by Morton Hunt
Revolution No. 2
What, another revolution so soon?
Well, not on the heels of the cognitive revolution, but not far behind it. This one, though long gathering force, would not burst forth until the 1980s, but we must look ahead to its emergence because much of what we will see happening in cognitive psychology will be affected by it. It was the cognitive neuroscience revolution.
That’s a relatively new name for an old school of thought about the mind, the biological approach to mental processes that sought to explain them in terms of neuronal processes and events. We saw a notable example of it in Hubel and Wiesel’s discoveries of retinal cells that respond only to specific shapes or directions of motion. That was recent, but the neuroscientific approach has antecedents going back at least to Descartes. Although he believed in the immateriality of mind, he conjectured, as we saw, that reflexes were caused by the flow of “animal spirits” through the nervous system, much as the movements of automata in the royal gardens were caused by the flow of water in pipes, and that memory was the result of the widening of the particular “pores of the brain” through which animal spirits had passed during learning.12 Similarly, a century ago the young Freud confidently asserted that all psychological processes could be understood as “quantitatively determined states” of the neurons, though he soon admitted with chagrin that the time was not ripe for such understanding.
The same hope, though, had continued to inspire many researchers. And during the past sixty years, and especially the past twenty-five, extraordinary advances in cognitive neuroscience have led some enthusiasts to assert that it will soon replace the psychological approach to the mind and that concepts such as needs, emotions, and thoughts will be replaced by physiological data. When such data are available, Paul Churchland, a philosopher of neuroscience, asserted in 1984,
we will set about reconceiving our internal states and activities, within a truly adequate framework at last. Our explanations of one another’s behavior will appeal to such things as our neuropharmacological states, the neural activity in specialized anatomical areas, and whatever other states are deemed relevant by the new theory.13
Most research in behavioral neuroscience in the decades immediately preceding its 1980s breakout as cognitive neuroscience was focused not on thought processes but on the physical events taking place (as thought occurs) in the “wetware”—the hundred billion or more neurons that make up the human brain. Cognitive neuroscientists—some were neurobiologists who had studied psychology, others were psychologists who had studied neurobiology—were interested in such phenomena as the flow of sodium and other ions into and out of the axon (main stem) of a neuron as electrical impulses pass along it; the molecular structure of the neurotransmitters (the chemicals produced in the synapses, the junctions at which the impulses are passed to other neurons); the bursts of neurotransmitter molecules that leap across the microscopic synaptic gaps from neuron to neuron carrying messages of excitation or inhibition; and the neuronal routes and networks activated by different kinds of stimuli and mental activities.
Behavioral neuroscientists (as they were then known), often white-coated, spent much of their time in operating rooms and laboratories, where, among other things, they surgically destroyed specific portions of animals’ brains to learn what aspects of behavior those parts control; they interviewed and tested people who had suffered brain damage; they measured and recorded the spikes of activity of single neurons and the overall patterns of brain excitation (“brain waves”) during various mental activities; they administered drugs that increase or decrease the production of particular neurotransmitters to determine what functions these perform; and they did chemical analyses of the brain tissue of laboratory animals and human cadavers to see what neurotransmitters were in either short supply or excess in individuals whose behavior is abnormal in some respect.
A good deal of their work, as we have already seen, involved testing patients with cerebral damage (most often strokes), pinpointing the affected brain area and identifying it as the cause of the patient’s diminished or lost perceptual and mental abilities. But much other neuroscientific research, though arguably valuable, had its comical overtones. One investigator implanted sixteen microelectrodes into the muscles of a male grasshopper in order to record the electrical impulses of its neurons during courtship. Others inserted microelectrodes into the left front leg of a cockroach and the foot of a snail to measure the neural impulses that produce movement toward some goal; the investigators regarded this as research on “motivated behavior.”14
Of all cognitive processes, especially in more advanced species, memory is the most basic, and for decades cognitive neuroscientists sought to identify how and where memory exists at the cellular level. A few examples of the ways in which they did so:
—As long ago as 1949, Donald Hebb, a Canadian psychologist, hypothesized that memories are stored by the modification of the synapses connecting neurons (an idea not unlike Descartes’s). The repeated activation of a synapse in a learning experience, he said, somehow strengthens the synapse and links the two neurons into a circuit or “memory trace.”15 Hebb’s hypothesis was more or less confirmed in 1973 when a British neurophysiologist, Timothy Bliss, and a colleague, Terje Lømo, measured the voltage in one neural pathway in the brain of a rabbit, then sent repeated bursts of electricity down the path, and afterward found that the pathway carried a higher voltage than before. The synapses had been strengthened by the electrical impulses. The implication was that that is what happens in learning.16 (Later research, as we will see, has added many details and complexities to the explanation.) —Also in the early 1970s, an American psychologist, William Greenough, raised rats in two environments, one containing toys, mazes, and other stimulating devices, the other without any. The rats in the stimulating environment developed heavier areas of cerebral cortex; the neurons in those areas had grown more dendrites and thus more synapses than those of rats in the dull environment. Later, by means of electron microscopy, Greenough and a colleague actually counted 20 to 25 percent more synapses in the affected cortical areas of the enriched-environment rats than those of the deprived ones. Learning had generated the extra connections; memory traces must somehow be recorded in them.17
