Home Page || Public Bookmarks || The Calvin Bookshelf
A book by
William H. Calvin
The Cerebral Symphony
Seashore Reflections on the
Structure of Consciousness

Copyright ©1989 by William H. Calvin.

You may download this for personal reading but may not redistribute or archive without permission (exception: teachers should feel free to print out a chapter and photocopy it for students).


Dynamic Reorganization:
Sharpening up a Smear with a Mexican Hat

[How] we come to analyze the world, without postulating the presence of a non-material central agent or homunculus... is a problem that is rarely faced by neuroscientists.... Perhaps the reason for this serious failure is that neuroscientists mostly think about individual units but not about the
population of neurons and their connections, the synapses of the brain.
the English neurobiologist J. Z. Young, 1988

To learn is to eliminate.

the French neurobiologist Jean-Pierre Changeux, 1983

Surely there is something better than an infinite regress of little people inside the head. Or the infinitely regressing agendas of "reductionism forever," constantly changing so that one never answers the interesting questions about how the mind works.
      Well, one thing that's better is the bookstore. The Shining Sea Bikeway is the only path I know that has excellent bookstores at each end, both in Falmouth and in Woods Hole. No matter which end you start from, you can take a break in the middle of your round trip by sitting down and browsing through the new books. A truly enlightened trail, with some foresightful soul having provided park benches every mile and, not far from each end, a truly civilized bookstore with a readers' table and chairs. And J. Z. Young's new book, Philosophy and the Brain.
      "Jay Zed," after providing the membrane biophysicists with the squid giant axon (he and Keffer Hartline did the first recordings from it here; Hartline was surprised, never having seen such a large electrical signal before, and suspected faulty equipment initially), went on to study the other familiar cephalopod mollusk, the octopus. It is the most intelligent of the invertebrates, sometimes compared to a laboratory rat in its behavioral versatility. And after modeling the octopus's visual memory, he suggested that memory may work, in part, by eliminating some neural connections -- by carving away material, in the same way those carved figureheads for old sailing ships were created.
      This was somewhat heretical: Everybody was used to thinking of growth and memory as a process of adding on material, the way children with modeling clay are likely to make a figurine by padding it out. Neurobiologists are like everyone else, always filling up one file cabinet after another, and when we said that memory was a cumulative process, we usually meant it in both senses of the word. J.Z.'s proposal had the interesting property of suggesting some limit to how much information could be stored with such a process: After all, a carver who doesn't know when to stop soon runs out of wood. If you lived long enough, you might run out of brain! And surely consciousness before that. You can see why the idea didn't immediately catch on.
      Because new memories had to be somehow superimposed on old memories, one imagined smaller figures being carved into larger ones, as in graffiti on a wooden figurehead; from a distance, it would loom like one thing, but focusing up closer would tell a different story. So adding on new memories had to conform in some way to the old memories: after all, carving a whale does make it hard to superimpose the form of a giraffe on it. Reorganizing your memories, in the manner that I periodically reorganize my file cabinets, might be a little hard if the information was stored by carving. Certainly neural carving raised the possibility that some experiences could produce irreversible effects.
      Much to my surprise, carving connections has provided one of the key insights required to understand how consciousness committees function, and for why one doesn't need labeled lines to get Grandmother's face recognition. It's all because of the Mexican hat -- what was originally called lateral inhibition by its neurophysiological discoverers, Georg von Békésy, Keffer Hartline, Floyd Ratliff, Stephen Kuffler, Robert Barlow, and such.

To begin personally, on a confessional note, I was at one time, at my onset, a single cell.... I do not remember this, but I know that I began dividing. I have probably never worked so hard, and never again with such skill and certainty.... At one stage I possessed an excellent kidney, good enough for any higher fish; then I thought better and destroyed it all at once, installing in its place a neater pair for living on land. I didn't plan on this when it was going on, but my cells, with a better memory, did.
      Thinking back, I count myself lucky that I was not in charge at the time. If it had been left to me to do the mapping of my cells I would have got it wrong, dropped something, forgotten where to assemble my neural crest, confused it. Or I might have been stopped in my tracks, panicked by the massive deaths, billions of my embryonic cells being killed off systematically to make room for their more senior successors, death on a scale so vast that I can't think of it without wincing. By the time I was born, more of me had died than had survived. It is no wonder I can't remember; during that time I went through brain after brain for nine months, finally contriving the one model that could be human, equipped for language.
the American physician and essayist Lewis Thomas, 1987

