The Object of the Sensory Cortex

Abstract

The phenomenon of ''transfer'' is used to draw attention to the point that whereas the brain stem animal can solve a variety of individual problems ad hoc, only the intact animal appears to be able to use the result of one problem to solve a closely related problem. It is suggested that this is because information is organized in the cortex as objects and concepts, rather than as a set of more or less elaborate features. The shortcomings of a hierarchical system of feature detectors as a model of cortical organization are examined, and compared with a distributive model. It is pointed out that our ways of thinking about complex tones – e.g. their pitch and harmonic structure – is related to the fact that sound suffers a Fourier transform at the cochlea, and it is this transform which is represented in the brain stem. It is suggested that an analogous integral transform may equally naturally represent the data as ''objects'' in the cortex. The resistance of such transforms to degradation by ablation, and the implications for single neurone behaviour are discussed.