Current Biology Vol 21 No 3

R116

resting length of w50 nm, consistent super-resolution acquisition times 5. Zaidel-Bar, R., Itzkovitz, S., Ma’ayan, A.,
Iyengar, R., and Geiger, B. (2007). Functional
with a recent atomic force microscopy become faster, it will become atlas of the integrin adhesome. Nat. Cell Biol. 9,
study of isolated talin in vitro [10]. possible to analyze active processes. 858–867.
One important caveat in interpreting Integrin-mediated adhesions in 6. Parsons, J.T., Horwitz, A.R., and
Schwartz, M.A. (2010). Cell adhesion:
these studies is that the technique particular are dynamic, force-sensitive integrating cytoskeletal dynamics and cellular
localizes the fluorescent protein, not machines that respond to changes in tension. Nat. Rev. Mol. Cell Biol. 11, 633–643.
7. Kanchanawong, P., Shtengel, G.,
the target protein per se. Adding ECM composition, topography and Pasapera, A.M., Ramko, E.B., Davidson, M.W.,
1–2 nm (the diameter of EOS plus linker) mechanics to determine cellular Hess, H.F., and Waterman, C.M. (2010).
is straightforward, but the geometry responses [6]. But the active, dynamic Nanoscale architecture of integrin-based cell
adhesions. Nature 468, 580–584.
of the EOS relative to the rest of the process by which cells read the 8. Betzig, E., Patterson, G.H., Sougrat, R.,
molecule is unknown. Thus, the EOS topography and mechanical properties Lindwasser, O.W., Olenych, S.,
Bonifacino, J.S., Davidson, M.W.,
moiety could be highly mobile or of the ECM are poorly understood. Lippincott-Schwartz, J., and Hess, H.F. (2006).
could be ‘stuck’ at a defined A detailed analysis of the location of Imaging intracellular fluorescent proteins at
orientation relative to the amino- or specific protein domains in nanometer resolution. Science 313, 1642–1645.
9. Shtengel, G., Galbraith, J.A., Galbraith, C.G.,
carboxy-terminal domain to which it adhesions during sensing would be Lippincott-Schwartz, J., Gillette, J.M.,
is attached. While a mobile EOS will a huge step forward. The biology Manley, S., Sougrat, R., Waterman, C.M.,
Kanchanawong, P., Davidson, M.W., et al.
simply broaden the Gaussian community can look forward to (2009). Interferometric fluorescent super-
distribution for Z-axis location, a fixed major advances in understanding these resolution microscopy resolves 3D cellular
orientation could introduce a small but and other complex subcellular ultrastructure. Proc. Natl. Acad. Sci. USA 106,
3125–3130.
systematic error. While the potential machines as a result of the resolution 10. del Rio, A., Perez-Jimenez, R., Liu, R.,
error is small relative to the 40 nm scale revolution. Roca-Cusachs, P., Fernandez, J.M., and
Sheetz, M.P. (2009). Stretching single talin
examined by Kanchanawong et al. [7], rod molecules activates vinculin binding.
it may not be negligible for some References Science 323, 638–641.
1. Hynes, R.O. (2002). Integrins: bidirectional,
applications. allosteric signaling machines. Cell 110,
This study points the way toward 673–687.
2. Schwartz, M.A., and Assoian, R.K. (2001). Robert M. Berne Cardiovascular Research
approaches that are likely to transform Center, Departments of Microbiology,
Integrins and cell proliferation: regulation of
our understanding of multi-protein cyclin-dependent kinases via cytoplasmic Cell Biology and Biomedical Engineering,
complexes in living or fixed cells. The signaling pathways. J. Cell Sci. 114, 2553–2560. and Paul Mellon Urological Cancer Research
3. Katsumi, A., Orr, A.W., Tzima, E., and
ability to localize specific components Schwartz, M.A. (2004). Integrins in
Institute, University of Virginia,
in three dimensions with 10–20 nm mechanotransduction. J. Biol. Chem. 279, Charlottesville, VA 22908, USA.
resolution represents a major advance 12001–12004. E-mail: maschwartz@virginia.edu
4. Danen, E.H., and Sonnenberg, A. (2003).
that will catalyze progress in many Integrins in regulation of tissue development
fields of cell biology. Moreover, as and function. J. Pathol. 201, 632–641. DOI: 10.1016/j.cub.2010.12.025

