|A book by|
William H. Calvin
UNIVERSITY OF WASHINGTON
SEATTLE, WASHINGTON 98195-1800 USA
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).
Making Mind from Mere Brain:
Taking Apart the Visual World
Indeed they were very close to the Lighthouse now. There it loomed up, stark and straight, glaring black and white, and one could see the waves breaking in white splinters like smashed glass upon the rocks. One could see the windows clearly; a dab of white on one of them, and a little tuft of green on the rock. A man had come out and looked at them through a glass and gone in again. So it was like that, James thought, the Lighthouse one had seen from across the bay all these years; it was a stark tower on bare rock.the novelist Virginia Woolf, 1927
Our brain is domineering when it comes to coping with reality, and so we sometimes see things not as they really are, sometimes invent categories that do not exist in nature, sometimes fail to see things that are really there. Romanticizing things is the least of the problems.
It's not so much that we only see certain wavelengths of light, only hear a limited range of frequencies, have an impoverished sense of smell compared to most mammals. But once accepted by the sensory receptors, information may be forced into a Procrustean bed created by our expectations. We may force reality into a certainty and definiteness that it doesn't naturally possess. If the information is probabilistic, we make it definite. You see this at the level of physics (the discoveries of quantum mechanics have backed us up somewhat, out of excessively rigid human categories for nature) and you certainly see it in how the brain processes the information that gets in.
We also ignore repeated inputs, tend to respond only to the new and unexpected. Though our brain asks the questions, it is more interested in answers that are new and odd. We see contrasts and borderlines, hear changes and movements. Prevent the eye from moving, and it will see nothing unless the object itself is moving. What does that say about "consciousness" if an object can disappear, returning only if it moves?
While attuned to categories, our brain seems to seek gestalts (and sometimes finds them even if they aren't there!). It constructs scenarios, most of them sheer nonsense and discarded the moment they are compared to our memories of the real world; as we sleep and dream, that testing isn't very efficient, and so this fantasy is largely unedited. Awake, we shape up our scenarios by additionally testing them against reality. And these scenarios, when trying to explain the past by constructing a story about it, may also prove useful in predicting the future.
So why is the brain so simultaneously Procrustean, but also inventive of the categories into which the incoming information is impressed? The nature of reality, as we perceive it, depends on what it is filtered through.
It is, of course, filtered through the whole nervous system, a distributed collection of cells extending a meter and more in length. True, there are some systems that extend over only a tenth that distance, such as the 10 cm of the visual system. Within such systems are multiple maps, which compress half of a sensory world into about 1 cm worth of cerebral cortex. Within each map are distributed many networks, each contained in about a cubic millimeter of tangled nerve cells. Those cells, the neurons, stretch over various distances, though many are contained in about 0.1 mm. The synaptic connections between them are a hundred times smaller, and seem to be the sites of modification, where memories are recorded. And all this is effected by molecules at least 10,000 times smaller than a synapse, some of which form structures such as membranes (and channels through the membranes), others of which move like messengers through the membranes.
So where in all this does my memory of a lighthouse reside? Or my plan to go there? Or my one-foot-in-front-of-the-other subprogram?[What] you are describing is not an object but a function, a role that is inextricably tied to some context. Take away the context, and the meaning also disappears.... When you perceive intelligently, as you sometimes do, you always perceive a function, never an object in the...physical sense.... Your Cartesian idea of a device in the brain that does the registering is based on a misleading analogy between vision and photography. Cameras always register objects, but human perception is always the perception of functional roles. The two processes could not be more different....IT IS CURIOUS HOW DISTRACTING little black lines in the sky prove to be, as if they had some privileged access to our brains -- like visual equivalents of advertising jingles hung from utility poles. And they do have priviledged access, another discovery that Steve Kuffler had a hand in.the mathematician Stanislaw Ulam, about 1970
Leaving the Church of the Messiah, one walks east along a winding road in Woods Hole that would be beautiful except for the proliferating poles stringing the skyline with metastasizing threads of metallic tumor. It's an area of expensive homes whose owners could afford to pay for burying the utility wires, but only people get buried around here. In another block down Church Street, the wires thin out and the view opens onto Vineyard Sound, a long, sandy beach stretching into the distance. At the end of the beach, a white pillar rises into the blue sky, capped with black. It is surely the Nobska Point lighthouse, and it looks somewhat familiar.