—In the late 1980s Daniel L. Alkon and his colleagues at the
National Institute of Neurological and Communicative Disorders and Stroke trained a sea snail, Hermissenda crassicornis, to respond to light in a way it does not normally do. Hermissenda instinctively swims toward light; also, when the water is turbulent it instinctively clenches its foot muscle in order to cling to a surface. Alkon combined these reactions. By flashing a light and simultaneously whirling the chamber in which he housed the snail, he conditioned it—taught it—to clench its foot muscle whenever it saw a flash of light. He then found that in certain of the snail’s photoreceptor neurons, molecules of PKC, a calcium-sensitive enzyme, had moved from the interior of the neuron toward its membrane, where they reduced potassium-ion flow—a partial explanation of memory in molecular terms.18
—Over several decades, James L. McGaugh and other researchers did a number of studies in which they injected epinephrine (a hormone produced by the adrenal gland) and other catecholamine neurotransmitters into rats after training them to run a maze. Epinephrine, in particular, causes the rats to remember longer what they learned than rats not so dosed. The explanation, deduced from other studies, seems to be that a byproduct of the epinephrine combats opioids, a group of neurotransmitters that serve useful purposes but plug up receptors on the receiving side of synapses. The result: more receptors remain open, the synapses function more efficiently, and memory is strengthened.19
These and many other research studies made cognitive neuroscientists feel sure that they were on the right track to explaining the many mysteries of psychology. Their approach promised to end, once and for all, the ancient debate about body and mind by explaining all mental processes in terms of material substances
and events. All high-level mental processes such as memory, language, and reasoning were only ions and molecules flowing hither and thither in the labyrinthine and infinitesimal plumbing of the brain.
But the great majority of cognitive psychologists, proud of their new dominance and excited by the amazing capacity of computers to mimic—and perhaps explain—human reasoning, were dismissive of cognitive neuroscience. In the 1950s, after Newell and Simon’s dramatic presentation of Logic Theorist, whatever connection had existed between cognitive psychology and neuroscience fell apart; Simon, in fact, authoritatively declared that to “understand cognition, one needn’t pay much if any attention to the underlying biology.”20
For the next twenty-five or so years, most cognitive psychologists agreed with him, insisting that neural events do not provide an adequate or useful explanation of cognitive phenomena. Few were dualists in the sense of believing in immaterial mind, but they asserted that psychological processes, though constructed of neural events, were properties of the organization or metastructure of those components, not of the components themselves, even as shelter is not a property of bricks, beams, and shingles but of a house built of them.
Nobel Laureate Roger Sperry, though himself a brain scientist, offered another analogy: a higher-order mental process is like a wheel rolling downhill—the rolling is determined by the “overall system properties” of the wheel, not by the atoms and molecules of which it is made.
The developmentalist Jerome Kagan used a different analogy: the elegant laws of planetary motion illustrate phenomena that are not expressible in terms of the atoms of which the planets are made.
Another analogy, this from the cognitive scientist Earl Hunt: “We can tell from physical measures that the left temporal region of the brain is active when we read, but we cannot discriminate the activity induced by reading Shakespeare from that induced by reading Agatha Christie.”
And a word from the cognitive psychologist George Mandler: “The mind has functions that are different from those of the central nervous system, just as societies function in ways that cannot be reduced to the function of individual minds.”21
Most cognitive psychologists thus believed that a word retrieved from memory could not be equated with the firing of millions of neurons and the resultant millions or billions of synaptic transmissions, but was the product of the pattern or structure of those firings and transmissions. The neurobiological study of memory, valuable as it was, did not tell us how we learn anything, recognize things we have earlier experienced, or retrieve items from memory as needed—the words we use in speech, to give one example. Such phenomena, they insisted, were governed not by the laws of cognitive neuroscience but by those of cognitive psychology.
Martha Farah, a distinguished neuroscientist and director of the Center for Cognitive Neuroscience at the University of Pennsylvania, recalls that in 1980, when she was a graduate student of psychology at Harvard, “I asked to take a course in neuroanatomy—and got lectured. I was ‘supposed to be studying how the mind worked, and looking at how the brain works was simply not relevant.’ That was the received wisdom in those days. The ’70s and ’80s were the last hurrah of brain-free psychology.”22
What ended the reign of brain-free psychology? Many things, including:
—the growing mass of data on neuronal transmission, on the functions of brain substructures, and on the molecular and other factors that strengthen synaptic connections in learning;
—the shortcomings of the computer model of cognition (it was becoming apparent that although computers could simulate some aspects of cognition, the mind processes information in vastly more complex ways than the linear step-by-step fashion of computer programs);
—the weakening resistance of some leading cognitive psychologists to valuable neuroscientific findings about brain processes; and
—the growing sense among neuroscientists by the late 1970s that they were doing far more than exploring brain biology and that their domain should be called “cognitive neuroscience.”