IRREVERSIBILITY has been an important idea about human beings for a long time. The Jesuits thought that they could produce lifetime obedience with suitable indoctrination of little boys; Charles Darwin noted that "a belief constantly inculcated during the early years of life, while the brain is impressible, appears to acquire almost the nature of an instinct." Defense attorneys invoke such reasoning to attempt to excuse their clients' actions, saying that their behavior was all the manufacturer's fault, i.e., the school system and "society" in general. Some psychologists claim that the child's brain is infinitely malleable by the environment in which it grows up -- but usually recognize that adults are more likely to become set in their ways as they grow older. Whatever the extent of the plasticity in various periods, it certainly seems likely that something so important will be carefully regulated, probably on a day-to-day basis (or night-to-night -- one proposed function of sleep is to adjust what is passed from temporary to long-lasting memory).
      Behavioral plasticity may be restricted to some "critical periods" in development; in certain respects, it really matters what your experiences are during some years. If the two eyes don't get a chance to work together within the first two years of life, chances are that one eye will become functionally blind; there will be nothing wrong with it optically, but it will act as if disconnected from the higher centers of the brain (this is the reason why cross-eyed babies need such early surgery).
      Something as well learned as a schema becomes the modal schematic of a series of similar experiences: We no longer remember when we first encountered the letter A, or when we took our first step. A snapshot schema, however, is a record of a unique event (in my generation: what you were doing when you heard about President Kennedy being shot). Unless there is some way of freezing it (the way those infantile memories of binocular experience were frozen by myelination and synaptic editing), we might expect such memories to be malleable. We shape up schemas to be the average shape of a lighthouse or the defining character of a Picasso. The lack of "fixing" means that every time we reactivate the schema (as when we recall it "to mind" in the absence of the original), we stand some chance of modifying the memory.

History is what you remember, and if you don't think it's being revised all the time, you haven't paid enough attention to your own memory. When you remember something, you don't remember the thing itself -- you just remember the last time you remembered it.
the Grateful Dead songwriter John Perry Barlow, 1984