Animal Cognition: Concepts from Ever since Herrnstein and Loveland

Apes to Bees [3] showed that pigeons could
discriminate novel stimuli (pictures with
people in them as opposed to
unpopulated pictures), there has been
New research shows that honeybees can classify arrangements of two visual
interest in whether animals have
patterns according to their spatial configuration. Can bees learn relational
concepts. The debate centres on the
concepts of ‘above’ and ‘below’? And are the underlying psychological
question of whether what appears like
processes comparable in humans and other primates facing similar tasks?
concept formation can instead be
explained by stimulus generalisation
Lars Chittka and Keith Jensen underlying most, if not all, higher [4,5] and discrimination by first order
cognitive competences when they perceptual features [6,7]. In one
The study of concept formation in result in judgments about x’’ [1]. We categorisation experiment,
animals is notoriously contentious. invite the reader to brainstorm for a few chimpanzees managed to sort a mix
On the one hand, it appears at the heart moments as to what the concept of of various different tools and food
of the question of whether animals ‘dog’ entails. It is much more than items into separate piles — despite the
can be capable of ‘abstract thinking’ a category encompassing a huge fact that exemplars within each
in a human-like sense; on the other variety of distinct breeds with sizes category had no obvious physical
hand, it is often difficult to rule out that ranging from the equivalent of a cat to resemblance to one another [8]. Had
what appears to be conceptual a pony, with distinct colour coats and the chimpanzees thus understood the
thinking cannot be explained by low- body proportions; the concept of ‘dog’ concept of ‘tools’ and ‘food’? It has
level cues. Machery [1] defines the also includes many types of semantic been pointed out that subjects could
term as follows: ‘‘A concept of x is information, various behaviour have simply classified objects by
a body of information about x that is patterns and ‘mentalities’, multiple whether or not they induce salivation
stored in long-term memory and that is ways in which dogs can be useful to [9]; and indeed there are many other
used by default in the processes their owners, and so on [2]. low-level cue explanations — after all,
Dispatch
R117

food and tools differ consistently in

surface texture, angularity of shape,
softness and other features.
Nonetheless, there are convincing Transfer
demonstrations of categorisation in test
many animal species, including bees,
that may not be so easily explained
by such low level processes [9,10].
To explore if animals can learn
relational concepts, such as sameness
and difference, subjects are often
tested with ‘matching-to-sample’ or
non-matching to sample paradigms.
They are shown a stimulus, and shortly
afterwards must choose between two
stimuli, one similar (or identical) and Trial 30
one different to the one previously
viewed. Many vertebrates [9], and
indeed bees [10], can solve such tasks.
ng

But impressive though they may be,

ini
Tra

the level of ‘understanding’ of the

concept of sameness or difference
needed for passing such tests might
be fundamentally simpler than in Trial 2
humans. Human subjects can
appreciate, for example, the ways in
which buses, ferries, and trains are
‘same’ (they are all means of public
transport) and in which ways they are
different. In typical matching-to-
sample tasks, it is the stimuli that are Trial 1
either similar or not. At its basic level,
this involves comparing an incoming
stimulus with the contents of working
memory, and the rule of whether to
choose ‘same’ or ‘different’ must be Left arm Right arm
stored in long term memory — but it Current Biology
does not necessarily require the
formation of a concept with the full Figure 1. Stimuli sequentially shown to bees on the back wall of a Y-maze.
range of implications of the term in Bees were trained to collect sucrose solution in one of the arms of the maze. Each pair of
human psychology. patterns simultaneously presented in the Y-maze contained the same visual stimuli, only
Now Avarguès-Weber et al. [11] distinguished by whether the ‘target’ occurred below or above the ‘referent’, which was always
the same. Bees encountered a series of different targets during training, and were rewarded
have performed experiments to see
only if they flew into the arm that contained the target below the referent (another group of
if honeybees (Apis mellifera) can bees was rewarded if they correctly picked the arm where targets consistently appeared
conceptualise above/below above the referent). The exact positions of targets and referents on the back walls of the
relationships. Bees were trained to arms of the Y maze were varied to ensure that bees could not learn a fixed configuration [11].
collect sucrose solution in one of two
arms of a Y-maze; if they chose the the maze could be located by a rule: so that they could not solve the task
wrong arm, they would receive quinine if the target was presented below the based on the presence or absence of
solution, a strongly aversive stimulus referent, there would a sweet reward the two items in each arm of the maze.
for bees. The arm of the maze for the bee; if the target occurred Avarguès-Weber et al. [11] also altered
containing the reward changed above, a bitter punishment was in the exact positions of the two items
unpredictably from trial to trial, and the store. (Another group of bees was to discourage subjects from learning
only way bees could identify the correct trained in the reverse way.) The bees a fixed compound of both. A trivial
arm was by choosing the correct swiftly learnt this task, and were even explanation for the findings might be
arrangement of two visual stimuli on able to pass a transfer test when that bees simply could not distinguish
the back walls of the Y-maze. One of presented with a completely unfamiliar the various targets, and therefore
these stimuli was the ‘referent’, which target, which, however, appeared treated them identically. However, the
was kept constant through all trials again in the correct location relative to authors ensured that the targets
of each individual bee (Figure 1). The referent. presented sequentially were of
other was the target (where the reward Thus, even though the same pairs different colours and shapes, and
was offered), but the visual appearance of stimuli were always presented in the indeed they demonstrated empirically
of the target changed multiple times right and left arm, the bees were able that bees could discriminate them
during training. The correct arm of to choose the correct arrangement, when trained to do so. These elegant
Current Biology Vol 21 No 3
R118

