisk for 1.5 sec. Targets were Sleep, Self-Report, and Cognitive Data
placed randomly throughout the sequence (with the exception Data from sleep diaries were averaged over the 7 days prior
that the initial stimuli, by definition, could not be targets, since to the first night in the sleep laboratory. Sleep diary data were
nothing preceded them). The task lasted 8 min, 7.5 sec. analyzed with a multivariate analysis of variance (MANOVA)
to examine group differences, as the diary variables were

Downloaded from https://academic.oup.com/sleep/article/36/9/1307/2577899 by guest on 17 May 2022

Scanning Procedures expected to correlate with one another. Planned follow-up
Functional MRI scans were acquired at the UCSD Center for a significant omnibus (assessed with Wilks Lambda)
for Functional Magnetic Resonance Imaging using a GE Signa included individual univariate analyses for each variable.
EXCITE 3.0T whole-body imaging system. Sessions were sched- Similar MANOVAs were conducted for self-report question-
uled 12 h after habitual wake time for each patient. Each scan naires, actigraphy, PSG sleep continuity data, and PSG sleep
session contained three primary scans: a field map, an anatomical architecture data. For the cognitive performance data, a group
image of the brain, and a set of functional images acquired during × difficulty analysis of variance (ANOVA) was conducted to
performance of N-back. Anatomical scans used a T1-weighted parallel the functional MRI analyses (see next paragraphs). The
fast spoiled gradient echo (FSPGR) pulse sequence (echo time omnibus was evaluated with the Greenhouse-Geisser correction
(TE) = 4 ms, flip angle = 90°, 1 mm3 resolution). The functional for sphericity violations. Significant interactions were followed
scans were sensitive to the T2*-weighted blood oxygenation level up by testing the effect of difficulty within each group.
dependent (BOLD) signal, and 192 gradient echo, echo planar
images were collected across 32 4-mm axial slices covering the Functional MRI Data
whole brain (repetition time = 2,500 ms, TE = 30 ms, image Two sets of group analyses were conducted. First, a group ×
matrix = 64 × 64, 4 × 4 × 4 mm resolution). difficulty ANOVA was conducted to test whether PIs showed
a different cerebral response to increasing difficulty relative
Postscan Questionnaires to GSs. For any region where the interaction term was signifi-
Immediately after each cognitive task, patients completed a cant, the follow-up consisted of examining the difficulty effect
computerized questionnaire containing the Karolinska Sleepi- for each group. We also examined the main effect of group.
ness Scale (KSS) and a series of 10-point Likert scales asking As a check on the internal validity of the N-back task, we also
about motivation to perform well, ability to concentrate, amount examined the main effect of difficulty in each group separately
of effort required to perform the task, and perceived task diffi- in an effort to ensure our modulation of task load produced
culty. After the scan, patients were taken to a quiet room and the expected modulation of brain activation25 and that areas
asked to provide information on strategies and perceptions showing a group × difficulty interaction generally overlapped
regarding their performance (100-mm VAS rating subjective with those typically modulated by changing load in healthy
sense of accuracy on each N-back condition from zero to 100%). sleepers. Second, a group × performance multiple regression
examined whether PIs used a different set of brain regions,
Sleep Data Processing relative to GSs, to obtain good performance. Based on pilot
Actigraphy was scored using the bedtime and rise times data in PIs, we restricted the analysis to the most difficult (i.e.,
reported by the patients as the boundaries of the rest interval. 3-back) condition.32,33 Significant interactions were followed by
This was done rather than using the automatic scoring algorithm examining performance effects in each group separately. For all
because the automatic selection of a rest interval can miss long functional MRI analyses, we used a search region approach.34-36
sleep latency periods or times after the final awakening that This involves restricting the analysis to a set of brain regions
the individual remains in bed, both common issues in PI. By known to be related to the task of interest. In this case, we chose
matching the actigraphy rest interval to the diaries, we were the set of regions based on the meta-analysis of N-back tasks
able to better examine sleep within the total time in bed reported from functional MRI studies conducted by Owen et al.25 and
by the patients. Because of missing data, the sample size for supplemented those regions with areas identified in more recent
actigraphy included 21 PIs and 24 GSs. PSG data were scored studies using the N-back in either sleep deprivation26,27 or sleep
according to American Academy of Sleep Medicine procedures. apnea samples.28-30 Specific areas included in the search region
included: bilateral inferior, middle, and superior frontal gyri,
Functional MRI Data Processing medial frontal gyrus, the entire cingulate gyrus, bilateral and
Individual functional MRI data were corrected for distor- medial superior parietal lobe, bilateral inferior parietal lobe,
tions using field maps and motion corrected by coregistering bilateral cerebellum, and bilateral thalamus. We identified
to a base image. Time series analyses applied a general linear clusters of activation as areas containing at least 11 contiguous
SLEEP, Vol. 36, No. 9, 2013 1309 Neural Correlates of Working Memory in Insomnia—Drummond et al
voxels (704 mm3), each activated at the P = 0.05 level, resulting analysis, postscan questionnaire data were correlated with
in an overall whole-brain alpha level of 0.001.37 As a post hoc BOLD signal data from the significant clusters found in the
group × difficulty ANOVA using bivariate correlations. Due
to missing data related specifically to subjective perceptions
Table 2—Sleep diaries (average of 1 week) of performance, sample size for those analyses included 20
GS PI Cohen PIs and 24 GSs. We also examined the bivariate correlations
Mean ± SD Mean ± SD P d among the clusters from the group × difficulty ANOVA in a
Total sleep time 471.2 ± 24.3 348.5 ± 41.1 < 0.001 -3.750 post hoc analysis.
Sleep latency 14.8 ± 6.4 44.2 ± 26.6 < 0.001 1.777
Number of awakenings 0.9 ± 0.9 2.2 ± 1.0 < 0.001 1.376 RESULTS
Wake after sleep onset 9.4 ± 10.8 50.0 ± 36.9 < 0.001 1.705
Sleep efficiency 94.8 ± 2.8 79.4 ± 9.2 < 0.001 -2.588 Sleep Data
Caffeine, servings/day 0.56 ± 0.93 0.53 ± 0.64 0.892 -0.039
Alcohol, servings/day 0.21 ± 0.37 0.32 ± 0.43 0.336 0.276 Sleep Diaries
The omnibus MANOVA was significant [F(9,40) = 24.35,