But as has been the case in various other sciences, it was a new tool (actually a set of tools) that transformed the domain of neuroscience and produced a second revolution in the cognitive sciences. The tools were an array of brain scan devices—machines that could produce various kinds of images of the working brain and, most importantly, of the physical changes or events taking place in it when mental processes were in progress.
Prior to the 1980s, physiologists had been able to use EEG (electroencephalography) to show the form of brain waves; this was useful in studying the differences in brain wave activity during various wakeful and sleep states and the distortions of the waves during epileptic seizures. The method, however, was poor at localizing the brain activity of specific cognitive processes because it reflected overall electrical activity, not that of specific regions or structures of the brain.23
But in the 1980s several dramatic advances were made. One was the development of PET (positive emission tomography) scanning after many years of experiments in measuring blood flow in the brain. In a PET scan the subject lies supine on a narrow table which rolls into a large tubular machine. A nearby cyclotron generates a weakly radioactive isotope with a half-life of only two minutes which is then injected into the patient. The scanner, sensitive to the isotope, records blood flow in a “slice” (narrow cross-section) of the brain, the isotope showing where the brain is active. From a number of slices, a computer assembles three-dimensional images of the brain. The PET scan can be used clinically to study physical damage to, and abnormalities in, the brain, but cognitive psychologists and neuroscientists soon began using PET scans to see what areas of the brain had increased blood flow during—and thus were involved in—various kinds of mental activity.24
In 1983 another important tool was introduced—CT (computed tomography) scanning, also known as CAT scanning (computerized axial tomography). It proved to be a valuable medical tool for assessing many kinds of physiological problems, but also for studying brain structure and identifying brain lesions. In the CT scan, the subject is, as in a PET scan, supine, and eased into the scanner, which has an X-ray source and a set of radiation detectors. The scanner sends radiation through the target part of the subject from various angles. The density of biological materials varies; accordingly, the data gathered by the detectors reveal the hidden structure, and are assembled by a computer program to yield X-ray pictures of the entire scanned target. CT scans were and are used primarily for clinical medical analyses, but the method also had some value for cognitive research on brain structure, although the results are not very distinct and lack fine resolution.25
By far the most important and latest new tool is the MRI (magnetic resonance imaging) scan. Again the subject is supine inside a scanning machine, which is about the size of a small SUV, and which, while making a horrendous racket, generates a powerful magnetic field that permeates the subject’s head. The magnetism, unlike the radiation of the CT scan, is harmless—and capable of revealing brain structure and activity far better than the CT scan.
It can do so because the protons of hydrogen, a major component of the water and fat in the brain, behave like tiny magnets and line up under the influence of the magnetic field (normally, their orientation, unaffected by earth’s weak magnetism, is randomly distributed). Then, radio waves, passed through the subject’s head, change the protons’ orientation, but the instant the radio waves stop, the protons bounce back to the orientation created by the magnetic field and, in so doing, emit energy signals. These, picked up by detectors, yield scans much clearer, and with far finer resolution (a spatial resolution of one millimeter and a temporal resolution of about one second) than any other scanning method.
Best of all, from the cognitive researcher’s viewpoint, if the subject performs some prescribed mental task while being scanned, the resulting fMRI (functional MRI) scan gives an intimate look at exactly which brain areas and substructures are active, and how active, during that k
ind of mental activity. Accordingly, the fMRI quickly became the workhorse of cognitive neuroscience. A dozen years ago, a mere handful of studies based on fMRI scans appeared in a year’s worth of research literature; today, the annual output is several thousand.26
What has all this done to psychology, the science of the human mind? That depends on who is assessing the situation.
Most psychologists, focused on mental processes rather than wetware, continue to use research methods that were available before the advent of scanning, but many of them also rely on the help of scanning. They no longer see cognitive psychology and cognitive neuroscience as distinct and unrelated fields. As Robert J. Sternberg, a notable cognitive psychologist, says, “Biology and behavior work together. They are not in any way mutually exclusive.”27 Some use stronger terms to appraise the impact of cognitive neuroscience: Psychologists Stephen Kosslyn and Robin Rosenberg write, “It is fair to say that neuroimaging techniques have transformed psychology, allowing researchers to answer questions that were hopelessly out of reach before the mid-1980s.”28
Does that suggest that cognitive neuroscience will become the psychology of the future? Not according to cognitivist Michael Posner, who has worked in both camps and whose work has been admired by researchers in both: “An impressive aspect of the anatomical methods such as PET and fMRI is how much they have supported the view that cognitive measures can be used to suggest separate neural structures,” and he stresses the importance of the contributions of both fields to understanding brain function.29
But some cognitive neuroscientists think it possible, even likely, that their field will come to dominate mental science. Martha Farah, when asked if cognitive neuroscience would eventually become the overarching theory of psychology, said, “Yes, because it’s a broader and more heterogeneous approach to studying the mind which encompasses cognitive psychology. It’s a molecular-cellular-systems explanation of how the brain acts during all the classical processes of cognitive psychology—how we learn, think, behave, why we differ from each other, the sources of personality. All these things are in principle explainable by various levels of brain activity at various levels of description.”30