Review published in Journal of Consciousness Studies, Vol. 11 no. 5/6, 2004

Christof Koch
The Quest for Consciousness: A Neurobiological Approach
Englewood, CO: Roberts & Company, 2004, 448 pp., $45, ISBN 0-9747077-0-8 (hbk).

Ten years ago, Francis Crick dedicated a popular book, The Astonishing Hypothesis, to Christof Koch. Now Koch has returned the favor in The Quest for Consciousness, which sums up a collaboration going back to the 1980s; hence in this review ‘Koch’ can be taken as shorthand for ‘Crick and Koch’ (as the author himself suggests). Their collective enterprise marries a sense of wonder—a ‘quest’ for the ‘astonishing’—with a resolute focus on the microanatomy of the brain. The resulting blend might be called ‘reductionism in Wonderland’.

It was Lewis Carroll's Alice who inspired the now-notorious quip addressed to the Crick's readers in 1994: ‘You're nothing but a pack of neurons.’ Though this is usually read (and often quoted) as the acme of brazen reductionism, those who persisted to the end of the book found a more cautious Crick: ‘the words nothing but in our hypothesis can be misleading if understood in too naïve a way’ (p. 261). Astonishing or not, the hypothesis was essentially a theoretical framework for an embryonic research program which he hoped might eventually contribute to solving the old mind-body problem. Nevertheless, Crick's final chapter declared eloquently that the more we understand the mechanics of the brain/mind, the more marvelous it becomes. ‘To say that our behavior is based on a vast, interacting assembly of neurons should not diminish our view of ourselves but enlarge it tremendously’ (260).

In his new book, Koch's expressions of wonder are a little more muted, and the hypothesis somewhat modified from the 1994 version, but the main outlines of the research program remain the same. The fruits it has borne are beautifully presented here in rich detail, lavishly illustrated and impeccably edited. Specialists will appreciate the wealth of empirical detail and documentation, but Koch's style—relatively informal, concise and straightforward, though perhaps less colorful than Crick's—should appeal to more general readers.

Like its predecessor, Koch's book focusses mainly on the visual system in the brain and its relation to the experience of seeing, on the assumption that all aspects and modalities of consciousness ‘employ one or perhaps a few common mechanisms’ (15). The visual system is better known than others (so far) because it is more accessible to empirical research—mainly because the visual system in macaque monkeys is virtually identical to its counterpart in humans, and we can use laboratory techniques on monkey brains that can rarely be used on humans. The clarity with which Koch conveys this circumstance points to one significant virtue of his book: he provides a vivid sense of what neurobiologists actually do in their laboratories, and why.

Koch establishes his own motivation with an anecdote that happens to be non-visual. It all started in 1988 with a toothache: ‘I pondered why it hurt,’ he writes; though he already knew quite a lot about the neuronal mechanisms generating the pain, ‘none of this explained why it felt like anything!’ (xv). At this point readers will naturally expect Koch's Quest for Consciousness to deliver some kind of answer to the question of ‘qualia’, the ‘Hard Problem’ of Chalmers. Yet after 300 pages devoted mostly to describing neuronal mechanisms in exquisite detail, Koch admits that ‘why qualia feel the way they do remains an enigma’ (310). Why? Not because the Problem is misconceived, but because we still need more data. This is the reductionist faith in a nutshell, and though Koch avoids dogmatic insistence on it, his ‘quest’ never deviates from it. He sums up the nature of this quest as follows (19):

I seek the physical basis of phenomenal states within brain cells, their arrangements and activities. My goal is to identify the specific nature of this activity, the neuronal correlates of consciousness, and to determine to what extent the NCC differ from activity that influences behavior without engaging consciousness.

Koch's framework thus fits within the global workspace family, along with other models that associate consciousness with a temporary ‘coalition’ of neuronal populations or circuits whose more or less synchronized activity persists long enough to command attention and guide action, until the next coalition takes over. His glossary (which is excellent and extensive) defines the NCC as the ‘minimal set of neuronal mechanisms or events jointly sufficient for a specific conscious percept or experience’. The distinctive features of Koch's theory are most clearly indicated by the words minimal, sufficient and specific.

Koch's minimalism is remarkable in that he considers it a form of emergence theory. ‘The working hypothesis of this book is that consciousness emerges from neuronal features of the brain’ (10); and in the footnote which defines his terms, Koch says ‘A system has emergent properties if these are not possessed by its parts.’ Yet he does not consider it scientifically viable to treat consciousness as an emergent property of the whole brain, because its functioning includes all sorts of ‘activity that influences behavior without engaging consciousness.’ Therefore consciousness must emerge from some part of the brain. Clearly this part (the NCC) does not have a simple fixed location, and membership in the coalition must be constantly changing from moment to moment, but the more narrowly we can specify the components of this shifting subsystem, the closer we are to understanding how consciousness emerges. What we need, then, is a ‘strategy for circumscribing the NCC’ (italics mine).