Downloaded from https://academic.oup.com/sleep/article/36/9/1307/2577899 by guest on 17 May 2022

P < 0.001, partial eta2 = 0.846]. In addition, all diary vari-
P values are univariate analyses not corrected for multiple comparisons; bolded
data = significant; P < 0.05. Cohen d effects sizes are positive when PI > GS ables showed a significant difference between groups in
and negative when PI < GS. GS, good sleeper; PI, primary insomnia; SD, the expected direction. All diary measures were within the
standard deviation. clinical range for the PIs. This information is presented in
Table 2.

Actigraphy
Table 3—Actigraphy (average of 1 week)
The omnibus MANOVA was significant [F(5,39) = 6.59,
GS PI Cohen P < 0.001, partial eta2 = 0.458]. In addition, all actigraphy
Mean ± SD Mean ± SD P d
variables showed a significant difference between groups in
Time in bed, min 488.1 ± 31.2 441.8 ± 53.3 0.001 -1.094
the expected direction, except wake after sleep onset (which
Total sleep time, min 425.5 ± 35.2 363.8 ± 41.8 0.000 -1.602
was nonsignificant) and time in bed (PI < GS). Generally,
Sleep latency, min 14.6 ± 8.7 25.6 ± 15.5 0.005 0.905 actigraphically measured sleep was not as impaired in PIs
Wake after sleep onset, min 34.4 ± 14.3 38.3 ± 16.1 0.384 0.262 as diary-measured sleep, though all variables except sleep
Sleep efficiency, % 86.8 ± 0.1 82.1 ± 0.04 0.002 -1.011 latency were still in the clinical range for PIs. This informa-
tion is presented in Table 3.
P values are univariate analyses not corrected for multiple comparisons; bolded
data = significant; P < 0.05. Cohen d effects sizes are positive when PI > GS Polysomnography
and negative when PI < GS. GS, good sleeper; PI, primary insomnia; SD, The omnibus MANOVA for sleep architecture vari-
standard deviation.
ables was not significant [F(7,42) = 1.53, P = 0.183, partial
eta2 = 0.203]. The omnibus MANOVA for sleep continuity
variables was also not significant [F(5,44) = 1.77, P = 0.140,
Table 4—Polysomnography (1 night) partial eta2 = 0.167]. This information is presented in Table 4.
GS PI Cohen
Mean ± SD Mean ± SD P d Self-Report Measures
Total sleep time, min 422.9 ± 30.5 390.8 ± 53.7 0.013 -0.762 The omnibus MANOVA was significant [F(10,39) = 67.02,
Sleep latency, min 18.1 ± 20.6 18.2 ± 11.4 0.987 0.005 P < 0.001, partial eta2 = 0.945]. All self-report measures
REM latency, min 87.7 ± 50.8 76.4 ± 24.0 0.320 -0.302 showed a significant difference between groups, except
Wake after sleep onset, min 30.3 ± 25.4 32.8 ± 30.8 0.748 0.092 the Beck Depression Inventory and the physical fatigue
Sleep efficiency, % 89.8 ± 6.4 88.6 ± 7.9 0.530 -0.18 subscale from the Multidimensional Fatigue Inventory
Stage 1 min 24.1 ± 11.6 20.5 ± 7.3 0.194 -0.382 (MFI). PIs reported insomnia severity in the moderate range
Stage 2 min 236.0 ± 41.6 207.8 ± 43.6 0.024 -0.66 on the Insomnia Severity Index, clinically impaired sleep
Stage 3 min 58.3 ± 34.3 67.1 ± 36.9 0.385 0.248 on the Pittsburgh Sleep Quality Index, greater sleepiness
REM min 104.6 ± 20.5 95.4 ± 33.0 0.243 -0.344 (although not above the clinical cutoff) than GSs, more
Stage 1% 5.8 ± 2.9 5.3 ± 2.1 0.537 -0.178
evening tendencies on the Horne-Ostberg Morningness
Eveningness Questionnaire (though both groups were in
Stage 2% 55.9 ± 9.3 53.2 ± 9.5 0.321 -0.283
the “neither morningness nor eveningness” category), and
Stage 3% 13.6 ± 7.9 17.4 ± 10.3 0.158 0.409
greater fatigue on all MFI subscales other than physical.
REM% 24.7 ± 4.4 24.1 ± 6.8 0.712 -0.107
This information is presented in Table 5.
P values are univariate analyses not corrected for multiple comparisons; bolded
Cognitive Task Data
data = significant; P < 0.05. Cohen d effects sizes are positive when PI > GS
and negative when PI < GS. GS, good sleeper; PI, primary insomnia; REM,
The group × difficulty interaction for accuracy on N-back
rapid eye movement; SD, standard deviation. was nonsignificant [F(2,96) = 0.389, P = 0.608, partial
eta2 = 0.008] and there was also no main effect of Group
SLEEP, Vol. 36, No. 9, 2013 1310 Neural Correlates of Working Memory in Insomnia—Drummond et al
[F(1,48) = 1.50, P = 0.227, partial eta2 = 0.030]. Similar results lower levels of motivation, concentration, and effort required
were found for response time interaction [F(2,96) = 1.19, compared with GSs. There was no difference for subjective
P = 0.309, partial eta2 = 0.024] and main effect of group task difficulty (Table 9). Self-reported concentration during
[F(1,48) = 0.278, P = 0.600, partial eta2 = 0.006]. 0-back was task performance correlated significantly with BOLD signal
not evaluated as part of the previously mentioned MANOVAs in pregenual cingulate for 1-back (r = -0.469, P = 0.018) and
in an effort to keep these analyses parallel with the imaging 2-back (r = -0.438, P = 0.028) for PIs, and for 1-back (r = -0.417,
analyses. When examining 0-back separately, there was no P = 0.038) in GSs. For PIs, self-report sleepiness during task
group effect for either accuracy or response time. This informa- performance also correlated significantly with BOLD signal in
tion is presented in Table 6. pregenual cingulate for 1-back (r = 0.552, P = 0.004), 2-back
(r = 0.410, P = 0.042), and 3-back (r = 0.415, P = 0.039). For
Functional MRI Data GSs, self-report sleepiness did not correlate with any brain
The main effect of modulating task difficulty for each group region. The omnibus MANOVA for subjective perceptions of
is shown in Figure 1. The group × difficulty interaction produced performance was not significant [F(4,37) = 1.74, P = 0.162,
five significant clusters (Table 7, Figure 2). In two right middle partial eta2 = 0.158]; information presented in Table 9. Self-
frontal gyrus clusters (MFG, BA46 and BA9), GSs showed reported performance on 1-back correlated significantly with

