In the last decade or two, Dominic ffytche and his colleagues in London have done pioneering research on the neural basis of visual hallucinations. Based on the detailed reports of dozens of subjects, they developed a taxonomy of hallucinations, including categories like figures with hats, children or small people, landscapes, vehicles, grotesque faces, text, and cartoonlike faces. (This taxonomy is described in a 2000 paper by Santhouse et al.)
With this classification in hand, ffytche went on to do detailed brain-imaging studies in which selected patients with different categories of visual hallucinations were asked to signal the beginning and end of their hallucinations while being scanned.
There was, as ffytche et al. wrote in a 1998 paper, “a striking correspondence” between the particular hallucinatory experiences of each patient and the particular portions of the ventral visual pathway in the visual cortex which were activated. Hallucinations of faces, of color, of textures, and of objects, for example, each activated particular areas known to be involved in specific visual functions. When there were colored hallucinations, there was activation of areas in the visual cortex associated with color construction; when there were facial hallucinations of a sketchlike or cartoonlike character, there was activation in the fusiform gyrus. Visions of deformed or dismembered faces or grotesque faces with exaggerated eyes or teeth were associated with heightened activity in the superior temporal sulcus, an area specialized for the representation of eyes, teeth, and other parts of the face. Text hallucinations are associated with abnormal activation in the visual word form area, a highly specialized area in the left hemisphere.
Ffytche et al. observed, moreover, a clear distinction between normal visual imagination and actual hallucination—thus, imagining a colored object, for example, did not activate the V4 area, while a colored hallucination did. Such findings confirm that, not only subjectively but physiologically, hallucinations are unlike imagination and much more like perceptions. Writing of hallucinations in 1760, Bonnet said, “The mind would not be able to tell apart vision from reality.” The work of ffytche and his colleagues shows that the brain does not distinguish them, either.
There had never before been direct evidence of such a correlation between the contents of a hallucination and the particular areas of cortex activated. We have long known, from observation of people with specific injuries or strokes, that different aspects of visual perception (color perception, face recognition, movement perception, etc.) depend on highly specialized areas of the brain. Thus, for example, damage to a tiny area of the visual cortex called V4 may knock out color perception but nothing else. Ffytche’s work is the first to confirm that hallucinations make use of the same visual areas and pathways as perception itself. (Ffytche has emphasized more recently, in papers on the “hodology” of hallucinations, that attributing hallucinations, or any cerebral function, to specific brain regions has its limitations, and that one must pay equal attention to the connections between these areas.)11
But while there are neurologically determined categories of visual hallucination, there may be personal and cultural determinants, too. No one can have hallucinations of musical notation or numbers or letters, for example, if they have not actually seen these at some point in real life. Thus experience and memory may influence both imagery and hallucination—but with CBS, memories are not hallucinated in full or literal form. When people with CBS hallucinate people or places, they are almost never recognizable people or places, only plausible or invented ones. CBS hallucinations give one the impression that, at some lower level, in the early visual system, there is a categorical dictionary of images or part images—of generic “noses,” for example, or “headwear” or “birds,” rather than of particular noses or headwear or birds. These are, so to speak, the visual ingredients called upon and used in the recognition and representation of complex scenes—elements or building blocks which are purely visual, without context or correlation with other senses, without emotion or particular associations of place or time. (Some researchers have called them “proto-objects” or “proto-images.”) In this way, CBS images seem more raw, more obviously neurological, not personal like those of imagination or recollection.