MALLEABILITY IS THE FLIP SIDE of irreversibility: Our memories are sometimes more modifiable than we imagine. Indeed, each time we recall something, we have an opportunity to modify that memory. This means that we can readily fool ourselves, unless we have disciplined ourselves to keep fantasy separate from reality. Even then, "brainwashing" techniques (of which some religious-conversion practices are the most familiar) may persuade us that the exact opposite is true. Ordinarily, human memories are pretty good, and we have come to insist (as a matter of social policy) upon one's responsibility for telling the truth -- but there are no guarantees biologically, just as there are no guarantees against mental illness.
      So how does this malleability, this plasticity, of memory arise -- and how is it occasionally "fixed" to resist further change? There are no general answers yet, but a spectrum of phenomena has been uncovered, ranging from dramatic examples of carving during childhood to subtle modeling of interconnections during adult experience "permitted" by neuromodulators.
      I was reminded of all this when I sat in on the neural systems course and heard Patricia Goldman-Rakic lecture on the frontal lobes (yes, indeed; frontal lobes have finally made it to MBL!). We usually think of the process of growing up as one of growth, of adding on to the body and filling it out with additional material. But development is also a process of removing some material. This isn't the familiar old saw about "you lose ten-thousand nerve cells every day" (the experts now say that isn't particularly true in higher centers, that nearly all the cell death in cerebral cortex is accomplished before you are born, even though some subcortical structures such as the substantia nigra lose a lot of cells during life -- and excess loss is a component of what we know as Parkinson's disease). But there is an evolutionary version of the carving principle, some neurobiologists now saying that new neural structures deep in the brain have been created by a process of removing intervening cells to define their shapes better, that this is how differentiation of "subdepartments" occurs.
      Nerve cells also die during one's lifetime, of course, as when injured by a bruise or the loss of oxygen supply or by escaped hemoglobin (when blood cells rupture, and the arterial walls as well, the hemoglobin may come into direct contact with nerve cells and kill them). These local areas of damage may never be noticed, as adjacent areas take over their function so well. When large areas of brain start losing a lot of nerve cells, however, you start having problems with memory, reasoning, speech, and all the rest.
      There was an old Gary Trudeau cartoon posted on the bulletin board in the student lab for the neural-systems course, from the series "In Search of Reagan's Mind," where the investigative reporter is wandering around inside the convolutions of a seemingly empty frontal lobe, shining his flashlight here and there and exclaiming, "Where is Ronald Reagan?" But neurologists don't find that so funny: Every day they look at magnetic resonance scans (those computerized pictures that seem to slice up the brain; they're particularly useful for judging the size of various structures) from patients who are losing a lot of frontal lobe and getting senile, exhibiting many of the signs and symptoms that Elaine had for a month and then got over.
      Alas, the senile patients with such massive loss don't recover: once brain cells are gone, they're not typically replaced (unlike the blood cells, which are totally replaced every 120 days, or the intestinal lining which is replaced every 3 days). Neurologists look at the brain images and see the infolded cortex developing wide, unoccupied valleys, and they sigh regretfully. In a normal brain, those valley walls are pushed together so tightly that you cannot see down into the groove. Strangely, many magazine photographs of "normal brains" are really the brains of people who had senile dementia -- that's because art directors like exaggerated features that reveal the deep folds, and so they pick the pictures that "look best." But those brains had been carved from within by the disease process that destroyed most of the cortical nerve cells and shrank the brain. You wouldn't want to have a brain like the ones that they, in their ignorance, like to picture.
      Is the disease process an exaggeration of normal developmental processes, something like cancer? Cancer is wild proliferation of cells, adding on more cells uncontrollably to create tumors and infiltrate other tissues, the way that strip cities invade farmlands insidiously. Is this loss of cells in senile dementia a disorder of a later stage of development where editing, not addition, is the dominant feature? That is just one of the many reasons why so many people around MBL are studying developmental processes in biology, trying to learn the normal rules of the game so as to better understand how things go wrong.
      When one sees a process in biology, such as cell proliferation, one is almost sure to see one or more additional processes, such as cell editing, that oppose it. Everything is usually a tug-of-war, with net movement occurring only when strengths no longer balance. Yet the two opposing processes are seldom symmetrical like the tug-of-war with a dozen people on each end of the rope: In biology you find situations more like a winch being used on one end of the rope, people on the other. Push and pull often come about in different ways, and it may not be obvious what is being "balanced."

We learn... that there's a utility in death because... the world goes on changing and we can't keep up with it. If I have any disciples, you can say this of every one of them, they think for themselves.
the pioneer neuroscientist Warren S. McCulloch

EDITING CONNECTIONS during childhood is a much less destructive process than the loss of entire cells during prenatal development or senile dementia. What we are talking about here is not nerve cells dying, but selectively breaking half the interconnections between cortical nerve cells.
      As we grow up, we lose close to half of all the interconnections in our cerebral cortex -- we gained connections until about eight months after birth, but after that comes this net loss. It can be even worse for some long-distance connections, e.g., the connections between the monkey's left brain and right brain decline by 70 percent between birth and sexual maturity. Mammalian brains have connections from all areas of cerebral cortex to the spinal cord at birth but, by maturity, all have been withdrawn excerpt for those from the usual somatosensory, motor, and premotor cortical areas.
      That is a lot to lose; if you'd told me (or any other neurophysiologist) this fact maybe two decades ago, I wouldn't have believed it. Nerve cells seem to start out by making lots of connections with other nerve cells (not quite "everything is connected to everything," but much more widely than anyone thought) -- and then something edits them, disconnects quite a few, shaping up the child's mind by whittling away. There's an important principle here: Make lots of overlapping connections, then narrow them down somewhat -- but not all the way down to unique "labeled lines." The necessity for such a disconnection principle was recognized a quarter-century ago by a philosopher, Daniel C. Dennett, in his 1965 doctoral dissertation at Oxford; he even recognized the analogy to biology's pre-Darwin convergent selection:

What is needed is for some intra-cerebral function to take over the evolutionary role played by the exigencies of nature in species evolution; i.e., some force to extinguish the inappropriate.... This would have the effect of pruning the initially unstructured connections along lines at least compatible with and occasionally contributory to the appropriate inherited links already endowed by species evolution.... The process is a repeated self-purification of function, gaining in effectiveness as more and more not inappropriate structure becomes established.