Downloaded from https://academic.oup.com/sleep/article/36/9/1307/2577899 by guest on 17 May 2022

increasing activation with increasing task difficulty, suggesting activation during 1-back within each of the default mode
greater reliance on these working memory regions as the task
became harder. In contrast, for three midline regions (orbital
frontal gyus BA11, pregenual cingulate gyrus BA32, and poste- A B
rior cingulate BA23), GSs showed greater deactivation as the 2L 2L
task became harder. Unlike the GSs, PIs showed no modulation Cohen d
of these regions with increasing task difficulty. Several brain 1.5
regions associated with the frontoparietal working memory
network, premotor areas, visual processing regions, and the
thalamus showed a significant main effect of group (Table 7,
Figure 3). In all of these, GSs showed greater activation than
PIs. For more than half of these regions, activation was not
significantly different from zero in PIs, suggesting areas typi- 48R 48R
-1.5
cally engaged during task performance were not engaged by PIs.
The group × performance analysis revealed four brain
regions showing a significant interaction, one of which did
not have a significant relationship with performance in either
group on follow-up and will not be discussed further. The other
three regions (right inferior frontal gyrus (IFG) BA44/45, right
inferior parietal lobe (IPL) BA40, and right cerebellum) each
Figure 1—Brain regions showing a significant main effect of task difficulty
showed a positive relationship between cerebral activation within each group. (A) Good sleepers. (B) Primary insomnia.
and 3-back performance in GSs and no significant relation-
ship with performance in PIs. Five clusters showed a
significant main effect of performance in this analysis,
with increasing activation associated with better perfor- Table 5—Self-report measures
mance, regardless of group membership (Table 8). GS PI Cohen
Intake Measures Mean ± SD Mean ± SD P d
Intercluster Correlations Insomnia Severity Index 1.1 ± 1.6 16.8 ± 3.0 < 0.001 6.725
Within the PIs, there were several significant posi- Pittsburgh Sleep Quality Index 1.7 ± 1.4 11.4 ± 2.2 < 0.001 5.362
tive correlations, and no negative correlations, between Epworth Sleepiness Scale 4.2 ± 3.3 7.3 ± 4.3 0.006 0.821
the two task-related clusters and the three default mode Horne-Ostberg MEQ 56.0 ± 7.7 48.6 ± 8.5 0.002 -0.904
clusters from the group × difficulty analysis: (1) middle Beck Depression Inventory 1.9 ± 3.7 3.6 ± 3.6 0.098 0.477
frontal gyrus BA9 with orbital frontal gyrus for 1-back MFI: general fatigue 6.0 ± 2.0 12.2 ± 2.2 < 0.001 2.97
(r = 0.550, P = 0.004); (2) middle frontal gyrus BA9 with
MFI: physical fatigue 6.4 ± 2.0 8.0 ± 3.6 0.065 0.557
posterior cingulate for 2-back (r = 0.537, P = 0.006);
MFI: reduced activity 6.4 ± 2.7 8.4 ± 3.2 0.021 0.675
(3) middle frontal gyrus BA46 with posterior cingu-
MFI: reduced motivation 5.8 ± 1.7 9.3 ± 3.3 < 0.001 1.394
late for 2-back (r = 0.410, P = 0.042); and (4) middle
MFI: mental fatigue 6.8 ± 3.34 11.6 ± 3.5 < 0.001 1.394
frontal gyrus BA46 with posterior cingulate (r = 0.463,
P = 0.020) for 3-back.
P values are univariate analyses not corrected for multiple comparisons; bolded
data = significant P < 0.05. Cohen d effects sizes are positive when PI > GS and
Postscan Questionnaires
negative when PI < GS. GS, good sleeper; MEQ, Morningness Eveningness
The omnibus MANOVA for self-report measures Questionnaire; MFI, Multidimensional Fatigue Inventory; PI, primary insomnia; SD,
was significant [F(5,44) = 6.212, P < 0.001, partial standard deviation.
eta2 = 0.414]. PIs reported greater levels of sleepiness and
SLEEP, Vol. 36, No. 9, 2013 1311 Neural Correlates of Working Memory in Insomnia—Drummond et al
clusters in the group × difficulty analysis: orbital frontal gryus DISCUSSION
(r = -0.490, P = 0.028), posterior cingulate 1-back (r = -0.501, This is the largest study to date examining cerebral activa-
P = 0.024), pregenual cingulate 1-back (r = -0.538, P = 0.014). tion with functional MRI during cognitive performance in PIs,
Self-reported performance on the 3-back also correlated with relative to well-matched GSs, and the first to characterize func-
activation during 3-back in pregenual cingulate (r = -0.634, tional MRI differences in working memory in PI. We examined
P = 0.003). 25 young to middle-aged PIs and 25 GSs, matched one-to-one
for age, sex, and level of education. Our overall
aim was to compare the two groups on both
Table 6—Cognitive performance measures performance and cerebral activation during a
GS PI Cohen N-back working memory task with multiple
Mean ± SD Mean ± SD P d levels of difficulty (allowing evaluation of
d’ 0-back 5.0 ± 1.0 4.7 ± 1.2 0.292 -0.273 whether PIs responded differently to increasing
d’ 1-back 4.4 ± 1.3 4.1 ± 1.5 0.395 -0.214 cognitive loads). Furthermore, we analyzed the
d’ 2-back 4.0 ± 1.4 3.6 ± 1.3 0.221 -0.296 relationship between performance and cerebral
d’ 3-back 3.2 ± 1.6 2.6 ± 1.3 0.128 -0.414 responses to examine if the better-performing PIs