A lighthouse is a version of the scarecrow, but meant to warn off sailors. And just as the scarecrow is a human mimic of sorts, so the lighthouse stands as a surrogate human, erected to send a human message in times of need. As lighthouses get "smarter" with the addition of automatic fog sensors that trigger their foghorns, they come to be more and more like a stationary robot. Their foresightful designers have endowed the lighthouse with ways of forecasting trouble, and appropriately responding. Lighthouses lack the versatility to be considered conscious, but the air-traffic control system might someday evolve into a partially conscious robot, constantly trying out "what if" scenarios and so heading off collisions between airplanes in much the same ways as human controllers now attempt to do.
WALKING UP THE BEACH IN THE WET SAND just above the lapping waves of Vineyard Sound, I saw various people from the MBL and WHOI, talking science. Bankers from Boston can also be found on this beach, though they do not draw diagrams of cells in the sand, so far as I know. The diagrams may be washed away by the next high tide, but the new concepts linger on. This beach has seen a lot of science happen, as has Stony Beach near MBL, which Lewis Thomas immortalized in his essay "The MBL," in Lives of a Cell:It is so crowded that one must pick one's way on tiptoe to find a hunching place, but there is always a lot of standing up anyway; biologists seem to prefer standing on beaches, talking at each other, gesturing to indicate the way things are assembled, bending down to draw diagrams in the sand. By the end of the day, the sand is crisscrossed with a mesh of ordinates, abscissas, curves to account for everything in nature.This beach on Vineyard Sound is instead spread over two blocks and has a higher proportion of sand castles than graphs. The lighthouse down at the end of the beach, as one walks closer and the perspective changes, evokes ever stronger sensations of déjà vu -- I seem to have some mental image that this picture is triggering.
You can hear the sound from the beach at a distance, before you see the people. It is that most extraordinary noise, half-shout, half-song, made by confluent, simultaneously raised human voices, explaining things to each other.
And then I remember the art museum where I've seen it before: it's a close relative of the lighthouse that Edward Hopper painted. And painted again and again, though they might be different lighthouses, because the coast guard built a number of them from the same 1870's design; they can be found all around Cape Cod and the offshore islands. This lighthouse looks even more solid than the painted versions, ready to take whatever the weather dishes out next.
I HAVE AN EDWARD HOPPER SCHEMA somewhere in my head, not to mention a Picasso schema, and a Henry Moore schema, and so on. A schema is a mental image. While schema was initially used to avoid using memory trace (another term invented to try to pin down an elusive concept), a schema is not just any memory. A sensory schema is a schematic outline that fits most of the likely variations that Hopper constructed; it fits in a fancier way than a cookie-cutter fits a Christmas cookie, but that's the general idea.
A collection of schemas -- say, my collection of mental images of artistic styles -- is something like a family of cookie-cutters, each shape being tried out on a particular Christmas cookie to see which comes closest. The collection of schemas seems to sit there in your head, constantly on the lookout for something that fits it. So that when you see a fragmentary Picasso sketch, and something shouts "Picasso!" inside your head, you've heard a schema speaking.
I also have a lighthouse schema, equally well activated by English lighthouses and Hawaiian lighthouses. It is only the lighthouses that are in the local 1870's style that additionally activate my "Edward Hopper" schema, and then only when seen from a perspective similar to the ones that Hopper favored (aerial photographs won't do). Such an "Edward Hopper schema" is a collective memory, containing something of the "essential" features of his style and subject matter that come through whatever variations are present in a particular example. For every word in your vocabulary, you probably have one or more schemas -- though I certainly have some schemas for which there is no word in my vocabulary, such as people that I've seen before but whose names I haven't learned yet. And some schemas aren't objects but movements: When we construct a sentence or spin a scenario, the units that we are chaining together are primarily sensory ("noun") and movement ("verb") schemas. You, me, food, rock -- as well as run, walk, gallop, shuffle, touch, bend, break.