for example, had to manipulate a cursor

onto a fixation point using a joystick,
no doubt a challenging task in itself.
For infants, there was no reward or
punishment whatsoever; instead
performance was assessed via looking
times depending on the apparent
novelty in the stimuli. Thus, while the
tasks themselves might be roughly
comparable, there were profound
differences in how subjects could go
Figure 2. A natural analogue of the above/below-ness concept for bees. about solving them, and presumably
Bees might be using the shape of otherwise near-indistinguishable hillocks as a landmark (the in motivation. And even if all these
referent) and then pick the familiar one according to whether the area beneath (a flower differences were eliminated, it would
meadow) contains targets, or not. Other bees might have trained themselves to forage from be hard to determine the exact strategy
targets above the reference — in this case, trees presenting their flowers above the horizon
by which subjects solved the task,
of a different hillock.
and indeed there might even be inter-
individual differences with respect to
experiments clearly show that bees can findings of the new study by the underlying strategy: even if
do much more than just associate the Avarguès-Weber et al. [11] make performance is the same, some
visual properties of a target with beautiful biological sense. individuals might solve a task using
a reward. This is one possible explanation; relatively simple heuristics while others
To understand why animals are good another is that bees had indeed might use a more abstract reasoning.
at certain laboratory tasks, it often conceptualised the relationship of A broader question is whether the
helps to explore what the nearest above-ness and below-ness. To performance of bees and primates
natural analogue to the experimental explore this possibility, it us useful to (including infants) qualifies for concept
setup is. Bees outperform many compare the experimental procedures formation in a stricter sense. While
species of vertebrates at colour that have been used in primates to test learnt rules of same/different, above/
learning [9] presumably because for the existence of such concepts. below, more/less relationships are
associating colours with rewards is An important difference between the crucial ingredients of concepts,
a fundamental part of a forager bees’ new study on bees and those are such rules themselves concepts?
lifestyle [12]. Their remarkable ability to conducted on baboons (Papio papio) As pointed out above, a sameness/
solve sameness/difference tasks in [14] and capuchin monkeys (Cebus difference judgment can take on very
a matching-to-sample paradigm might apella) [15] is that primates were asked different levels of complexity, and the
also be related to natural foraging to further discriminate above and same is true for above/below
behaviour, where a strategy to win-stay below relationships when the referent relationships. Once humans have
or lose-shift is employed in deciding also changed. Both capuchins and understood the concept of this
which flower species to visit, baboons succeeded (excluding the relationship, they can appreciate the
depending on recent rewards [13]. In possibility that the referent is simply ways in which the sky is above them,
the new experiment, let us replace the used as a ‘landmark’), but bees remain the king is above the knight, the way
‘referent’ with a natural landmark — say to be tested. A striking difference in which ‘above/below’ relates to
the horizon profile of a hillock, where observed in those tests that were more ‘better/worse’, and so on. Transfer of
there are several similar hillocks directly comparable, though, was that relational rules across domains or to
nearby. The task for a bee to pick the it took the primates several hundreds novel situations is rarely explored in
hillock with a rich flower meadow or indeed thousands of trials to master animal studies (but see [10]) — for
beneath — i.e., pick the landmark, if it similar tasks, whereas the bees took example, would bees transfer their
has a target beneath — any target, no only a few dozen. Human infants at learnt preference for target above
matter the flower species or the exact three months of age fail altogether to a reference to the higher of two
location (Figure 2). In nature, as in show a concept of above and below flowering species in their natural
the laboratory experiments by unless the targets remain the same three-dimensional environment?
Avarguès-Weber et al. [11], the exact [16], though by the time they are Because of the proximity of the term
location of the landmark relative to 6 months old, infants perform ‘concept’ to ‘abstract thinking’, we
the potential target will vary depending comparably to bees and monkeys [17]. must be careful not to conclude too
on the approach angle, distance to But while the tasks are seemingly swiftly that fulfilling basic prerequisites
target, and so on. Thus, we need to similar, are these performances indeed of concept formation is indicative of
consider the possibility that bees might comparable? Bees were allowed to the full range of cognitive phenomena
solve the task by a relatively simple move freely to and from the apparatus that come with conceptual thinking in
algorithm: fly to the referent (the and could thus solve the tasks in humans [5]. Similar behavioural
landmark) if there is something — a manner equivalent to a spatial performances by different animals,
anything — in the ventral visual field as navigation task, and motivation and the fact that current definitions
you’re approaching the referent (or the remains high in forager bees since they permit application to a range of
dorsal visual field if trained to pick forage for the colony and thus never species’ abilities, don’t necessarily
stimuli above the referent). When satiate. Primates were tested in highly mean that the underlying processes
viewed in a naturalistic setting, the non-biological settings — the baboons, are similar (or similarly complex).
Dispatch
R119