Downloaded from https://academic.oup.com/sleep/article/36/9/1307/2577899 by guest on 17 May 2022

Response time 0-back, ms 517.3 ± 115.5 553.0 ± 244.7 0.514 0.198 used the same brain regions as better-performing
Response time 1-back, ms 583.8 ± 212.4 581.2 ± 221.7 0.967 -0.012 GSs, or whether they would require recruitment
Response time 2-back, ms 774.2 ± 255.3 685.1 ± 294.7 0.259 -0.324 of additional or different brain regions to achieve
Response time 3-back, ms 796.7 ± 210.3 800.5 ± 232.3 0.952 0.017 optimal performance. Our hypotheses were that
Response time weighted mean, ms 661.338 ± 142.202 645.2 ± 206.2 0.749 -0.093 PIs would show intact performance relative to
GSs, and this would be associated with increased
P values are univariate analyses not corrected for multiple comparisons. Cohen d effects cerebral responses representing compensatory
sizes are positive when PI > GS and negative when PI < GS. GS, good sleeper; PI, primary activation. These behavioral hypotheses were
insomnia; SD, standard deviation. supported, as PIs showed equivalent perfor-
mance to GSs on all measures and all levels of
difficulty for the N-back. The functional MRI
hypotheses, however, were not supported.
Table 7—Significant brain regions in N-back group × difficulty analysis
Indeed, all of the brain regions showing a
Size Centroid Effect significant main effect of group during N-back
Anatomical Location BA (mm3) coordinates size
performance showed diminished activation in
Group × difficulty interaction
R Middle frontal gyrus 46 1,024 45.3R, 39.4A, 20.9S 0.423
PIs, compared with GSs (Table 7, Figure 3).
R Middle frontal gyrus 9 1,024 47.2R, 28.6A, 34.4S 0.430 These regions included widespread areas within
L Orbital frontal gyrus 11 1,024 2.1L, 38.9A, 16.6I 0.429 the frontoparietal working memory network,
L Pregenual cingulate 32 3,456 4.3L, 33.2A, 17.4S 0.386 motor and visual processing regions, thalamus
B Posterior cingulate 23 2,688 3.7R, 52.8P, 23.8S 0.461 (nuclei with connections to cortical working
Group main effect memory regions), and cerebellum. These are the
L Frontal pole 10 1,088 23.8L, 64.5A, 5.5S 0.763 brain regions most commonly reported in studies
R Middle frontal gyrus 46 1,088 44.9R, 42.1A, 18.4S 0.852 of verbal N-back tasks.25 Examining each group
L Middle frontal gyrus 9 1,152 36.4L, 30.5A, 31.6S 0.706 separately, GSs showed significantly positive
R Middle frontal gyrus 9 960 46.4L, 31.1A, 30.5S 0.782 activation within all but two areas showing a
L Inferior frontal gyrus 47 1,152 42.8L, 23.2A, 8.3I 0.839 group main effect. In contrast, PIs only showed
R Inferior frontal gyrus 47 832 42.1R, 21.1A, 5.6I 0.853 significantly positive activation in four of these
L Cingulate/medial frontal gyrus 24/6 5,824 1.2L, 3.1A, 47.8S 0.838 areas (bilateral premotor cortex, bilateral MFG
R Premotor area 6 1,984 27.6R, 1.5P, 57.5S 0.915
BA9). The other areas either showed no signifi-
L Premotor area 6 1,536 27.0L, 1.6P, 58.2S 1.042
L Suppl motor area 6 896 21.1L, 15.0A, 54.8S 0.823
cant activation (10 of 18 areas) or a significant
B Thalamus L pulvinar 4,672 3.7L, 21.0P, 7.2S 0.911 negative activation (4 of 18 areas). Thus, it
B DMN appears PIs did not engage many regions of the
L Inferior parietal lobe 40 2,752 53.4L, 33.4P, 29.0S 0.768 brain typically used during performance of this
R Inferior parietal lobe 40 1,280 64.7R, 33.0P, 28.7S 0.806 working memory task, and did not activate any
R Superior parietal lobe 7 2,048 30R, 47.8P, 54.7S 0.919 other areas to compensate for not activating the
L Posterior cingulate 29/30 2,176 2.5L, 55.4P, 8.5S 0.792 typical network.
B Precuneus 7 8,320 2.9L, 58.9P, 51.8S 1.128 In addition to not activating regions typically
L Cuneus 31 896 16.7L, 71.7P, 30.8S 0.794
activated during performance of the N-back
R Cerebellum 768 14.2R, 80.5P, 19.7I 0.663
task, PIs did not show the modulation of acti-
vation with increasing task difficulty one would
All clusters survived the cluster thresholding method described in the text. Whole-brain
expect and that GSs showed here (Figures 1 and
alpha = 0.001. Effect size = mean Cohen d within the cluster. B, bilateral (i.e., the cluster
crossed the midline); DMN, dorsal medial nucleus of the thalamus; L, left hemisphere; r, right
2). Specifically, GSs showed increased cerebral
hemisphere; BA, Brodmann’s Area. responses in right dorsolateral prefrontal cortex
(BA46 and BA9) and classic N-back working
SLEEP, Vol. 36, No. 9, 2013 1312 Neural Correlates of Working Memory in Insomnia—Drummond et al
 A B C
2L 1-back 2-back 3-back