EPICURUS AND LUCRETIUS considered mental images to be simulacra, and Aristotle compared them to "the imprint left by a seal on a wax tablet." They hadn't seen cookie-cutters. And now we have some even better analogies -- indeed, potential mechanisms. The lighthouse beacon winking on and off (even in broad daylight, the 150-watt light bulb looks bright because of that old 1825 Fresnel lens that focuses its photons) reminds me of another Steve Kuffler classic. His 1953 work was on how nerve cells in the eye respond to such lights winking on and off; I'm unlikely ever to forget it, as reading it about six years after it was published was one of the major reasons that I had forsaken physics for physiology. It is still a classic, required reading for the aspiring graduate student. It demonstrates a cellular version of the schema -- or at least one of its building blocks.
The story starts even earlier, in 1938, when H. K. Hartline studied frog retinas and their response to winking lights. He found that each optic-nerve "wire" responded to far more than just one spot on the retina -- it was a whole patch (which he called the cell's "receptive field"). It would fire a train of spikes in a staccato manner whenever the light was turned on, but then settle down, not acting much different than in darkness. But when the light was finally turned off, it gave another burst of activity, a cellular version of "Hey! Who turned the lights off?" This OFF-response was a puzzle, though not unrelated to the problem of feeling your wristwatch's absence just after you remove it from your wrist and all those flattened hairs start popping back up.A friendly caution: You are approaching some conceptually difficult material in the remainder of this chapter, and at places in the following two chapters. Ordinary mortals will likely, at some point, feel disoriented -- and for good reason. Everyone who has learned this material has complained too. Graduate and medical students traditionally stumble on "receptive fields" -- but then, like art students who finally learn how to "see" a particular painting, recover and soon sort out a mental picture of how a brain cell sees the world. The "viewpoint" turns out to be all-important, both the cell's viewpoint and the scientist's varied viewpoints used to analyze the cell's properties. We're still debating how a committee of cells, like a jury of critics, can have a collective viewpoint! None of this material is essential, in that you could skip over it and still understand the rest of the book, but I include it because so many of those who come to understand it are, let us say, inordinately impressed. It provides a nice, concrete example (almost too nice!) of how the brain goes about its business. Indeed, to those familiar with business accounting, the time history of the cell (such as those OFF-responses) is like cash flow tracking, and the Mexican hat "receptive field" represents another way of looking at much the same data, showing where the income and expenses originated in a given time period and how much profit was paid out to shareholders. Anyone familiar with accounting practices may have a better background for appreciating this material than the typical neurobiology graduate student.
This "Somebody turned the lights off" report (an OFF-response) is an exaggerated aspect of "temporal contrast," a sensitivity to changes in light level rather than absolute light level. It is why flashing or moving lights grab your attention better than steady lights. Keffer Hartline (another luminary of MBL) also noted in passing that the background was important in influencing the vigor of these staccato responses, rather as the brightness of the sky behind the lighthouse affects how bright the beacon seems. Robert Barlow (once Hartline's graduate student, and now a veteran of the Woods Hole science scene) noted this spatial-contrast effect in 1953, as did Stephen Kuffler. Temporal contrast and spatial contrast turn out to be the major building blocks of visual perception, with color contrast an important added feature in some animals such as the primates. Our ancestors made their living by being able to spot ripening fruit high up in the trees against a background of waving leaves; that green spot you sometimes see from the Nobska lighthouse alongside the setting sun is one of the collateral consequences.
Kuffler dug deeper and wound up showing the underlying mechanism of spatial contrast: each of the cat's retinal ganglion cells (which are typically third in the chain of cells between photoreceptor and brain) was receiving input from thousands of photoreceptors (via intermediate cells called bipolar cells and amacrine cells), rather as a funnel collects raindrops from a wide area and concentrates it into a narrow stream. There has to be some funneling, as there are about a hundred photoreceptors for each "wire" back to the brain (the retinal ganglion cell axon; about a million of them constitute an optic nerve). And it isn't a matter of a hundred photoreceptors connecting to each ganglion cell: Messages from thousands of photoreceptors are funneled into each ganglion cell, except some cancel out the actions of others.
