High Quality
Open the downloaded document, and select print from the file menu (PDF reader required).
Acta Neuropsychiatrica 2007: 19: 139–148 All rights reserved
DOI: 10.1111/j.1601-5215.2007.00204.x
#
2007 The Authors Journal compilation
#
2007 Blackwell Munksgaard ACTA NEUROPSYCHIATRICA
Review article
When brains expand: mind and the evolutionof cortex
Kirkcaldie MTK, Kitchener PD. When brains expand: mind and theevolution of cortex.
Objective:
To critically examine the relationship between evolutionaryand developmental influences on human neocortex and the properties of the conscious mind it creates.
Methods:
Using PubMed searches and the bibliographies of severalmonographs, we selected 50 key works, which offer empirical support fora novel understanding of the organization of the neocortex.
Results:
The cognitive gulf between humans and our closest primaterelatives has usually been taken as evidence that our brains evolvedcrucial new mechanisms somehow conferring advanced capacities,particularly in association areas of the neocortex. In this overview of neocortical development and comparative brain morphometry, wepropose an alternative view: that an increase in neocortical size, alone,could account for novel and powerful cognitive capabilities. Other thanhumans’ very large brain in relation to the body weight, themorphometric relations between neocortex and all other brain regionsshow remarkably consistent exponential ratios across the range of primate species, including humans. For an increase in neocortical size toproduce new abilities, the developmental mechanisms of neocortex wouldneed to be able to generate an interarchy of functionally diverse corticaldomains in the absence of explicit specification, and in this respect, themammalian neocortex is unique: its relationship to the rest of the nervoussystem is unusually plastic, allowing great changes in corticalorganization to occur in relatively short periods of evolution. The factthat even advanced abilities like self-recognition have arisen in speciesfrom different mammalian orders suggests that expansion of theneocortex quite naturally generates new levels of cognitive sophistication.Our cognitive and behavioural sophistication may, therefore, beattributable to these intrinsic mechanisms’ ability to generate complexinterarchies when the neocortex reaches a sufficient size.
Conclusion:
Our analysis offers a parsimonious explanation for keyproperties of the human mind based on evolutionary influences anddevelopmental processes. This view is perhaps surprising in its simplicity, butoffers a fresh perspective on the evolutionary basis of mental complexity.
Matthew T. K. Kirkcaldie
1
,Peter D. Kitchener
2
1
Department of Physiology, School of MedicalSciences, The University of New South Wales,Randwick, New South Wales, Australia; and
2
Department of Anatomy and Cell Biology, TheUniversity of Melbourne, Melbourne Victoria,AustraliaKeywords: cerebral cortex; cognition; prefrontalcortex; neuroanatomyDr Matthew T. K. Kirkcaldie, Department of Physiology,School of Medical Sciences, Rm 308a1, WallaceWurth Building, The University of New South Wales,Randwick, NSW 2052, Australia.Tel:
1
61 2 9385 2560;Fax:
1
61 2 9385 1059;E-mail: m.kirkcaldie@unsw.edu.auBoth the authors contributed equally.