      Least destructive of all would be simple modifications in the strength of the synaptic connections between cells, being able to diminish the strength to nothing without actually disconnecting it (a "silent synapse"). Maybe physical disconnection is one way of "fixing" the memories encoded by such reductions-to-nothing; if the connection remains physically there, some future retuning of the system might destroy the old memory that relied on the weak connection.
      So there seem to be a variety of ways of editing brains: killing whole cells (as happens in prenatal development and senile dementia, and possibly in songbirds that learn a new song every year), disconnecting some interconnections between selected cells (as happens during the tuning-up to the environment seen in childhood), and simple increases and decreases in synaptic strengths (as certainly happens in short-term memory throughout childhood and adult life). Long-term memories, of the multitrial varieties we call schemas, likely involve both altered synaptic strengths -- and sometimes the creation of additional synapses by an existing cell budding off a new axon branch and attaching to another cell.
      New synapses? All that the childhood halving of synaptic numbers in cerebral cortex means is that there is a difference between the rate at which new connections are being made and the rate at which old synapses are being disconnected. Up to eight months after birth in humans, the creation rate exceeds the destruction rate; afterward, slightly more are disconnected than new ones are formed. But no one knows how many synapses are being destroyed in the average week: All we know is the cumulative difference between creation and destruction rates, which yields a 35-50 percent loss during childhood. We have no way of tracking individual cortical synapses over time in a given animal, though brain imaging techniques that visualize proteins involved in making new synapses should eventually give us a clue about creation rates. Sprouting to make brand-new connections gives us an additional process to modify for memory's sake: We have little idea of how frequently this happens in adult life, or if there are favored sites for sprouting, or how it might be regulated.
      Worrying about fixing snapshot schema may, of course, be needless: There is no evidence that humans were designed by evolution to be faithful recorders of events. It is true that one-trial learning exists, particularly for the tastes of foods that make you sick. But there is nothing about that which says that the memory must remain forever fixed. Modification is probably the rule, not the exception.
Nothing seems more possible to me than that people some day will come to the definite opinion that there is no copy in the...nervous system which corresponds to a particular thought, or a particular idea, or memory.
the English philosopher Ludwig Wittgenstein (1889-1951)

      Information is not stored anywhere in particular. Rather, it is stored everywhere. Information is better thought of as "evoked" than "found."
the American cognitive scientists David Rumelhart and Donald Norman, 1981