Each of those little "wires" in the optic nerve sends messages akin to a bank statement when it tells you how much interest was paid this month. You have to imagine, instead of an eye, a giant bank that is busily mailing out a million statements every second. What maximizes the payout on a single wire? That depends on the bank's rules, and how you play the game.
The plus-and-minus, push-and-pull arrangement (which neurophysiologists tend to call "excitation" and "inhibition" instead) is not unlike the deposits to and withdrawals from my bank account. The output (the interest paid out) is naturally proportional to the net balance. There is often a minimum requirement ("5 percent interest paid on all balances over $1,000"). Nerve cells don't compound interest (since they pay it out each time) -- but they have other nice features that bank accounts lack. I sure wish that my bank account paid an "adaptation bonus" whenever I resumed depositing money after a fallow period! Some nerve cells even have "OFF-responses" as a rebound from inhibition, equivalent to a bank paying a bonus when you stop a long series of withdrawals. I must remember to suggest that to my bank; they can advertise it as "the brain's way of doing business."
Kuffler discovered the parts of the retinal mosaic from which "deposits" and "withdrawals" each originated, something that Hartline couldn't resolve in frogs. A single retinal ganglion cell really receives two funnels, a wide funnel and a more potent narrow funnel; one inhibits and the other excites the retinal ganglion cell. It's not unlike the sources and sinks of a checking account: We usually funnel into our account a small number of large deposits, and write a much larger number of small-amount checks.
A uniform illumination of the back of the eye, such as one gets from looking at the blue sky, stimulates both funnels, and the messages cancel much of the time. A small spot of light such as the lighthouse beacon might fill the small funnel but only half fill the larger one. If the large but weak funnel was connected with inhibition, and the small but potent funnel with excitation, the net difference would cause a vigorous response. Fill up more of the large but weak inhibitory funnel by using a larger spot of light, and the excess of excitation over inhibition would decline and the staccato response would decrease to a whimper.
The overall result after subtraction looks something like a doughnut, with a center and a surround. An even better analogy is the Mexican hat, peaked in the middle and flanked by wide shallow troughs: That's the size of the interest paid out for a small spot of light at each position in space. When you adapt to darkness (your night vision improves after about twenty minutes in the dark), that wide inhibitory funnel is disconnected from the retinal ganglion cell as a way of "running wide open", getting maximum sensitivity by not subtracting anything. Again I wish that my bank would let me boost my interest payments by simply not writing any checks for a few months. Disconnecting the inhibitory funnel in dim light doesn't similarly result in big problems later, unless you count the momentary blindness that occurs when you turn on the bathroom light in the middle of the night (bathroom lights were not a big problem as humans evolved during the ice ages).
So such a cell is a specialist in small spots of light -- indeed, the diameter of the optimal spot is about the size of the hatband in that Mexican hat analogy. Such a cell discriminates against larger spots, hat-brim width or more, sometimes refusing to tell the brain about them at all. But that's all right, as there are other nearby cells specializing in big spots, having a somewhat different arrangement of funnel sizes, just as there are hats with narrow brims for fat heads. We tend to start thinking about such cells as specialists like cookie-cutters of a certain size, just sitting there trying out each new spot of light to see if it's the right size. But the Mexican hat arrangement is the better analogy, as it tells you what would happen to spots that weren't quite the right size.
Now some bank accounts don't have the few-deposits-but-many-checks characteristics of my checking account; they're more like a small business that receives many mail orders and has only a few suppliers to pay. Some retinal cells are like that, with their funnels hooked up just the opposite of the previous examples. Their weak but wide funnel is excitatory, and their narrow but potent funnel is inhibitory. And so they seem to specialize in black spots on white backgrounds rather than white spots on darker backgrounds; they're better at detecting the insects on the sand beach (and the black periods ending sentences) than the white spots such as lighthouse beacons against the night sky. This yields an inverted Mexican hat curve: A black spot imaged onto the inner of the two-funnel arrangement doesn't cause any inhibition, and so the excitatory response from the white background imaged onto the larger funnel has nothing to balance it. Thus one gets a vigorous staccato burst to a black spot on a lighter background.
