Structure Versus Function
Before proceeding, it is necessary to clear some semantic underbrush: Though the process presented here is based on a "functional" approach, the hard distinction between structure and function survives in the tenacious tradition of the language of dualism. That is, structure and function are seen as the realization, if you will, of brain and mind, respectively. Every brain function, however, must have its structural underpinnings, so the more tangible distinction, and the one more accessible to experiment, is one based on the timescale of change and the ease with which change can be induced. Most of what we consider in terms of brain function involves typically rapid, transient changes in the electrical activity in the brain, activity which may leave little in terms of residual imprint. Most of what we consider in terms of brain structure is that which remains essentially unchanged over longer time constants. This is a continuum, and over much of the range in timescale, one can appropriately describe a phenomenon either in the vocabulary of structure or that of function. One analogy that comes to mind is the redefinition by David Bohm of a noun as a "slow verb".

Another way of looking at the structure/function duality is in the division between hardware and software in computers. On one hand, we have the true hardware, the semiconductor devices and ancillary items needed to service and operate them. On the other hand, we have the operating system software. Though this can be changed, it is generally modified only rarely and deliberately. At the next level is the applications software. A number of different modules may be drawn upon (brought to consciousness?) at a given moment, and there is in fact considerable "interaction" with the outside world which may make "functional" changes in the application software; and there may even be some adaptation to what the user typically wants. At the top level is the phenomenology of what is created with the applications software, which has typically a very transient quality (e.g., imagery). One could argue that at each level we are dealing with physical electrons moving around from site to site (structure), but that would be cumbersome, and not really to the point. Similarly, one could talk about software failures in terms of "electron deficiencies" in certain memory locations. This is both true and absurd as a model for software failures. Every level has its appropriate terminology, referring progressively to structure, function, and objects (gestalts).

The categories distinguished here can find their analogues within the brain. However, the boundaries are not as discernible and the distinctions between structure and function even less definitive. Nevertheless, let us push the analogy forward a little further: A similarity can be drawn between our brain's neuromodulator systems and the operating system software of a computer. There is persistence in the workings of our neuromodulator systems that puts them on a different timescale than the applications software (which might involve the processing of a visual image, for example). Yet it would not be correct to regard the characteristics of a person's neuromodulator systems as immutable (even absent any drug intervention). Over time, it is clear that environmental influences, for example, can effect changes in neuromodulator function. A person may become more or less hypervigilant over time; he may become more depressed or anxious. He could also, however, achieve "spontaneous" recoveries from depression, which can be simply interpreted as autonomous normalization of neuromodulator functioning.

The distinction, therefore, between categories of structure and function is not based so much on issues of transience versus immutability, per se, but rather on a multiplicity of factors: the timescale of change; how matters have been historically viewed; and the level of abstraction which is appropriate to the discussion. This whole issue is currently very much in flux, and somewhat confused. We have, for example, the following from Michael S. Gazzaniga, director of the Center for Neuroscience at the University of California in Davis: "When someone remembers something, is there a structural-or discrete anatomical-change in neuronal synapses? Or is it functional change, which would simply reflect reprogramming of the pattern of neuronal discharges in the nervous system?" (Gazzaniga, 1995).

Here the posited "structural" change could equivalently be talked about in terms of function, and the posited "functional" change (which clearly must be sufficiently robust to persist long-term if it is to represent a memory) can be talked about in terms of structure (altered synaptic coupling strengths).