Consciousness presents a special challenge forneuroscientists; despite increasingly detailed con-firmation that it is produced by the action of thebrain, it has been considered an inappropriate, orat least unwise, area of study. Omitting the brain’smost important function has placed neuroscientistsin a strange situation. If we maintain that con-sciousness is, in principle, not amenable toinvestigation, then we are effectively dualists – or
Ô
closet dualists’ (1), if we also adopt the fashionthat dualism is untenable. Otherwise, we may betempted to endorse the claim that consciousnessarises from some complexity of neuronal interac-tion, as an
Ô
emergent feature’ of the highlyrecursive information processing transformationsafforded by neuronal interactions. But as KevinKelly points out in his postscript to
Out of Control
(2),
Ô
emerges’ can be just another way of saying
139
Ô
happens’, in which case an
Ô
emergent’ explanationfor consciousness might be little more than anappealtomagic.Morecautiously,we couldtaketheposition that the problem of consciousness is bestavoided; that the gulf separating genes, membranes,synapses and circuits from embodiment, unity,agency and continuity (3) is unbridgeable.There has, nevertheless, been a growing interestin consciousness paralleled by an increasingsophistication in describing brain activity duringspecific conscious experiences. Some correlates,such as electroencephalograph (EEG) activity indifferent states of consciousnesses, are well estab-lished, whereas others such as functional magneticresonance imaging (fMRI) or exploiting definableconscious percepts in binocular rivalry (4) haveonly recently enhanced our knowledge. Despitebetter empirical data, many fundamental questionsabout the forms and distribution of consciousnessin animals remain unanswered. Our own con-sciousness is irrefutably self-evident, and we arecomfortable in ascribing this consciousness toother people, but we have little insight into otherspecies: are they conscious in a rudimentary orprototypic form? Do they have affective experi-ences, and if so, are they comparable to humanemotions? Some insight has been gleaned bystudying conscious behaviour in animals, forexample, the mirror test (recognizing a reflectionas the self, instead of another individual) is passedby some species: humans, some of our closerelatives (5), elephants (6) and dolphins (7). Thiscapacityseemscloselyalignedwith
Ô
theoryofmind’:the ability to conceptualize the viewpoint of, andattribute mental states to, other individuals. Sucha test is relatively specific but leaves unanswered thedifficulty of explaining what consciousness is like inanimals exhibiting empathetic behaviours to othermembersof theirspecies (8). Even ifwe discount theapparent self-consciousness of such behaviour, howdoweconceptualizewhattheseanimalsexperiencedinthemselves?Manymammalsexhibitsophisticatedperceptual and social behaviour, including re-cognition of conspecifics and their actions. AsGriffin (9) has asked, would there really be withinthe conscious perception of the activity of otheranimals a
Ô
perceptual black hole’ where its ownactionswouldberepresented?Itissimilarlydifficultfor us to conceive of forms of consciousnesses thatsolely elaborate the experience of the present, withlittle conception of how the future may be shaped,and a memory embodied only as behaviouraltendencies shaped by the unrecalled past.The biggest difficulty in understanding conscious-ness is understanding how there can be any sort of awareness or phenomenality. This has been termed
Ô
the hard problem’ (10) in recognition that con-sciousness is fundamentally unlike everything else weknow of: there seems no way to relate it to anyproperty or process, or any member of the set of allthings that are not consciousness. The study of thebrain in relation to consciousness is thereforenecessarily confined to the
Ô
easy’ problem – theneural events associated with the generation of phenomenality, the physiological correlates of con-sciousness. A contrasting view is that consciousnessonly
seems
unique but is in fact nothing more thanthe result of complex brain activity, as much expli-cable in terms of the physiology of biological entitiesas any other complex and superficially mysteriousprocesses. Dennett (11) points out that while ourintuition that consciousness is special may bepowerfully compelling, intuition is a poor guide tothe underlying nature of complex processes andsuggests that we let go of this fondly held, un-supported belief, and set about better understandingthe biological processes that generate conscious-nesses: the easy, and only, problem. Calling theproblem
Ô
easy’ implies only that it comprises instan-ces or combinations of known phenomena; a sepa-rate issue is whether practical considerations mightnonetheless prevent a full explanation (12).Fortunately, progress in understanding how thebrain generates consciousness may not requirea choice between philosophical stances. The easyproblems of consciousness are practical and usefulpursuits, regardless of whether we believe that theirsolution offers a complete explanation of consciousexperience. We might suspect that there are crucialorganizing principles or fundamental phenomenayet to be apprehended and included in the analysis,but even if a detailed account of the brain’s activitycannot fully explain the associated phenomenality,such an account will almost certainly be anessential
part
of the comprehensive explanation.This view stems from Hubel and Wiesel’s seminalinsight (13) about the visual system: that theexperienced features of visual percepts coincidewith the features extracted by the computationalactivity of neurons. There is still considerableuncertaintyastowhatanalysesneuronsinthevisualsystem are actually performing (14), but it seemsextraordinarily unlikely that a later understandingof how perceptual consciousness arises will showthat this activity is arbitrary and unrelated to thegeneration of the associated percepts.