The map (such as the Mexican hat) of what's connected to a single cell is what is now called the "receptive field" of that cell -- but one can also think of it as a template, the cell always being on the lookout for images that fit its favored template. Just as cookie-cutters the size of the smaller funnel won't fit a larger cookie, so receptive-field centers (the hat-band size) tend to represent the optimal size of a stimulus spot. Specialists in small black spots on lighter backgrounds are sometimes called "bug detectors," because they spring into activity when the image of a fly is presented to them. But they also respond pretty well to thin black lines and usually to edges that cover only half of a funnel. So they are not exclusively "bug detectors": The moving black bug is simply their optimal stimulus. If you think of the bug as the "essence" of the cell's function, you miss out on the whole range of less optimal, but still effective, stimuli that also run the cell.
In 1959 it was reported by Maturana, Lettvin, Pitts, and McCulloch that frogs were even more discriminating. Their retinal ganglion cells didn't bother telling the frog's brain about anything that wasn't moving food like flies, or a tilting horizon, or some other key feature of their environment. Presented with a novel visual feature like a house or car, they often didn't bother mentioning it at all to the brain (unless, of course, it was moving quickly toward them, something that every visual system manages to sense). The frog's retina seemed to be a special-purpose computer, not a TV camera that indiscriminately reports everything it sees. That was troubling: Why was the frog throwing away all that more detailed information?
Doesn't he realize that he'll need it to read with, when he "grows up" evolutionarily? This is, of course, our point of view; the frog is interested in more immediate things, like catching flies for dinner, and has apparently tuned up his eye to maximize the detection of little moving black spots. Our point of view might seem pretty parochial to the frog too. It would have a hard time understanding our reductionist preoccupation with taking things apart into ever-smaller pieces: discriminating between the fly's wings and feet might seem a little excessive to the pragmatic frog.Much of today's public anxiety about science is the apprehension that we may be overlooking the whole by an endless, obsessive preoccupation with the parts.SPRINGY SAND? Walking on the sand beach on the eastern side of the lighthouse can be like walking on a forest floor carpeted with pine needles where the floor gives and rebounds a little. But here on this sandy beach one positively bounces. Why?the American physician and essayist Lewis Thomas
Behind the holists' very proper concern with the need to consider the system as a whole lies also a fear that reductionism attempts not to just explain but to explain away; that by reducing humans to an assemblage of working parts, their humanity has been in some measure also reduced. This is an important fear, and one which should be treated with respect.the British neurobiologist Steven P. R. Rose
The "underlying mechanism" is not hard to discover, since it is literally underlying and can be uncovered by digging down a little with one's toe: Great masses of seaweed have been washed ashore during a recent storm, some stuffed with little air bladders that keep them floating up near the light just under the ocean's surface. Then a windstorm washed sand up the beach to cover up these mats of cushioned seaweed. And so one bounces along the sand like some improbable ballet dancer.
Eventually, the seaweed cooks beneath the hot sand. Vegetation that gets buried deeper can turn into oil; our cars are propelled on the cooked vegetation of great tropical swamps that got themselves buried many millions of years ago. That some of the great oil fields are now located up above the Arctic Circle shows you how much the continents drift around, as all that oil was originally buried when it was within about 20° of the equator.
This also shows you what scientists mean by "reductionist" approaches: Springy sand attracts a scientist who wants to know what the resiliency is all about. Its underlying mechanism, discovered by "digging deeper," is buried seaweed. Then the agenda shifts: Why is the seaweed bouncy? Its underlying mechanism is little air bladders. And why is that? It's due to the way light is filtered by water, making things dark in the ocean depths: If you want to photosynthesize, you want to stay as close to the surface as possible. And hence built-in flotation devices. Why photosynthesize...? And so forth, with a constantly shifting agenda that maintains little of the original question about resiliency (which, after all, was answered right off).