Cingulate Cortex BA32 R MFG BA46

0 16

Bold signal (a.u.)

Bold signal (a.u.)

Cohen d -1 14
1.5 -2 12
10
-3 8
-4 6
-5 4
-6 2
GS PI 0
48R GS PI

Bilateral Posterior Cingulate Cortex R MFG BA9

0.0 14
-1
Bold signal (a.u.)

Bold signal (a.u.)

Downloaded from https://academic.oup.com/sleep/article/36/9/1307/2577899 by guest on 17 May 2022

12
-3 10
-5 8
6
-7
4
-9 2
-11 0
GS PI GS PI

L Ventromedial PFC
-1
Bold signal (a.u.)

-3
-5
-7
-9
-11
-13
-15
GS PI
Figure 2—Brain regions showing significant group × difficulty interaction for N-back. (A) Functional magnetic resonance images of five clusters showing a
significant interaction; 1. L OFG BA11; 2. Pregenual cingulate gyrus BA32; 3. B Posterior cingulate cortex BA23; 4. R MFG BA46; 5. R MFG BA9. (B and
C) Parameter estimates for clusters. (B) Regions with significantly greater deactivation in GSs as the task became harder with no significant task-related
modulation in PIs. (C) Regions with significantly increasing activation in GSs as the task became harder with no significant task-related modulation in PIs.
GS, good sleeper; PI, primary insomnia. Error bars indicate standard deviation.

memory areas, as the task became more difficult. In contrast, modulate this network, showing persistent (rather than reduced)
PIs did not show a modulation with task difficulty in these activity throughout these default mode regions during task perfor-
clusters. Similarly, GSs showed greater deactivation in the mance. Impaired modulation of default mode network may be
pregenual cingulate cortex, ventromedial prefrontal cortex, and caused by either abnormal disengagement of the network during
posterior cingulate as the task became increasingly difficult. In the baseline condition of the task or by abnormal engagement of
contrast, PIs again failed to show this modulation of activation the network during the active conditions.
with task difficulty. Importantly, each of these interaction clus- An example of the first explanation for impaired default
ters overlapped entirely with regions showing a main effect of mode modulation in PI may be found in our recent study (Czisch
difficulty in GS (Figure 1). Thus, these interactions occurred in et al.39) comparing a group of “sleep-deprived but alert” patients
brain regions normally modulated by task difficulty. with those same patients after a normal night of sleep, during
Of particular interest are the three clusters showing greater deac- performance of an auditory oddball task. The sleep-deprived
tivation in GS with increasing task difficulty. These brain regions but alert patients also showed impaired modulation of default
are among those often reported as part of the default mode.38 The mode regions, including areas overlapping those reported here,
default mode is a set of brain regions thought to be generally acti- with intact performance. Data suggested the findings reflected
vated when the brain is not otherwise engaged in task-focused increased effort needed to focus on the task during the base-
behaviors. As attention is diverted to task- or goal-directed behav- line condition. This need for increased cognitive resources after
iors, disengagement from default mode regions is necessary for sleep deprivation therefore produced a tonic disengagement of
optimal goal-directed performance. As would be expected, GSs the default mode network. That, in turn, affected the relative
showed greater disengagement (reflected in greater levels of BOLD signal in such a way as to reduce the normal differences
negative activation) of these regions as the N-back task became in default mode activation between baseline and active condi-
harder and greater attention and working memory resources were tions.39 If PIs also required greater sustained cognitive resources
required. PIs, however, demonstrated an impoverished ability to to perform the baseline condition of the N-back task, it would
SLEEP, Vol. 36, No. 9, 2013 1313 Neural Correlates of Working Memory in Insomnia—Drummond et al
also lead to tonic disengagement of default mode regions during with regions showing a group × performance interaction or
baseline and thus less room to exhibit even greater disengage- main-effect performance. This observation is consistent with
ment as the task got harder. That would explain the lack of the idea that less relative disengagement of default mode
modulation by task difficulty of default mode network regions during active conditions (perhaps due to tonic disengagement
for the PIs (Figure 1B). Czisch et al.39 further reported compen- during baseline) is associated with greater task-related activa-
satory increases in activation in task-related regions associated tion (perhaps representing compensatory activation in those
with intact performance in their patients Here, activation was subjects), although the post hoc nature of the analysis makes
positively correlated between default mode and task-related such a conclusion speculative at this point.
regions in PIs, however, unlike in Czisch et al., this tradeoff Rather than abnormally high levels of default mode disen-
was not associated with better performance in PIs and none of gagement during the baseline, an alternative hypothesis may
these regions showing a group × difficulty effect overlapped be a lack of disengagement during the harder conditions. We
have previously reported, in the context of a sustained attention
task placing a premium on speed, acute total sleep deprivation
8I 6S reduces the ability to disengage default mode areas during task
performance,40 again effectively reducing the normal differ-

