A New Model of Global Brain Function

A NEW MODEL OF GLOBAL BRAIN FUNCTION HARRY BLUM* Some like to understand what they believe in. Others like to believe in what they understand . Stanislaw Lec [?] I. Why Another Model? Despite the deep concern ofphysiology, psychology, philosophy, and mathematics, our understanding of intellect and brain remains a central enigma. Neither the long historical involvement with these problems nor the intensity of recent efforts has sufficed to give us cohering insights to the large body of disparate data. We do not know what memory is, where memory is, how we get in and out ofmemory, nor what memory abstracts from the world. Yet, the faith we have in our scientific culture, orperhaps the investment we have in its methods, or perhaps the difficulty we have ofputting aside our cultural preconceptions seduces us, so that we do not come to grips witli the depths ofour ignorance in these areas. A field so rich in conceptual possibilities and so fraught with experimental difficulties can be dealt with onlyvia theoreticalpreconceptions. However, we are the captives of our intellectual ontogeny, and our most insidious preconceptions are implicit. Any attempt to generate new constraints in such a rich field ofpossibilities is extremely speculative and smacks more ofart than ofscience. Yet, we must educate ourselves on the broad possibilities in order to open our eyes honestly to actualities. Our experimental sciences and our formal models give us images ofman that do not begin to conform with what we see when we really look at him. Man cannot be understood simply as a logical or a statistical organizer ofhis environment. He goes through life full of logical contradictions that he rarely faces. * Air Force Cambridge Research Laboratories, Bedford, Mass. This is an expansion of a paper presented at the Symposium on the Analysis of Central Nervous System and Cardiovascular Data Using Computer Mediods held in Washington, D.C., October 29-30, 1964. NASA SP-72, Washington , D.C., IOÖ5. 381 (Indeed, he would be a Hamlet of monstrous proportions, unable to act at all, ifhe had to check each action for logical consistency with all his other actions.) And he is an extremely poor statistical processor, except where he has been clued into the key elements ofa problem. Yet, he does something right—something that is not illuminated by our preconceptions ofinformation processing; something that is far better comprehended by those in the arts and humanities than by those in the sciences. But science will not be altered by mere philosophical objections. And properly so. One can only divert it by presenting more fruitful insights and alternatives . In this model, I am attempting to get at the roots ofour preconceptions at the deepest level that I see them and to do this in a way that suggests—I believe—new scientific directions. We find an excellent example ofthe preceding objections to our scientific views in our experimental work on electroencephelographic data, a major topic for which this symposium was convened. We have seen here a number of masterful and elegant analyses ofthese data by some ofthe foremost researchers in the field. The experiments have been extremely sophisticated, conceived to extract a maximum ofinsight from the experimental and computational techniques we have at our command. Yet, the necessity for preconceiving the problem has applied here also. Although the computational techniques are centered around harmonic and statistical analysis, notions shown to have great mathematical generality, the results have an elusive quality reminiscent ofPtolemaic astronomy. Nowhere is cognitive function apparent. Exponential increases in the quantity ofdata processing are necessary to explore another epicycle. And even with the additional epicycle, corrective terms for statistical prediction, rather than incisive functional insights, are likely to result. (Where would our orbital universe be if Ptolemy had had a computer?) It is no wonder that there is a lingering dissatisfaction with the application of mathematical techniques developed for studying noise and randomness to this problem. They can only have been taken up as a last resort. Exploring qualitatively a wide range of simple models is essential if we are to find meaningful alternatives to the complexity of such uncommitted approaches. Ifthese models are to be incisive, they must reflect constraints that are acting on...