There are a lot of scientists who started out hoping to understand the mind who haven't been able to answer such questions right off, and so have gotten distracted by underlying mechanisms and the subtle meandering of the agenda. The only way to understand something thoroughly is to investigate all its parts, but sometimes that isn't enough: Sometimes you have to figure out how they work in committees, as when they regulate your body temperature or blood pressure to remain within a narrow range of desirable levels. And so "understanding how we see something" may involve a lot of analysis of the parts, but also figuring out how new properties emerge from combinations.
Understanding emergents is generally harder than understanding the pieces, and an imperative of science (or at least of scientists continually faced with raising money and training graduate students) is to "make progress" via asking answerable questions. And some of our most respected scientists are the ones who switch from subject to subject, constantly exploiting the new techniques for "looking deeper." Only if they are disabled from laboratory work by illness or retirement do they ever get around to "trying to piece things together." Yet the history of science is written by both types, the experimenters like Hooke, Franklin, Boyle, and Faraday, and the great synthesizers like Newton, Mendeleyeff, Einstein (Neils Bohr had to think for years before coming up with the model of the atom that proved so successful). In eighteenth- and nineteenth-century biology, the great synthesizers were even more prominent: Linnaeas, Lamarck, Darwin. Neurobiology is young, and its history is mostly the history of the meandering reductionists.
The important thing in science is not so much to obtain new facts as to discover new ways of thinking about them.VISUAL PERCEPTION MECHANISMS had been a mystery before all the wire-tapping neurophysiology on frogs, cats, and monkeys; optics had yielded to investigation, and some shrewd guesses (particularly Thomas Young's 1802 trichromaticity theory) had been made about underlying color vision mechanisms, but no one had shown how the brain went about taking apart an image into its component parts.the English physicist William Lawrence Bragg (1890-1971)
Many students of animal behavior have become so fascinated with its directedness, with the question "What for?" or "Toward what end?" that they have quite forgotten to ask about its causal explanation. Yet the great question... "How?" [is] quite as fascinating as the question "What for?" -- only they fascinate a different kind of scientist. If wonder at the directedness of life is typical of the field student of nature, the quest for understanding of causation is typical of the laboratory worker. It is a regrettable symptom of the limitations inherent to the human mind that very few scientists are able to keep both questions in mind simultaneously.the Austrian ethologist Konrad Lorenz, 1960
Now we had an impressive explanation of how the nervous system went about that, breaking the image into components via how the excitation and inhibition were wired up. And a flurry of new research started, analyzing the next stages of analysis back in the brain proper. We were particularly interested in a visual area at the back of the head, which is where those messages in the optic nerve seemed to be addressed. If that area, called "primary visual cortex" or "Area 17" or sometimes just "V1," wasn't working because of damage, people seemed to be functionally blind.
Two of Steve Kuffler's immediate followers were the young ophthalmologists-turned-neurophysiologists, David Hubel and Torsten Wiesel. When Kuffler went to Harvard Medical School in the late fifties, he hired the pair -- and then went back to studying his first love, synapses (in 1961, together with Josef Dudel, he discovered a major new principle we call presynaptic inhibition, which results in a percentage-type computation more like multiplication or division than the usual excitation-inhibition analogous to addition-subtraction).
Kuffler had worked on the third layer of cells in the chain between photoreceptors and brain (he'd have worked on the photoreceptors and second-order cells, but they were too difficult with 1953 techniques). Hubel and Wiesel discovered a gold mine when they started looking at the brain structures farther along in the visual pathways, the fourth-order cells in the thalamus, and the fifth- and higher-order cells in the cerebral cortex. Other neurophysiologists were busy investigating how sensations from the skin and ears were analyzed in the brain, but the same techniques applied to the visual pathway had an important advantage: The visual system, more so than any other of our senses, has been streamlined by one evolutionary adaptation after another. And we humans tend to think visually, compared to most other animals, giving human neurophysiologists an advantage in trying to understand such stages of analysis of what the eye initially "sees" with the first-order photoreceptors.
We understood how cameras worked, and how television cameras worked, and some of us probably expected the eye to produce a picture of what it saw for presentation to a viewer back in the brain somewhere. Though we knew that the fine mosaic of about 100 million photoreceptors receiving the optical image had to repackage some of that information, because there are only a million "wires" leading back to the brain, we somehow still expected that a faithful image would be recreated at the other end, much as a television tube recreates what the TV camera saw and encoded in radio waves.
