08 November 2001

Nature 414, 151 - 152 (2001); doi:10.1038/35102640

JOHN C. MARSHALL

John C. Marshall is in the University Department of Clinical Neurology, Radcliffe Infirmary, Woodstock Road, Oxford OX2 6HE, UK.

The New Phrenology: The Limits of Localizing Cognitive Processes in the Brain
by William R. Uttal
MIT Press: 2001. 255 pp. $39.95, £27.50

Phrenology is the vulgar Victorian term that refers to a set of conjectures first advanced by the Viennese physician Franz Joseph Gall (1758–1828). In addition to his exemplary clinical skills, Gall was the greatest neuroanatomist of his day, and he also set the agenda for the subsequent development of cognitive neuroscience. In brief, Gall argued that, just as the body from the neck down is comprised of distinct organs, each associated with particular physiological functions, so likewise is the body from the neck up. According to this hypothesis, the brain itself consists of many 'mental organs', each dedicated to a particular cognitive, emotional or voluntary function.

As William Uttal writes in the first chapter of The New Phrenology, the gallist paradigm thus asks three questions: "1. Can the mind be subdivided into components, modules, or parts? 2. Does the brain operate as an equipotential mass, or is it divisible into interacting but separable functional units? 3. Can these components — assuming they exist in some valid psychobiological sense — be assigned to localized portions of the brain?"

If these questions are interpreted fairly broadly, there can be few neuroscientists who would not today answer yes on all counts. But, as usual, the devil is in the detail. And it is those particulars that Uttal claims to examine in this highly polemical tract. Most interestingly, Uttal seems to be one of the few neuroscientists (his background is in electrophysiology) who are inclined to plead that the localization of cognitive processes in the brain is a chimera. Or, more precisely, he contends that there are so many "limits and constraints" on positive answers to any of the gallist questions that the reply might as well be no.

It is, of course, true that many different taxonomies of cognitive processes have been proposed over the past 2,000 years. And it is also true, as Uttal points out, that most of these classifications are pretty silly. Yet one might have expected that anyone concerned with the psychobiological categorization of psychic functions would have seriously considered the mental modules deployed by late-nineteenth-century neurologists to classify the cognitive disorders that could result from relatively focal damage to the human brain. The capacity to recognize objects in one's environment is compromised in the agnosias; the ability to assess where those objects are located is perturbed in a wide range of topographic disorders; the performance of skilled actions is impaired in the apraxias; remembrance of the past is clouded in the retrograde amnesias; the ability to communicate is weakened in the aphasias. In all these instances, careful clinical studies have established that different components of each of these mental modules can be independently impaired after brain damage.

Detailed neuropathology at autopsy (and, more recently, the in vivo visualization of structural lesions by such techniques as magnetic resonance imaging) then showed that distinct lesion sites could often be associated with different types of impairment to each of the cognitive subdomains in question.

Uttal is well aware that "the brain is not homogenous but made up of regions that are differentiated by function". And he is equally aware of some of the clinico-anatomical correlations that have been established in behavioural neurology. But he then proceeds to give "a totally different conceptualization of the localization problem". This restatement "offers, in place of a specific function being precisely localized (that is, instantiated, represented or encoded) in a particular place, the idea of one centre contributing to the operation of a complex system of nodes and loci that are collectively responsible for the behaviour." Well, yes: this is precisely what every behavioural neurologist and neuropsychologist has argued since (at least) Carl Wernicke's fractionation of the 'aphasic symptom complex' in 1874.

Sadly, Uttal does not engage with the details of this literature, despite acknowledging how clinical evidence supports the claim that complex cognitive processes such as speech and calculation "involve widely dispersed regions of the cerebral cortex". Rather, he rushes down the sexier path of functional neuroimaging to discuss such techniques as electroencephalography, positron-emission tomography and functional magnetic resonance imaging. Here he draws attention to failures of replication in which seemingly slight changes in experimental design, stimulus materials and task characteristics can lead to very different patterns of brain activation.

These are all genuine problems, but what needs to be debated here is the extent to which they reflect our failure to design good experiments rather than the brain's failure to instantiate distributed localization of function. Uttal also notes, correctly, that individual differences in brain organization can undermine the reliability of associations between structure and function. Again, however, this is a challenge to discover the range and limits of variation rather than a cause for despair. A judicious conjunction of psychophysical experimentation, functional neuroimaging, and lesion studies will no doubt eventually elucidate the basic functional architecture of cognition. We can then start on the really difficult issue of showing how, not where, neuroanatomical and neurochemical structures and processes instantiate our best guesses about the organization of the mind.

Many of the problems of functional localization that Uttal outlines are real enough, but he offers no new ways of solving them and no convincing arguments that some other approach would resolve them.