Downloaded from https://academic.oup.com/sleep/article/36/9/1307/2577899 by guest on 17 May 2022

Cohen d
ences in default mode activation between baseline and active
1.5
conditions. This hypothesis, as opposed to the earlier one, is
supported by the finding showing decreased ability to concen-
trate and increased sleepiness during the task correlated with
greater activation of default mode regions. These subjective
responses to default mode intrusions during task performance
may, in turn, manifest as the subjective cognitive complaints
so common in PI. Indeed, subjectively worse performance
20S 34S also correlated with activation of default mode regions, further
supporting this hypothesis.
Finally, we must entertain a more purely physiological expla-
nation for the PI-GS activation differences. In particular, if there
0.0 are group differences in resting cerebral blood flow (CBF),
they could influence the BOLD signal.41-43 Although no studies
have examined resting CBF in PIs, Nofzinger et al.44 reported
increased glucose metabolism in waking PIs, Huang et al.21
reported resting state functional connectivity data consistent
with cerebral hyperarousal in PI, and van der Werf et al.45
Figure 3—Brain regions showing significant main effect of group for reported increased cortical excitability in PI measured with
N-back. 1. L IFG BA47; 2. R IFG BA47; 3. L Frontal Pole; 4. B Thalamus; transcranial magnetic stimulation. If such preliminary findings
5. R MFG BA46; 6. L IPL BA40; 7. LMFG BA9; 8. R MFG BA9; of hyperarousal extend to CBF in PI, and if CBF was increased
9. R IPL BA40. All regions showed greater activation in response to throughout much or all of the brain in PI, one could potentially
N-back demands in good sleepers, relative to primary insomnia. find a reduced magnitude in the relative BOLD signal (i.e., both
decreased activation of task-related areas and
decreased deactivation of default mode network)
Table 8—Significant brain regions in N-back group × performance analysis without necessarily affecting performance.
Future studies would benefit from a direct
Size Centroid Effect
Anatomical Location BA (mm3) coordinates size examination of resting-state CBF and/or oxygen
Group × performance interaction metabolism in PI in direct relation to BOLD
R Inferior frontal gyrus 44/45 704 50.2R, 5.7A, 26.6S 0.132 signal responses during task performance.
R Inferior parietal lobe 40 1,088 28.9R, 10.1P, 20.7S 0.097
R Inferior parietal lobe 40 704 30.0L, 14.8P, 19.9S 0.117 Limitations and Future Directions
R Cerebellum nodule 960 64.2R, 27.4P, 30.0S 0.176 In summary, a few limitations of this study
Performance main effect should be noted. First, we studied only PI. Given
B Frontal pole 10 10,688 7.3L, 59.3A, 5.4I 1.015 the prevalence of insomnia in psychiatric disor-
L Middle frontal gyrus 9 2,880 40.9L, 14.7A, 33.4S 0.942 ders, it would be potentially useful to examine
L Posterior parietal cortex 39 1,152 29.9L, 74.2P, 34.7S 0.874 how much variance is accounted for in future
R Cerebellum tuber/declive 2,496 43.8R, 65.9P, 229.I 0.912
neuroimaging studies in psychiatric popula-
L Cerebellum declive 2,304 15.5L, 69.3P, 24.2I 0.946
tions by the psychiatric condition, relative to by
insomnia. For example, Neylan et al.23 reported
All clusters survived the cluster thresholding method described in the text. Whole-brain
insomnia severity contributed more unique vari-
alpha = 0.001. Effect size = mean eta-squared within the cluster for group × performance
interaction and mean Cohen d within the cluster for performance main effect. B, bilateral (i.e.,
ance to hippocampal volume measurements
the cluster crossed the midline); L, left hemisphere; R, right hemisphere; BA, Brodmann’s Area. in veterans with posttraumatic stress disorder
than did Clinician Administered PTSD Scale
SLEEP, Vol. 36, No. 9, 2013 1314 Neural Correlates of Working Memory in Insomnia—Drummond et al
(CAPS) scores. However, it is unclear if that
Table 9—Postscan questionnaires
would be the case for any functional neuro-
imaging measure or even another structural Control PI Cohen
imaging measure. Second, we did not recruit Scan Session Questionnaires Mean ± SD Mean ± SD P d
for specific insomnia subtypes, and thus Sleepiness (1-10) 3.2 ± 2.3 6.3 ± 1.9 < 0.001 1.424
it is unclear if our findings relate to early, Concentration (1-10) 8.2 ± 1.8 6.2 ± 2.6 0.002 -0.938
middle, and/or late insomnia. Most patients Task difficulty (1-10) 5.6 ± 2.5 5.2 ± 2.2 0.557 -0.168
reported both early and middle insomnia, Motivation (1-10) 8.9 ± 1.3 7.1 ± 2.5 0.003 -0.921
making it impossible to explore this question. Effort required (1-10) 8.9 ± 1.6 7.4 ± 2.4 0.015 -0.73
Although this finding may actually increase Subjective performance 0-back (0-100) 98.9 ± 2.4 95.3 ± 6.9 0.042 -0.772
external validity of the study, future research Subjective performance 1-back (0-100) 87.7 ± 18.6 89.6 ± 12.9 0.710 0.118
may wish to more specifically examine Subjective performance 2-back (0-100) 79.4 ± 23.3 77.1 ± 18.6 0.726 -0.11
insomnia subtypes. Third, our patients were, Subjective performance 3-back (0-100) 54.0 ± 25.1 55.3 ± 30.2 0.877 0.047
on average, in the moderate severity range
according to the ISI. Only three patients P values are univariate analyses not corrected for multiple comparisons; bolded data = significant;

Downloaded from https://academic.oup.com/sleep/article/36/9/1307/2577899 by guest on 17 May 2022

scored in the severe insomnia range. Thus, it P < 0.05. Cohen d effects sizes are positive when PI > GS and negative when PI < GS. GS, good
is not clear if these results would generalize sleeper; PI, primary insomnia; SD, standard deviation.
to those with more severe forms of insomnia.