But there was that troubling result from the frog studies, all that more-detailed-than-a-fly information that was simply thrown away. Kuffler had fortunately studied the cat's retinal ganglion cells rather than the frog's (he needed a big eye because of the size and complexity of the experimental apparatus); clearly, they were at least reporting the boundaries between all objects. Potentially, the cat's brain ought to be able to read a newspaper. And as Hubel and Wiesel soon showed, a monkey's retinal ganglion cells were rather like a cat's, except for being better at fine resolution and at color. What, however, happened in the brain cells to which the optic nerve reported? The main line from eye to brain goes through the thalamus, whose lateral geniculate nucleus is an elaborately layered structure (bent knee-like, hence geniculate) that had long intrigued neuroanatomists. Surely something fancy happened there.
Hubel and Wiesel recorded from those geniculate cells. And at first glimpse and even second glimpse, they were indistinguishable from the retinal ganglion cells upstream from them -- nothing new seemed to be happening, as if they were just some sort of relay station (this terminology is borrowed from the old Pony Express, where tired horses were exchanged for rested ones, the messages in the saddlebags passed on unaltered). The geniculate's receptive fields were the same center-surround doughnut shapes as the dual-funnel arrangement had produced in the retinal ganglion cells. About all that the first-pass analysis of the geniculate showed was that diffuse light was an even poorer stimulus than it was to the retinal cells: Excitation and inhibition tended to totally cancel out in the geniculate cells, whereas retinal cells often had some net excitation or inhibition when simply looking at an unbroken field like a cloudless blue sky. Not until one uses colored lights as stimuli does one discover the geniculate cells doing something that retina doesn't do: In some cells, the center-surround organization disappears.
When Hubel and Wiesel looked up in the cerebral cortex at some of the still higher-order cells in the processing chain, things were obviously quite different than in the retina or geniculate. Cortex was no relay station; the messages were rearranged here to make new patterns. Most obviously, the circularly symmetric receptive fields of the cells feeding into the cortex were somehow modified to make elongated receptive fields. The optimal stimulus for retinal and geniculate cells was a white spot on a dark background (or in other cells, a black spot on a light background) of a certain size -- but round, always round. In the cortical cells, round spots might evoke a response, but the best stimuli were lines and elongated edges.
Now given that switching from spots to lines and edges decreased the responses of the geniculate cells forming the input to the cortex (compared to the optimal spots), better responses to lines was a surprise. But a given cortical cell arranged those inputs so as to sum together the activity of many input cells whose receptive field centers were not all in the same place: Indeed, they were all strung out in a line, thanks presumably to the genetic instructions that had wired up the thalamus and cortex during prenatal development.
While some cortical cells obviously preferred horizontal white lines, others preferred vertical ones. And others preferred various in-between angles, so long as they stayed within about 5° to 10° of their optimal angle. As the tilt of a line shifted (as, for example, the horizon tilts when the Vineyard ferryboat rolls in a wave), one group of cells fell silent but another group sprang into activity. Each cell seemed to be a specialist in a particular tilt.
And while some cells were best at edges such as the sea-sky boundary, some preferred narrow black lines, and others liked narrow white lines (just as one might predict from the two major classes of their input's Mexican hat arrangements, regular and inverted). So the cortex had a lot of variety, but it was orderly: Neighboring cells tended to have the same preferences for tilt, until suddenly reaching a neighbor that liked a quite different tilt. In all of the layers of the cortex, the tilt preference was the same -- so if one sampled one neuron after another, they would all like the same tilt. But sometimes the recording probe would stray into an adjacent "column" that preferred a different tilt -- and so we began to talk about "orientation columns" in the cortex.
Because a cell would respond somewhat to a small spot of light, one could patiently explore its receptive field, making a "map" of it from which one could predict the optimal stimulus: a narrow black line at 45°, or an edge oriented vertically, and so forth. And when they tried out the predicted optimal stimulus, the cell indeed sang in its loudest voice. But some cells refused to respond to small spots, and so Hubel and Wiesel tried lines and edges -- and they'd usually work, once they twisted them around to the preferred tilt. It was as if these cells simply had a high threshold, as if they require a ten-thousand-dollar balance before paying interest. Spots simply aren't big-league hitters by the time that the information gets to the cortex, according to this interpretation.