POINT-TO-POINT REPRESENTATION is the notion that maps are connected in orderly ways, as by a pipeline from the tip of your little finger to your somatosensory cortex's little finger region. Or from a photoreceptor on the retina to the corresponding place in the visual cortex's map of the visual world that "represents" that direction from the eye. I suppose that it was reasonable to expect this, but we've known for a long time that it wasn't that simple. For example, the visual world is represented by several hundred million photoreceptors in the two eyes, but they have to get funneled down into several million optic-nerve axons. So, we said, maybe the fine grain is only several million points instead.
      But then it turned out that we can detect line spacings that are finer than the spacing between photoreceptors (what is called "hyperacuity"): We're even better than hundreds of millions! How can this be? It is because "Mexican hat" committees do the job, not pipelines. It is a population of cells at work, not just a single cell lighting up while the others keep quiet.
      The only way to understand how information is stored in the brain may be to understand what the information is being used for; that's always been obvious for learning new skills but isn't so clear for our more detached kinds of knowledge such as words. Certainly for the more familiar kinds of computers, you need not understand the program to understand storage techniques -- but processing and storage are all mixed up in nervous systems. The brain circuitry that analyzes the information is likely to be used to store it as well. So all that pruning of synapses is likely to subserve an analysis or performance function as well as a storage function. If many-to-many is the circuitry connecting sensory and performance regions of the nervous system, rather than many-to-one-to-many, we will simply have to learn to think in population terms -- just as Darwin did when contemplating transformations of one animal species into another.
      The most familiar transformations in our everyday experience are associated with hearing: The treble and bass controls on a radio that augment or reduce the high and low ends of the spectrum. Some hi-fi setups even have equalizers so that a half-dozen different parts of the spectrum can be adjusted separately. You transform what's really there into an altered version that is more pleasing. The brain is doing such transformations, and adjusting them, all the time -- but internally, without twiddling knobs.
      Our nervous system is indeed in the business of transforming things, not in the TV camera's "faithful reproduction" business. Sometimes what you see doesn't correspond perfectly to what you feel touching the same objects (so which is "reality"?). What we "see" when we look at a seashore scene is not what a TV camera would record -- it is subtly different because of the transformations taking place that help extract the information our brains need to make decisions.
      Some of the differences from reality are simply called "illusions," as they are unwanted side effects of the processing: Look at a waterfall for a moment, at the waves of water tumbling down, and then look at the trees nearby. The trees will seem to be moving upward! Look between your fingers at a bright light: You will see some little black lines partway between your fingers. They are not interference fringes but illusions called Mach bands, a side effect of a contrast enhancement transformation that occurs at several levels of visual processing in eye and brain. When we say we "see" something, we are simply reporting on one intermediate stage of a multi-stage set of transformations -- probably just the stage that is accessible to our language cortex.
      The transformations aren't always the same: they are adjusted as conditions alter. In the moonlight, one can see pretty well (though in monochrome). But try to read a newspaper in such light and you'll discover that the type is indistinct, as if irredeemably out of focus. Try to catch a thrown ball at dusk, and you'll discover that your visual images are too slow to keep track of ball position. The retina has readjusted some of the inhibitory mechanisms in the retina that enhance both spatial and temporal resolution, choosing to improve low-light sensitivity at their expense. If you were missing those spatial and temporal transformations in the daylight, you wouldn't be able to either to read or catch balls: What we normally "see" is enhanced in some respects, degraded in others, and has unrealistic features added. So much for "reality"!
      And, of course, different animals are tuned up in different ways. Those primitive Limulus wandering around offshore have about ten eyes, positioned at various strategic points around that horseshoe-shaped shell (including one on the tail that specializes in day-night rhythms!); most of the eyes probably use inhibition for contrast enhancement. And Limulus is extraordinarily good at detecting faint shadows even in moonlight, even when two stories deep offshore Stony Beach. Robert Barlow, a second-generation neurobiologist who studies lateral inhibition at MBL (in neurobiology, one can trace genealogies back to just a few pre-1940 workers: Barlow was a student of Keffer Hartline's, whereas my wife and I are third-generation neurobiologists, both students of another student of Hartline's, Charles F. Stevens), says he can swim around in scuba gear at the full moon, and when his dim shadow falls across a Limulus on the dark bottom, the animal will change course. He can make it crawl along a zigzag course by simply casting a shadow on its left side, then its right side, etc