CONCLUSIONS 3. Fernandez-Mendoza J, Vgontzas AN, Bixler EO, et al. Clinical and

This study suggests PIs have a reduced ability to engage some polysomnographic predictors of the natural history of poor sleep in the
general population. Sleep 2012;35:689-97.
task-appropriate brain regions and reduced ability to modulate 4. Ancoli-Israel S, Roth T. Characteristics of insomnia in the United
task-irrelevant (i.e., default mode) brain regions during working States: results of the 1991 National Sleep Foundation Survey. I. Sleep
memory performance. If either of our potential explanations for 1999;22:S347-53.
this is correct (PIs require greater attention and/or effort during 5. Leger D, Morin CM, Uchiyama M, Hakimi Z, Cure S, Walsh JK. Chronic
insomnia, quality-of-life, and utility scores: comparison with good sleepers
baseline or PIs are unable to disengage from default mode in a cross-sectional international survey. Sleep Med 2011;13:43-51.
during active conditions), these phenomena may be experienced 6. Zammit GK, Weiner J, Damato N, Sillup GP, McMillan CA. Quality of
by PIs as increased cognitive difficulties. Our post hoc analyses life in people with insomnia. Sleep 1999;22:S379-85.
support this hypothesis. Moreover, both of these explanations 7. Sarsour K, Kalsekar A, Swindle R, Foley K, Walsh JK. The association
between insomnia severity and healthcare and productivity costs in a
suggest a dysfunction in the attention system in PI that has not health plan sample. Sleep 2011;34:443-50.
been previously identified. Future studies should more directly 8. Leger D, Guilleminault C, Bader G, Levy E, Paillard M. Medical and
address hypotheses arising from these data. socio-professional impact of insomnia. Sleep 2002;25:625-9.
9. Chien KL, Chen PC, Hsu HC, et al. Habitual sleep duration and insomnia
and the risk of cardiovascular events and all-cause death: report from a
ACKNOWLEDGMENTS community-based cohort. Sleep 2010;33:177-84.
The authors thank Lesley Wetherell for her assistance with 10. Wright KM, Britt TW, Bliese PD, Adler AB, Picchioni D, Moore D.
processing the functional MRI data. The VA San Diego Health- Insomnia as predictor versus outcome of PTSD and depression among
care System’s Center for Excellence in Stress and Mental Iraq combat veterans. J Clin Psychol 2011;67:1240-58.
11. Perlis ML, Giles DE, Buysse DJ, Tu X, Kupfer DJ. Self-reported sleep
Health provided infrastructure support for this project. Work disturbance as a prodromal symptom in recurrent depression. J Affect
was performed at the VA San Diego Healthcare System and Disord 1997;42:209-12.
University of California San Diego, San Diego, CA. 12. Kessler RC, Berglund PA, Coulouvrat C, et al. Insomnia, comorbidity,
and risk of injury among insured Americans: results from the America
Insomnia Survey. Sleep 2012;35:825-34.
DISCLOSURE STATEMENT 13. Bramoweth AD, Taylor DJ. Chronic insomnia and health care utilization
This study was funded by an investigator initiated grant from in young adults. Behav Sleep Med 2012;10:106-21.
Actelion Pharmaceuticals, Ltd. Dr. Drummond declares grant 14. Rosekind MR, Gregory KB. Insomnia risks and costs: health, safety, and
funding (for this study) and consulting fees from Actelion Phar- quality of life. Am J Manag Care 2010;16:617-26.
15. Orff HJ, Drummond SP, Nowakowski S, Perils ML. Discrepancy
maceuticals, Ltd. Dr. Walker declares grant funding (for this between subjective symptomatology and objective neuropsychological
study) and consulting fees from Actelion Pharmaceuticals, Ltd. performance in insomnia. Sleep 2007;30:1205-11.
The other authors have indicated no financial conflicts of interest. 16. Vignola A, Lamoureux C, Bastien CH, Morin CM. Effects of chronic
insomnia and use of benzodiazepines on daytime performance in older
adults. J Gerontol B Psychol Sci Soc Sci 2000;55:P54-62.
REFERENCES 17. Hauri PJ. Cognitive deficits in insomnia patients. Acta Neurologica
1. Ford DE, Kamerow DB. Epidemiologic study of sleep disturbances
Belgica 1997;97:113-7.
and psychiatric disorders. An opportunity for prevention? JAMA
18. Edinger JD, Means MK, Carney CE, Krystal AD. Psychomotor
1989;262:1479-84.
performance deficits and their relation to prior nights’ sleep among
2. Roth T, Coulouvrat C, Hajak G, et al. Prevalence and perceived health
individuals with primary insomnia. Sleep 2008;31:599-607.
associated with insomnia based on DSM-IV-TR; International Statistical
19. Fortier-Brochu E, Beaulieu-Bonneau S, Ivers H, Morin CM. Insomnia
Classification of Diseases and Related Health Problems, Tenth Revision;
and daytime cognitive performance: a meta-analysis. Sleep Med Rev
and Research Diagnostic Criteria/International Classification of Sleep
2011;16:83-94.
Disorders, Second Edition criteria: results from the America Insomnia
20. Altena E, Van Der Werf YD, Sanz-Arigita EJ, et al. Prefrontal
Survey. Biol Psychiatry 2011;69:592-600.
hypoactivation and recovery in insomnia. Sleep 2008;31:1271-6.