Then they got another surprise: Some such cells would respond to the line even if they moved it sideways somewhat. For every cell in retina, geniculate, and (so far) cortex that had been examined, that maneuver would cause the cell to become uninterested, shutting up because the stimulus had wandered out of its receptive field center or off onto a weak portion of its periphery. While they couldn't move the tilted line simply anywhere on the retina and have it work, there was a region perhaps 10-15° wide where the cell still responded to the line. But try tilting the line away from its favored orientation, and the cell would promptly shut up, no matter where the line's center was located. Hubel and Wiesel called these "complex cells," and the ones with the predictable maps "simple cells."
This was an exciting development, as the complex cells appeared to be "generalizing" on the concept of "line tilted at 45°" regardless of location. Psychologists had long been pushing generalization as a difference between lower and higher animals: Some species will learn to treat an erect and an upside-down triangle as "the same thing," and other species will always treat them as different, refusing to "generalize on the concept of triangle." Complex cells were generalizing, not about triangles but about one of the parts of the triangle, a tilted line.
So were there higher-order cells in the brain that specialized in triangles, no matter what their size or tilt, no matter whether black-on-white or white-on-black, no matter whether the center was filled or not? Were there triangle detectors in higher animals' brains, just as frogs had good "bug detectors" even out in their retinas?
We know already how to hear isolated cells among the millions of elements in the system. I cannot imagine a more important task than the reconstruction of the symphony.the neuropsychologist Hans-Lukas Teuber
STEPHEN KUFFLER NEVER WON the Nobel Prize, despite his involvement in this string of successes in uncovering one fundamental aspect after another of how inhibition is used. The betting among the neurobiological community in the late 1970's was that the next Nobel Prize in matters neurobiological would go to the triumvirate of Kuffler, Wiesel, and Hubel. But Nobel Prizes are never awarded posthumously, and Steve Kuffler died while working at his desk in his Woods Hole home on 11 October 1980, at the age of sixty-seven. As was his custom, he had been swimming at Stony Beach that morning.
In 1981 Hubel and Wiesel did indeed receive the Nobel Prize in Physiology or Medicine. And as in all such matters, they were particularly noteworthy representatives of a whole community of successful scientists. Neurobiology grew up in those years: There was no such word in 1959, just various neurophysiologists and neuroanatomists and neuropharmacologists and developmental specialists, etc., coming from backgrounds in medicine, biology, psychology, and the physical sciences. Today, students can find a variety of graduate programs in neurobiology, and there are even a few places with undergraduate majors in neurobiology. But mostly we remain a diverse group of emigrants from other fields, something of a melting pot. Everyone learns the reductionist techniques, as they remain the best way to train graduate students, but a few wander off into trying to piece it all together, to see the big picture, maybe discover some emergent properties.
What we get on the retina, whether we are chickens or human beings, is a welter of dancing light points stimulating the sensitive rods and cones that fire their messages into the brain. What we see is a stable world. It takes an effort of the imagination and a fairly complex apparatus to realize the tremendous gulf that exists between the two.
Consider any object, such as a book or piece of paper. When we scan it with our two eyes it projects upon our two retinas a restless, flitting pattern of light of various wave lengths and intensities.... But we are never conscious of the objective degree of all these changes unless we use... a peephole that makes us see a speck of color but masks off its relationships. Those who have used this magic instrument report the most striking discoveries. A white handkerchief in the shade may be objectively darker than a lump of coal in the sunshine. We rarely confuse the one with the other because on the whole the lump of coal will be the blackest patch in our field of vision, the handkerchief the whitest, and it is the relative brightness that matters and that we are aware of. The coding process begins while en route between the retina and our conscious mind.the art historian E. H. Gombrich, 1959
Email || Home Page || Table of Contents || Endnotes for this chapter || continue reading Next Chapter
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.