I JUST SAW A FISHING BOAT return with a whole class of MBL students packed into its stern, standing room only. I can't believe they collected anything, as they were only out one hour. And there wasn't room for any fishing gear on the stern, so tightly packed was it with students. Sightseeing? Well, at least they saw the salt water from a boat -- that's more marine experience than most students get here these days.
      MBL is an oddity among marine stations: The staple course taught by nearly all marine labs is comparative invertebrate zoology. But it's not taught at MBL anymore. At other marine research stations, the tide tables are prominently posted and the researchers are likely to go out and collect their own animals from one or another of a fleet of small boats tied up at the dock. Life at such labs revolves around the tides. One sees rubber boots, set out to dry alongside the special nets and traps, diver's tanks and weighted belts lying out on the docks, instructions posted about how to reach the nearest decompression facility in case struck by the "bends." But the Marine Biological Laboratory is rather urban: The animals appear in aquarium tanks, delivered by ex-fisherman employees. With the exception of some people such as Bob Barlow, if you asked the typical MBL researcher when the best low tide of the season was going to be, you'd draw a blank stare. Or maybe, "What tides?" The MBL is now "marine" only at one remove for many of the researchers.
      Some of the research at MBL could now be done in the middle of New York City with the animals delivered by air freight in Styrofoam picnic chests -- an option that was not available in MBL's formative years. There are still many notable exceptions to that statement, such as the biophysics done on the fragile squid and the developmental biology on various eggs of marine organisms, and that research remains the hard core of MBL biology. But there are certainly some researchers around here who could get by on ice chests and couriers.
      So why do they continue coming to Woods Hole? MBL is an expensive item for most researchers, not only for their supporting budgets but personally, as the housing around here is (thanks to the better-paid Boston bankers competing for it) so expensive that researchers' savings accounts suffer. It can't be scenery -- going to the beach is far easier and cheaper elsewhere on the East Coast, and the people who come here to work typically go elsewhere for serious vacations. So what is the real reason that so many researchers still go to all the trouble of crating up their labs, suffering with the rental truck, the sore back, and the equipment that inexplicably stops working when disturbed -- and then repeating the tasks several months later?
      It's that MBL is, scientifically, a very special place quite aside from the setting and the animal availability that were associated with its origins. You learn important things here in three weeks, things that you'd seldom learn back home or at a convention center meeting amid ten thousand milling scientists. People are set up and working here; you can get a demonstration of something ten minutes after you hear about it, check things out in the extensive library after lunch, and try the modification out on your lab rig that afternoon. The rumor mill in techniques and preliminary results is a fundamental part of doing science, even if it is hard to document. Its buildings and animals are vital, but it is as a social institution that MBL is so influential in biological science. No other marine lab has a comparable level of free exchange of important ideas, certainly not built into its basic program of courses, conferences, and rich diet of special lectures.
      One of the pleasures of summer in Woods Hole is the breadth of evening lectures not only on science but on art, history, and public affairs. There are concerts several times a week (the philosopher Geoffrey Hellman just played the Brahms' Intermezzo in B minor last night as the encore to a superb piano concert of Mozart, Berg, and Beethoven). "Try to learn something about everything, and everything about something" was Thomas Henry Huxley's epitaph; he would have liked Woods Hole.
      What Woods Hole has now is, I suspect, partly a legacy of generations of nonworking wives with a lot of time and energy to organize. This was a form of scaffolding, now largely removed as most educated women pursue their own careers, but historians may come to see educated nonworking wives (and the occasional nonscientist husband of an MBL researcher) as essential intermediaries for the present cultural milieu of Woods Hole, what makes the social life here more than a place to meet people, more than the usual buffer to diffuse fatigue and hostilities. They've created a milieu that makes learning something about everything extend well beyond its usual liberal arts boundaries, and often bridge C. P. Snow's two cultures with grace. The number of families around here with second- and third-generation scientist-physician-musicians is one indicator of its success.
      Buildings are buildings, but this feat of social engineering was far harder to achieve. And it is potentially fragile, capable of being wiped out by empire-building government agencies who want to keep "their people" on a short leash tied to Washington, D.C., or by an economy-minded Congress bent on cutting costs by consolidating facilities for "efficiency." MBL has no institutional backing -- it's an independent nonprofit corporation owned by its seven-hundred scientist members, and its finances are always worrisome. But places like MBL are not defense contractors producing products, nor even designers with a definable output in terms of blueprints -- they are think tanks, first and foremost. In the physical sciences, think tanks require an office building, lots of salary money, and much computer time. MBL just happens to be a bioscience think tank that instead requires a small fleet of fishing boats and a battalion of librarians.