Accordingly, Koch provides a very detailed overview of cortical architecture to serve as the background against which the NCC can be defined. He also devotes considerable space to nonconscious brain functions, including a whole chapter on ‘the zombie within’. This refers not to the philosopher's zombie (a notion which Koch finds ‘sterile’) but to ‘zombie agents, that carry out routine missions’ (206). These ‘cortical reflexes’ must be eliminated in order to pin down the NCC. Koch also eliminates the ‘enabling factors’, such as biological mechanisms which maintain a general state of arousal. More controversially, he eliminates the primary visual cortex (V1): it is certainly necessary for the normal functioning of the visual system, but V1 activity is also observed in the absence of any conscious percept, so V1 is not sufficient for conscious seeing. The same goes for the retina; these parts of the visual system are ‘pre-NCC’. Further up in the visual hierarchy, we can also eliminate the dorsal stream, which enables ‘blindsight’ in patients whose ventral stream has been damaged: they do not consciously see anything but can manipulate objects as if they did.

The higher we go in the visual hierarchy, the more specific are the percepts or objects with which a neuron's firing is correlated. Koch calls this specificity ‘logical depth of computation’ because it takes several layers of computation, including feedback loops, to derive it from retinal input. Taking this pattern to its linear conclusion would posit ‘grandmother neurons’ whose activity directly represents a consciously recognized object—an idea that often elicits ridicule from researchers. Koch, though, comes very close to embracing it when he reports the discovery of ‘Bill Clinton neurons’ (by Itzhak Fried, who recorded the firing of single neurons in an epileptic patient while viewing a series of 50 images). However, Koch explains that no single neuron could do the job alone; it would take a coalition to bring Clinton's presence in a scene to consciousness.

Koch also flirts with embracing another oft-derided concept, the homunculus in the brain. It is essential to his theory that the neuronal coalition which supplies specific content to consciousness must project to the ‘planning areas’ in the frontal lobes. (This is another reason for eliminating V1 from the NCC, since it does not send axons to that region in the monkey brain.) So we are justified in using the metaphor of a ‘homunculus residing in the front of the forebrain, receiving information from the sensory regions in the back, and relaying its output to the motor systems’ (302). The twist is that this is a nonconscious homunculus. Following Ray Jackendoff, Koch claims that ‘neither the process of thought nor its content is knowable by consciousness’ (297). Thus he eliminates not only the lower levels of the ‘computational mind’ but the higher levels as well, leaving only the ‘intermediate’ to furnish the NCC. Also eliminated is the post-NCC, the ‘penumbra’, which Koch invokes (Chapter 14) to account for the meaning of qualia: the NCC of the moment will have outputs to other cells representing a myriad associated concepts, habits and so forth, which are lit up to some extent by the heightened activity in the NCC; some of these may participate in the next ‘winning coalition’, whereupon the associations become conscious.

Koch's brand of empiricism is not only minimalist but microscopic: he argues that the key to consciousness may lie in the differences between types of neurons (hundreds have been identified already) and/or in the fine structure of the cortex, its layers and columns and interconnections. His approach is also temporally microscopic, focussing on the short time span within which a coalition can arise and maintain itself. Thus Koch has little to say about diachronic matters such as the development of the organism or its brain. This would explain, for instance, why he rejects the ‘enactive’ approach which emphasizes the primacy of movement in generating consciousness. Koch dismisses this idea mainly because paralyzed people can still be conscious. The fact that the paralytic's NCC are functional components of an organism with an ontogenic history of movement (and an evolutionary heritage of motility) lies beyond the bounds of his theory, presumably because this kind of fact is not observable in the laboratory.

Koch does recognize the need to place consciousness in a larger functional context (since he does not regard it as epiphenomenal), and deals with this in Chapter 14, but he is careful to mark these ideas as ‘speculative’. He argues that consciousness provides the organism with an ‘executive summary’ of situations which demand attention because the ‘zombie agents’ cannot cope with them automatically. Koch does mention some other views on the nature and function of consciousness, and (near the end) some other neurobiological theories, but in each case he tells us just enough about the alternatives to explain why he has no use for them. For example, he acknowledges a kinship with Edelman and Tononi's ‘dynamic core’ hypothesis, only to comment that he finds it a bit too ‘holistic’. This book offers ‘only cartoon-like pocket sketches’ (4) of competing approaches to consciousness, but it certainly presents a full-color and high-resolution picture of brain structure at the cellular level. Readers oriented to higher levels may find Koch's definition of the NCC (and of consciousness) too narrow, yet still find much that is rewarding—even astonishing—in The Quest for Consciousness.

Gary Fuhrman

[ gnoxic studies (home) ]