There might be differences in how apes 7. Tomasello, M., and Call, J. (1997). Primate relations by tufted capuchin monkeys (Cebus
Cognition (Oxford: Oxford University Press). apella). J. Comp. Psychol. 118, 403–412.
(including humans) and monkeys form 8. Savage-Rumbaugh, E.S., Rumbaugh, D.M., 16. Quinn, P.C., Polly, J.L., Furer, M.J., Dobson, V.,
concepts [18], and other animals, Smith, S.T., and Lawson, J. (1980). and Narter, D.B. (2002). Young infants’
including bees, could use yet other Reference — the linguistic essential. Science performance in the object-variation version of
210, 922–925. the above-below categorization task: a result
solutions towards similar behavioural 9. Pearce, J.M. (2008). Animal Learning and of perceptual distraction or conceptual
problems. Cognition (Hove & New York: Psychology limitation? Infancy 3, 323–347.
Press). 17. Quinn, P.C., Cummins, M., Kase, J., Martin, E.,
10. Avarguès-Weber, A., Deisig, N., and Giurfa, M. and Weissman, S. (1996). Development of
References (2011). Visual cognition in social insects. Annu. categorical representations for above and
1. Machery, E. (2010). Precis of doing without Rev. Entomol. 56, 423–443. below spatial relations in 3- to 7-month-old
concepts. Behav. Brain Sci. 33, 195–206. 11. Avarguès-Weber, A., Dyer, A.G., and Giurfa, M. infants. Dev. Psychol. 32, 942–950.
2. Zentall, T.R., Wasserman, E.A., Lazareva, O.F., (2011). Conceptualization of above and below 18. Thompson, R.K.R., and Oden, D.L. (2000).
Thompson, R.K.R., and Ratterman, M.J. (2008). relationships by an insect. Proc. R. Soc. Lond. Categorical perception and conceptual
Concept learning in animals. Comp. Cogn. B. 10.1098/rspb.2010.1891. judgments by nonhuman primates: The
Behav. Rev. 3, 13–45. 12. Chittka, L., and Raine, N.E. (2006). Recognition paleological monkey and the analogical ape.
3. Herrnstein, R.J., and Loveland, D.H. (1964). of flowers by pollinators. Curr. Opin. Plant Biol. Cogn. Sci. 24, 363–396.
Complex visual concept in the pigeon. Science 9, 428–435.
146, 549–550. 13. Raine, N.E., and Chittka, L. (2007). Flower
4. Shettleworth, S.J. (2010). Cognition, Evolution constancy and memory dynamics in Queen Mary University of London, Research
and Behavior (Oxford: Oxford University Press). Bumblebees (Hymenoptera: Apidae: Bombus). Centre for Psychology, School of Biological
5. Penn, D.C., Holyoak, K.J., and Povinelli, D.J. Entomologia Generalis 29, 179–199. and Chemical Sciences, Mile End Road,
(2008). Darwin’s mistake: explaining the 14. Depy, D., Fagot, J., and Vauclair, J. (1999).
Processing of above/below categorical spatial
London E1 4NS, UK.
discontinuity between human and nonhuman
minds. Behav. Brain Sci. 31, 109–178. relations by baboons (Papio papio). Behav. E-mail: l.chittka@qmul.ac.uk
6. Chater, N., and Heyes, C. (1994). Animal Process 48, 1–9.
concepts: content and discontent. Mind 15. Spinozzi, G., Lubrano, G., and Truppa, V. (2004).
Language 9, 209–246. Categorization of above and below spatial DOI: 10.1016/j.cub.2010.12.045