WHAT KINDS OF TRANSFORMATIONS take place in brains? Perhaps they will help us understand the brain's versions of blueprints and libraries and computing. The sensory-processing examples are the best known transformations: While vision's lateral inhibition is perhaps the best-studied, the same principles are seen in skin sensation and hearing. Most exhibit a version of those Mexican hat arrangements where one region is excitatory but a surrounding wider region is inhibitory, leading to an optimal spot size -- any larger and the cell becomes uninterested.
      In the moonlight, such inhibition is turned off in some mammalian eyes (certainly cats and probably ours as well) to increase sensitivity -- another one of the things that Stephen Kuffler discovered. After our eyes adapt to darkness, big spots of light are even more effective than the formerly optimal-sized spots -- the cell can no longer tell what size a spot is. Which is why you can no longer read anything with a less-than-headline-sized typeface in the moonlight. After all, in such dim light, a photoreceptor has to give a detectible response to a single photon -- whereas in daylight it is bombarded by a million times as many. Inhibitory surrounds are one way of regulating sensitivity over that millionfold range.
      Originally we called this "surround inhibition." But then those inverted Mexican hat cells were discovered with excitatory surrounds and inhibitory centers -- so we began to talk instead about "center-surround antagonism," or simply "lateral inhibition." Whatever you call it, it's seen at virtually every level of the visual system, from retina to secondary cortical areas. The skin senses use it. Hearing uses it; a cell maximally sensitive to middle C may be inhibited by tones a half-octave above or below. Lateral inhibition's contrast-enhancing transformations require a broad wiring, each cell receiving excitation from a wide area, and inhibition from an even wider area. Or vice versa. funnel
      So is this related to the "everything is connected to almost everything" wiring of prenatal development, with pruning of far-flung connections used to narrow the connections down to a cone? Probably. Certainly the basic center-surround architecture is present by birth in the primate visual system. For at least four stages of processing, each cell's view of the world is from a funnel of converging inputs. The same thing is true of skin sensation: a funnel of excitation, a wider funnel of inhibition. In theory, this should mean that the higher-order cells back in the brain receive from ever-widening areas of skin surface, and potentially can respond to half the body surface (if all the inhibitory surrounds are turned off at each stage!).

REMEMBER THAT CORTICAL MAPS are maps based on the estimated centers of cortical-cell receptive fields, not on their total size (much less their potential size!). If all the inhibitory surrounds are working full strength, receptive-field centers will seem small, and a center point easy to define. And so it won't be too much of an exaggeration to say that there is a point-to-point correspondence of the skin (or retinal) surface to the cortical surface -- that we may, in short, talk meaningfully of a cortical "map" of the sensory surface. But if all the inhibition were turned off, the "map" might be pretty hard to detect because of the gross smear of anatomical connections, almost half of the total sensory surface potentially converging upon each cortical cell.
      This widespread anatomical basis for the much narrower functional specialization is the reason why a cortical cell can shift function, coming to specialize in a different finger than it formerly did. Everyone thought that cortical maps were pretty fixed -- maybe they are different in different individuals, but that they were fixed during the lifetime of an individual. But in the 1980's, we were all shocked to hear (from Michael Merzenich, Jon Kaas, Randy Nelson, and their colleagues) that somatosensory cortical maps were a day-to-day affair, changing size if the hand was exercised more; if a particular fingertip was regularly rubbed on something (say, a casino croupier always fingering the deck of cards with his forefinger), more cells in the somatosensory cortex would come to specialize in that finger. And conversely, the size of the average receptive-field center for a cell specializing in that finger would become smaller. 3rd finger exercise
      Usually when this happened, the new forefinger cells would come from cells that formerly specialized in adjacent fingers -- but sometimes from cells that formerly specialized in the face! The face's connections to such versatile cells were turned down to nothing, while the forefinger's connections were enhanced -- and so a "retrained worker"!
      But they also noticed that some changes in cortical boundaries seemed to occur spontaneously from week to week, even though the monkey wasn't being trained and was just moving about his cage. For example, the boundary between face and hand cells in cortex moved from week to week, back and forth -- some weeks, the cells near the boundary were face specialists, other weeks the very same cells were thumb specialists. To neurophysiologists, this was approximately as if you had told us that the state line between California and Oregon was moving a few miles back and forth from week to week for no apparent reason.
      When the researchers trained monkeys to hold one forefinger against a vibrating surface, they observed threefold increases in the number of cortical cells specializing in the tip of that forefinger. But the other finger boundaries shifted as well. When "California" was overly exercised, the California-Oregon boundary shifted -- but so did the Oregon-Washington and Washington-Canada boundaries! The whole hand map rearranged itself to accommodate the fingertip exercise; the expansion wasn't just at the expense of the immediate neighbors in the usual boundary-dispute manner. Furthermore, it wasn't just the other fingers that were squeezed; much of the expansion was at the expense of the face and wrist representations, as the total "hand" enlarged. Historical trends may have remade the map of Europe over the centuries, but whatever remakes cortical maps can be much quicker. One almost has to think of maps as ephemeral, about as permanent as the arrangement of papers atop my desk.
      So learning to play the piano probably does remake your brain, in a very real sense. But so too might almost any other activity that repeatedly stimulates a hand (or, presumably, foot). I told an anthropologist friend about this, because she always goes barefoot -- her feet receive much more detailed sensation that way than when encased in a shoe. Does that increase her foot representation at the expense of other body parts? Does it decrease her other sensory abilities? No one knows yet. Is this plasticity why many stroke victims get much better in the weeks and months after their strokes, with uninjured regions of cortex taking over the jobs originally done by the injured cortex? Is this why blind people can hear more acutely? Tune in next year.