SLEEP, Vol. 36, No. 9, 2013 1315 Neural Correlates of Working Memory in Insomnia—Drummond et al
21. Huang Z, Liang P, Jia X, et al. Abnormal amygdala connectivity in 35. Eyler Zorrilla LT, Jeste DV, Paulus M, Brown GG. Functional abnormalities
patients with primary insomnia: evidence from resting state fMRI. Eur J of medial temporal cortex during novel picture learning among patients
Radiol 2011;81:1288-95. with chronic schizophrenia. Schizophr Res 2003;59:187-98.
22. Riemann D, Voderholzer U, Spiegelhalder K, et al. Chronic insomnia and 36. Jonelis MB, Drummond SP, Salamat JS, McKenna BS, Ancoli-Israel
MRI-measured hippocampal volumes: a pilot study. Sleep 2007;30:955-8. S, Bondi MW. Age-related influences of prior sleep on brain activation
23. Neylan TC, Mueller SG, Wang Z, et al. Insomnia severity is associated during verbal encoding. Front Neurol 2012;3:49.
with a decreased volume of the CA3/dentate gyrus hippocampal subfield. 37. Forman SD, Cohen JD, Fitzgerald M, Eddy WF, Mintun MA, Noll DC.
Biol Psychiatry 2010;68:494-6. Improved assessment of significant activation in functional magnetic
24. Altena E, Vrenken H, Van Der Werf YD, van den Heuvel OA, Van Someren resonance imaging (fMRI): use of a cluster-size threshold. Magn Reson
EJ. Reduced orbitofrontal and parietal gray matter in chronic insomnia: a Med 1995;33:636-47.
voxel-based morphometric study. Biol Psychiatry 2010;67:182-5. 38. Raichle ME, MacLeod AM, Snyder AZ, Powers WJ, Gusnard DA,
25. Owen AM, McMillian KM, Laird AR, Bullmore E. N-back working Shulman GL. A default mode of brain function. Proc Natl Acad Sci U S
memory paradigm: a meta-analysis of normative functional neuroimaging A 2001;98:676-82.
studies. Hum Brain Map 2005;25:46-59. 39. Czisch M, Wehrle R, Harsay H, et al. On the need of objective vigilance
26. Choo WC, Lee WW, Venkatraman V, Sheu FS, Chee MW. Dissociation monitoring: effects of sleep loss on target detection and task-negative
of cortical regions modulated by both working memory load and sleep activity using combined EEG/fMRI. Front Neurol 2012;3:1-12.
deprivation and by sleep deprivation alone. Neuroimage 2005;25:579-87. 40. Drummond SPA, Bischoff-Grethe A, Dinges DF, Ayalon L, Mednick SC,
27. Lythe KE, Williams SC, Anderson C, Libri V, Mehta MA. Frontal and Meloy MJ. The neural basis of the psychomotor vigilance task. Sleep

Downloaded from https://academic.oup.com/sleep/article/36/9/1307/2577899 by guest on 17 May 2022

parietal activity after sleep deprivation is dependent on task difficulty and 2005;28:1059-68.
can be predicted by the fMRI response after normal sleep. Behav Brain 41. Bandettini PA, Wong EC, Binder JR, et al. Functional MR imaging using
Res 2012;233:62-70. BOLD the approach: dynamic characteristics and data analysis methods.
28. Aloia MS, Sweet LH, Jerskey BA, Zimmerman M, Arnedt JT, Millman In: Le Bihan D, ed. Diffusion and perfusion magnetic resonance imaging.
RP. Treatment effects on brain activity during a working memory task in New York: Raven Press, Ltd., 1995:335-49.
obstructive sleep apnea. J Sleep Res 2009:404-10. 42. Ogawa S, Tank DW, Menon R, et al. Intrinsic signal changes accompanying
29. Prilipko O, Huynh N, Schwartz S, et al. Task positive and default mode sensory stimulation: Functional brain mapping with magnetic resonance
networks during a parametric working memory task in obstructive sleep imaging. Proc Natl Acad Sci U S A 1992;89:5951-5.
apnea patients and healthy controls. Sleep 2011;34:293-301A. 43. Brown GG, Eyler Zorrilla LT, Georgy B, Kindermann SS, Wong EC,
30. Sweet LH, Jerskey BA, Aloia MS. Default network response to a working Buxton RB. BOLD and perfusion response to finger-thumb apposition
memory challenge after withdrawal of continuous positive airway after acetazolamide administration: differential relationship to global
pressure treatment for obstructive sleep apnea. Brain Imaging Behav perfusion. J Cereb Blood Flow Metab2003;23:829-37.
2010;4:155-63. 44. Nofzinger E, Buysse DJ, Germain A, Price JC, Miewald JM, Kupfer DJ.
31. Cohen MS. Parametric analysis of fMRI data using linear systems Functional neuroimaging evidence for hyperarousal in insomnia. Am J
methods. Neuroimage 1997;6:93-103. Psychiatry 2004;161:2126-9.
32. Orff HJ, Almklov E, Drummond SPA. Association between subjective 45. van der Werf YD, Altena E, van Dijk KD, et al. Is disturbed intracortical
daytime complaint and BOLD activation in patients with primary excitability a stable trait of chronic insomnia? A study using transcranial
insomnia. Sleep 2010;33:A218. magnetic stimulation before and after multimodal sleep therapy. Biol
33. Anderson D, Almklov E, Orff HJ, Drummond SPA. Insomnia severity Psychiatry 2010;68:950-5.
level impacts cerebral activation during working memory. Sleep
2011;34:A190-1.
34. Ayalon L, Ancoli-Israel S, Drummond SP. Obstructive sleep apnea
and age: a double insult to brain function? Am J Respir Crit Care Med
2010;182:413-9.

SLEEP, Vol. 36, No. 9, 2013 1316 Neural Correlates of Working Memory in Insomnia—Drummond et al