Cytokinesis: Thinking Outside the Cell morphology defects in him-4 mutants.

In the C. elegans gonad, germ cells
form as part of a syncytium, with
How might the extracellular matrix contribute to cytokinesis? In a recent report,
incomplete cytokinetic furrows
evidence is presented that the conserved extracellular matrix protein
partitioning single nuclei that share
hemicentinHIM-4 is required for cytokinesis in worms and mice.
a common cytoplasm [8]. These nuclei
are eventually segregated completely
Shawn N. Jordan1, Sara Olson2, conserved role in cytokinesis. as the cell matures into an oocyte and
and Julie C. Canman1,* Hemicentin is a relative of a multigene physically separates from the
family of proteins called fibulins, syncytium [8]. HemicentinHIM-4 labeled
The extracellular matrix (ECM) secreted proteins that assimilate into with green fluorescent protein
functions akin to a cellular exoskeleton, the ECM and form higher order localized in a ring structure at the base
providing structural support and sites structures, such as elastic fibers [3]. of these incomplete furrows in the
of attachment for the cells it surrounds. The hemicentin protein contains gonad. In the absence of
However, the ECM is much more than a single, highly conserved von hemicentinHIM-4, the gonad in aged
just a scaffold. With established roles in Willebrand A domain, a long stretch worms became disorganized due to the
cell migration, tissue separation, and of immunoglobulin repeats, epidermal formation of multinucleated germ cells.
cell signal transduction [1], the ECM is growth factor domains, and Temporal analysis revealed that while
clearly a dynamic player in many a fibulin-like carboxy-terminal module nascent membrane partitions
cellular functions. [4]. In the roundworm C. elegans, appeared to have a normal structure,
Cytokinesis is the physical division hemicentinHIM-4 has been found in they soon became destabilized and
of one cell into two daughter cells that cell-matrix adhesion sites known eventually collapsed.
occurs at the end of the cell cycle. as hemidesmosomes and at multiple The recruitment of hemicentinHIM-4 to
Cytokinesis is accomplished by connecting junctions throughout the the tips of the membrane partitions in
constriction of a contractile ring body [5]. Cells surrounding the the C. elegans gonad was dependent
composed of filamentous actin and the hermaphroditic worm gonad also on the highly conserved Rho family
motor myosin-II (together actomyosin) secrete hemicentinHIM-4, and functional guanine nucleotide exchange factor
[2]. In order to divide, a cell must recruit disruption of the him-4 locus in (GEF) ECT-2. During cytokinesis, ECT-2
and coordinate a host of regulatory and C. elegans leads to a high incidence activates Rho and initiates assembly of
structural proteins to the division plane, of males due to defects in segregation the actomyosin ring at the constriction
but existing models of cytokinesis do of the X chromosome, resulting in XO site [2]. In an ECT-2-deficient
not consider a contribution from male progeny [6]. The him-4 mutants background, the majority of
secreted extracellular proteins. also display pleiotropic defects in hemicentinHIM-4 remained in the
The secreted ECM protein motility, adhesion, behavior, and gonad pseudo-coelomic fluid outside of the
hemicentin (HMCN1 or fibulin-6 in morphology [5]. gonad. Taken together, these results
Homo sapiens, HIM-4 in In a paper published in a recent issue suggest a role for hemicentinHIM-4 in
Caenorhabditis elegans) is of particular of Current Biology, Xu and Vogel [7] maintaining the cytokinetic membrane
interest as it may have an evolutionarily took a closer look at the gonad partitions in older worms, and that its