Neuronal activity and consciousness
It has been noted that one of the difficulties instudying consciousness is the lack of a fixed pointfrom which to theorize (15). The lack of consensus
Kirkcaldie and Kitchener
140
on the form that a theory of consciousness mighttake may also obstruct the interpretation of lessovertly theoretical approaches, like attempting tocharacterize the neurophysiological activity relatedto aspects of consciousness. The obvious, anddemonstrable, characterization of neurons as
Ô
information processing automata’ connotes thatthe neurophysiological correlates will need to beresolved to this computational level for any insightinto the mechanism of phenomenality. Supportingsuch a conception are two of the most significantinsights into neural function: the ionic basis of theaction potential (16) and the extraction of higherorder visual features by connections of neurons inthe visual pathway (13). These insights rank amongthe greatest achievements in science, but they maynot be the only conceptual tools needed to extendneurophysiological studies to the correlates of consciousness. It could be, for example, that thereare higher order arrangements of these neuralproperties that the current computational viewmay not reveal (17) (analogously, the primaryamino acid sequence of an enzyme does not explainits catalytic properties). Even at the level of singleneurons, the characterization of informationprocessing by neurons may be open to differentinterpretations (18): the discovery of retrogradedendritic action potentials in cortical pyramidalneurons (19) has supported the alternative conceptof the action potential as a reset mechanism, actingto synchronize activity rather than express the sumof synaptic inputs. It might also be conjecturedthat features at other levels could be critical – canwe, for example, overlook complex electrodynamicphenomena generated by concerted neuronalactivity [see, e.g., Freeman (20)] in our eagernessto focus on the action potential?When relating neuronal activity to the moreglobal functions of the brain, we soon run intointerpretational difficulties in analyzing neuronsbeyond those which initially process sensory inputs.An example is the representation of perceptualobjects by neurons that are not involved in theinitial grouping and segmentation of the retinalimage – itself a task of breathtaking complexity.Understanding the way in which visual perceptscould correspond to the concerted activity of vastnumbers of individual neural responses is concep-tualized as an encoding problem;
sparse
and
distributed
coding offer contrasting models of howpercepts or their antecedents might be encoded insuch activity. Sparse encoding is an extension of thehierarchical feature extraction posited for lowerlevels of visual processing, such that particularcombinations of low-level features trigger responsesin a few specific neurons. In the extreme case, thereare single neurons that respond to a single distin-guishable object of perception. Alternatively, dis-cernible percepts may arise from unique spatial andtemporal distributions of neuronal activity. Thedifficulty of deciding between such explanations ishighlighted by a recent investigation (21) of themedial temporal lobe in surgical patients, in whichexperimenters recorded single-unit responses tovisual stimuli of well-known landmarks and people.The authors interpreted their results as supportiveof sparse representation, but a most surprisingaspect of the data was that an extraordinarily highproportion of recorded units (14% of the total andabout one third of those with invariant responses torepeated stimuli) had markedly selective responses.Several neurons showed responses to multiplestimuli that were different representations of a singleindividual (such as photographs showing differentviews, a hand-drawn portrait and the written nameof the actress Halle Berry) but not to otherindividuals. Such specificity could be consistentwith a sparse code for an abstract representation of individuals but does not explain why so many of theneurons encountered were responsive to the rela-tively few objects in the stimulus set. The set of stimuli used was limited (extremely small, given therange of all possible visual stimuli) and althoughBerry would, by virtue of film and publicity, likelybe familiar to the American subjects tested, it wouldbe expected that neurons at the top of a sparsehierarchy representing the actress would be encoun-tered only rarely. Part of the difficulty in revealingthe
Ô
neuronal code’ of abstract visual percepts is thatthe role of these neurons within the entire neuralresponse to the stimuli is unknown: the highlyselective response exhibited could relate to percep-tual representations
and
to top-down visual pro-cesses, attentional mechanisms, memory andmotivational factors (21). The authors of the studyalso suggest that what is revealed by these responsesmay be something like place cells of the hippocam-pus, which generate a dynamic reference system forposition in the environment (22). Such
Ô
Halle Berryneurons’ would analogously participate in dynamicbindings within a conceptual space, leaving ques-tions of sparse and distributed representationunresolved, and offering little insight to theirrelationship to conscious percepts.
Does quality equal quantity?
An important consideration in comparing humanminds to those of other animals (especially otherprimates) is the extent to which phenomena such asHalle Berry neurons, demonstrations of theory of
Mind and the evolution of the neocortex
141
Add a Comment