CORTICAL MAPS ARE EPIPHENOMENA anyway, since they serve no known purpose other than as guides to the neurophysiologist in the placement of recording electrodes. It is the nerve cells themselves, like the industries of a city, that are the functional pieces; while street maps do serve a function for strangers to the city, I cannot think of a comparable function served by the brain maps we produce. What is so interesting about these ephemeral maps is that they indicate that there is a lot of retraining of workers in the brain, that the number of cells assigned to a task (such as analyzing sensation from the forefinger) can be modified on a week-to-week (and probably a day-to-day) basis.
      And most of the plasticity seems to be at the cortical level; the maps of the fingers in the monkey's thalamus, the relay station just before cortex, are not altered in a similarly dramatic way. The thalamic nerve cells specializing in the forefinger, however, seem to send axon branches all over the hand region of somatosensory cortex, not just to the forefinger's current patch of cortex. This is what probably allows the rapid retraining of a cortical cell: It just switches from suppressing everything except forefinger to suppressing everything except thumb. Those cells that switch back and forth between thumb and face presumably have anatomical connections from both, with one or the other set suppressed. The alternative explanation, that the thalamic axons sprout new connections and that the old synapses disconnect, is not ruled out -- but the rapidity of the changes is faster than such sprouting processes usually happen in the nerves to muscle.
      Dynamic remapping, going on all the time, suggests that there is an ongoing competition of some sort that results in the work getting spread around to the available workers. Rather than a lifetime structure, it may be more like a free-wheeling economy -- perhaps a certain rate of neural unemployment is used to make sure niches are explored and filled. Are there monopolistic practices, used to seize power and keep down the newcomers? Is there some central direction, a circulating hormone that functions like a five-year plan to steer migrant workers in some directions more than others?
      And while economic analogies are perhaps more familiar, it seems likely that the true analogies are going to be to things even more primitive than economics: Self-organization, darwinian evolution, and ecosystems. The familiar computer analogies to which we retreat when seeking an analogy for brain functioning seem totally inappropriate: Computers have a memory that is kept separate from the processor; they do what they are told to do, rather than seeking out new niches.
      Ever since Darwin and Wallace, we've known that we need to understand evolutionary principles in order to understand how we came to be, that long road from monkey to ape to hominid to human. For almost as long, the analogies between ontogeny and phylogeny have made us aware that darwinian-like processes (such as all that prenatal cell death) are a major part of getting from fertilized ovum to an adult. But now it looks as if anyone who wants to understand day-to-day brain functioning (the nature of perception and cognition, the basis of memory, the organization of behavior) had also better bone up on Darwin.

What we call a mind is nothing but a heap or collection of different perceptions, united together by certain relations and suppos'd, tho' falsely, to be endow'd with a perfect simplicity and identity.
the Scottish philosopher David Hume (1711-1776)
Philosophy has succeeded, not without a struggle, in freeing itself from its obsession with the soul, only to find itself landed with something still more mysterious and captivating: the fact of man's bodiliness.
the German philosopher Friedrich Nietzsche (1844-1900)
The Cerebral Symphony (Bantam 1989) is my book on animal and human consciousness, using the setting of the Marine Biological Labs and Cape Cod. AVAILABILITY is limited.
Email || Home Page || Table of Contents || Endnotes for this chapter || continue reading Next Chapter