6 Introduction: from trepanning to functional neuroimaging

The symposium aims to outline how modern neuroimaging techniques, such as positron emission tomography (PET) and functional magnetic resonance imaging, can illuminate the functional anatomy of higher mental processes in the living human brain. Currently, the value of these techniques is hotly debated. Some cognitive scientists see functional brain imaging as merely a very expensive form of phrenology in which the performance of gross psychological tasks (e.g. recognizing a face) is associated with the physiological activity of equally gross regions of the brain (e.g. occipito-temporal cortex). Others, by contrast, see functional neuroimaging as the first major step towards understanding the organizational principles that are responsible for how activity within distributed networks in the brain can support mental functions. I shall not enter this controversy here, but rather try to set the problem of structure/function relationships in the historical context of lesion studies. Successful trepanning can be dated back to the Mesolithic era, and evidence from an Egyptian surgical papyrus (c. 1700 bce) already links damage to the brain with loss of speech. More generally, in On The Sacred Disease, Hippocrates (460? – 377 bce) can claim that neither the diaphragm nor the heart ‘takes any part in mental operations, which are completely undertaken by the brain’. What Hippocrates does not tell us is which parts of the brain are responsible for which mental operations. For that kind of account of mind/brain relationships, we must turn to Heriophilus of Alexandria (c. 300 bce) whose ventricular theory of cognition was more fully elaborated by the Early Church Fathers in the 4th and 5th century of the Common Era. In this model, the primary psychological faculties postulated by Aristotle (The Common Sense; Imagination; Reasoning; Memory) were successively localized in the cerebral ventricles, with ‘information-flow’ between these ‘cells’ regulated by the vermis. This notion of the punctuate localization of cognitive functions in specific regions of the brain persisted in the ‘organology’ of Franz-Joseph Gall (1758–1828). But in addition to moving the cerebral substrate of cognition from the ventricles to the cortex, Gall also replaced Aristotle's faculties with a new set of localized structure/function relationships, including the organs of size, shape, colour, number, music and language. It was Gall's localizations and, more importantly, his view of what could be localized, that initially (and very successfully) drove 19th and 20th century behavioural neurology. It is less clear that classical localization theory can adequately support 21st century neuroscience. Empirical and conceptual problems with classical views had already begun to emerge in the late 19th century. These problems led to an intensive search for ways of measuring the physiological activity in the living human brain that is correlated with ‘different kinds of intellectual work’ (Lombard, 1879). The most sophisticated of these techniques – multichannel cerebral thermometry (Broca, 1879) – is the direct conceptual precursor of PET and fMRI.