Beninger (2000), who sharpened my understanding of incentive learning many years ago when he was my graduate student, contributes his expertise on the role of dopamine in reward-related learning.
2. On comparing his review with others, I began to wonder whether a personal brain-washing, or mind-bending, was a prerequisite for appreciating the approach to theory underlying the book. I remembered that about 25 years ago Rick and I submitted a theoretical paper to the Psychological Review and, rooting among my archives, somewhat to my surprise I discovered a copy. With it were the rejection letters of two reviewers who, according to the editor, "know a lot more about the nervous system than most psychologists... and I think both are genuinely sympathetic to efforts toward theory." Both, however, denounced our efforts as idle speculation, unworthy of the Psychological Review.
3. On re-reading our manuscript, I was surprised at the extent to which it anticipated ideas in The Autonomous Brain. For example, we recognized that the way a stimulus is interpreted by an animal changes, via the action of selective attention, depending on the animal's momentary need. Our model of incentive learning was similar to the one in the book, including a tentative suggestion, based on the anatomy of the dopaminergic paths, that the basal ganglia are important for such learning. After a quarter of a century it is difficult to remember which of us contributed what ideas, so perhaps I should be wondering who was bending whose mind at the time. Certainly our discussions helped to clarify my ideas about incentive learning.
4. Idle speculation or not, Rick went on to a successful career investigating the relation between incentive learning and the motor system, and the role of dopamine and other neuroactive substances in synaptic modification. The late Gordon Mogenson, another McGill graduate student who worked with me, was later instrumental in demonstrating the importance of the ventral striatum in reward (Mogenson et al.,1979; Mogenson, Jones and Yim, 1980).
5. Unsurprisingly, Beninger's comments emphasize the chemistry of synaptic modification, an area to which he has made important contributions. I am grateful for these additions to the story in the book but I am rather sorry that Beninger's views about certain themes in The Autonomous Brain (Milner, 1999) that I thought might be controversial are too similar to mine to provoke much debate. Beninger is certainly correct in detecting a materialist outlook. The book also displays antagonism to the empiricism that dominated psychological ideas and practice for most of the 20th century. Beninger picks up a point that had not occurred to me in support of a nativist view - the fact that there are not enough neurons to accommodate all the possible combinations of olfactory sensitivity - but I suppose an ardent empiricist might argue that only combinations that are frequently experienced result in the formation of connections.
6. The days are long gone when a psychologist could postulate, with little chance of argument, that the visual system of the human infant invents itself out of a chaotic bundle of neurons. A similar philosophy is still to be found, however, among designers of connectionist neural networks. But even though nativism is enjoying a revival, encouraged by developments in genetics, it has not reached the point at which anyone might suggest that we have innate knowledge of the meaning of all of our percepts. We still have to learn the significance of the letters of the alphabet. As far as vision is concerned, the nativist position is that the mechanism by which the visual system produces consistent, discriminable neural activity based on shape must be innately programmed. It could hardly be otherwise considering the extreme complexity of the circuits required to achieve normal perception, coupled with the fact that most of the circuitry develops prenatally, in the absence of visual stimulation.
7. In most instances such perceptual activity must acquire meaning through associations formed by experience. The big question is: What do we mean by meaning? The view expressed in the book is that meaning stems from associations with innate motivation systems, either directly or via a chain involving other concepts and responses. The adult human stores a vast number of associations and can form new ones almost instantly. It seems likely that the simultaneous distribution of information throughout the system would overwhelm any purely parallel network, but the requirements for sequential communication are more modest. Forty-five distinct sounds, arranged in sequences of 4 or less (see Cowan, 1999) can produce more than 3.5 million different words. Even if only 10 percent of them are pronounceable, the average person would find them more than adequate for communicating information. Using sequential codes millions of groups of neurons, each incorporating a four step sequence decoder, can be individually addressed via a common 40 to 50 conductor party line. A widely dispersed group of neurons could all link themselves to the same address code by firing together whenever that code was being disseminated. All behavior can be reduced to sequences of activity of a relatively small number of effectors. Hence, it seems likely that groups of neurons representing different concepts may be associated with each other via sequential codes generated in the motor system. It is to this suggestion that Beninger refers so succinctly in his paragraph 6, with his statement "Milner argues that ideas are associated with each other via the response system."
8. A record of the order in which things occur is of vital importance for behavior, but until recently the problem has been poorly understood and, apart from some repudiated speculation about chained associations, ignored by most psychologists. As the complexity and diversity of synaptic chemical aftereffects is becoming known it provides the basis for more soundly based speculation about mechanisms for dealing with temporal order, especially for short intervals.
9. For longer intervals order may be stored by a different mechanism somewhat more similar to a tape recorder. Recent discoveries concerning the continuous supply of new neurons into the hippocampus and the prefrontal, inferior temporal and posterior parietal neocortex of primates (Gould, et al, 1999a; 1999b) may answer the question raised in Chapter 10 of the Autonomous Brain concerning the ordering of episodic memories. It is possible that the memory of today's parking place of one's car is partly stored in newly matured neurons, thus distinguishing it from memories of where the car was parked on previous occasions. Such a mechanism would also allow for retrieval of information on the basis of when it was recorded.
Beninger, R. J.(2000) Material Autonomy PSYCOLOQUY 11(054) ftp://ftp.princeton.edu./pub/harnad/Psycoloquy/2000.volume.11/ psyc.00.11.054.autonomous-brain.6.beninger http://www.cogsci.soton.ac.uk/cgi/psyc/newpsy?11.054
Cowan, N. (1999) The Magical Number 4 http://www.cogsci.soton.ac.uk/bbs/Archives/bbs.cowan.html
Gould, E., Reeves, A.J., Fallah, M., Tanapat, P., Gross, C.G. & Fuchs, E. (1999a). Hippocampal neurogenesis in adult Old World primates. Proceedings of the National Academy of Sciences, U.S.A., 96, 5263-5267.
Gould, E., Reeves, A.J., Graziano, M.S. & Gross, C.G. (1999b). Neurogenesis in the neocortex of adult primates. Science, 286, 548-552.
Milner, P.M. (1999a) The Autonomous Brain. Erlbaum, Mahwah NJ
Milner, P.M. (1999b) Precis of "The Autonomous Brain" PSYCOLOQUY 10(071) ftp://ftp.princeton.edu/pub/harnad/Psycoloquy/1999.volume.10/ psyc.99.10.071.autonomous-brain.1.milner http://www.cogsci.soton.ac.uk/cgi/psyc/newpsy?10.071
Mogenson, G.J., Jones, D.L. & Yim, C.Y. (1980). From motivation to action: Functional interface between the limbic system and the motor system. Progress in Neurobiology, 14, 69-97.
Mogenson, G.J., Takigawa, M., Robertson, A. & Wu, M. (1979). Self- stimulation of the nucleus accumbens and ventral tegmental area of Tsai attenuated by microinjections of spiroperidol into the nucleus accumbens. Brain Research, 171, 247-259.