It is always entertaining to look into the crystal ball and predict the future—entertaining but very speculative. The study of the brain in all its aspects, from genes to neurons to consciousness, is expanding at an almost exponential rate. Much of this new knowledge will impact our understanding of memory and the ways that memory processes might be altered or even enhanced in the future.

Some people are disturbed by the notion that we can alter genes. There is even opposition in some parts of the world to the use of food products genetically engineered to be more resistant to disease or to have a longer shelf life. Actually, people have been doing genetic engineering for thousands of years: selective breeding. The huge variety of dogs, from the Chihuahua to the Great Dane, have all been selectively bred from ancestral wolves. The only

difference between this kind of genetic manipulation and genetic engineering is that we can now alter the genes directly.

Memory abilities, like other human abilities and characteristics, have a significant genetic basis, called heritability. But genetic influences are by no means complete. As with other human attributes, memory abilities are influenced by genes and by development and experience, that is, by the environment.

In a classic study done many years ago at the University of California at Berkeley, Robert Tyron selectively bred rats to be maze-bright or maze-dull. After many generations the two groups diverged completely in their ability to learn mazes: The worst-performing maze-bright rat did better than the best-performing maze-dull rat. Maze learning in rats does indeed have a genetic basis.

In another experiment, discussed in Chapter 9, many mutated flies were much impaired in olfactory learning, but a genetically engineered fly with the enhanced biochemical CREB function learned much faster than normal flies. Similar results hold for mice. Impairing the function of the NMDA receptor molecule on neurons in the hippocampus impairs maze-learning ability. Joe Tsien, of Princeton University, engineered mice with enhanced NMDA function in hippocampal neurons, and these mice were super maze learners.

But what about people? Thanks to our knowledge of genetics and metabolism, some serious forms of mental retardation can now be prevented. A marvelous success story concerns a condition called phenylketonuria. The labeling on diet drinks warns that a sweetener called aspertame is used. This is actually the amino acid phenylalanine, which is present in certain foods. Some infants are born with a genetic defect in their ability to normally metabolize phenylalanine. Instead, they convert it to a toxic substance that can kill nerve cells. If untreated, these infants will develop brain damage and mental retardation.

As it happens, this disorder can be diagnosed in newborn babies by a simple urine test; many newborns are now routinely tested. If the disorder is present, the treatment is simply to

avoid all foods containing phenylalanine, and this prevents the disorder.

With modern genetic engineering techniques it may be possible to alter this genetic defect in phenylketonuria directly and thus not only prevent the disorder but also cure it. A number of attempts have been made to treat human genetic disorders directly using genetic engineering methods—methods developed in animal studies—but the results have been mixed.

There are reports that a single gene may be involved in determining high intelligence. Should we genetically engineer all infants to be smart? This raises serious ethical and moral questions. Aldous Huxley’s novel Brave New World is a frightening prediction of what might happen when genetic engineering is applied to people in a totalitarian society. Even the notion of selective breeding of humans, as the Nazis attempted in World War II, is abhorrent to most of us.

Individual people, of course, engage in a form of selective breeding when they marry. It sometimes happens that older wealthy or famous women marry much younger, attractive men, and vice versa. The beautiful dancer, Isadora Duncan, once propositioned the famous writer George Bernard Shaw to have a child with her. “Think of the incredible outcome,” she said, “a child with my body and your brain.” “But madam,” he replied, “what if it had my body and your brain?”

These are questions that extend beyond science. Part of the problem is that far too little is known about the genetic basis of complex human characteristics. Perhaps someday it will be possible to genetically alter humans so that they will all be healthy, intelligent super-learners. But should we? The ethical questions remain.

The activity of neurons in the brain generates electrical signals that can easily be recorded from the surface of the scalp, using the electroencephalogram (EEG). Considerable information is con-

tained within the EEG record, which averages the activity of many neurons. We can tell if someone is aroused or resting, if they are directing their attention, or whether they are awake or in either of the two sleep states. Patrick Suppes of Stanford University and Zhong-lin Lu of the University of Southern California and their colleagues appear to have succeeded in decoding “thoughts” from EEG records. They presented subjects with 48 different sentences about European geography, and in each case recorded brain EEG activity for many different electrodes on the surface of the scalp. Using complex mathematical analysis of the EEG records, they were able to correctly recognize 90 percent of the brain waves generated by the 48 different sentences! Although this is just a beginning, it is conceivable that it will someday be possible to decode thoughts from records of brain activity.

If this could be done, a person could communicate directly—electrically—with a computer. In fact, monkeys have been made to do just that. Electrodes are implanted directly into areas of their cerebral cortex, which gives much better resolution of the electrical activity than records from the scalp. Miguel Nicolelis of Duke University has recorded the activity of many individual neurons from electrodes in a region of the motor cortex when a monkey subject is making an arm movement. He then uses the recording via a computer to generate the same movement of an artificial arm. In this way he is able to have the monkey control the movement of the artificial arm just by thinking about the movement.

Richard Andersen and his colleagues at the California Institute of Technology have gone even further. There is a posterior region of the cerebral cortex in monkeys and humans in front of the visual cortex that is critical for sensory-motor integration. It functions as the place in the cortex where intentions to act are formed—that is, high-level cognitive plans for movements, including eye movements, reaching movements, and grasping movements. In monkeys, Andersen recorded the activity (action potentials) from neurons in this region that responded prior to the animal making arm and eye movements and was able to computer decode the activity patterns of the neurons in terms of what

movements the animal intended to make. He then used such recordings to correctly generate the movements the animal was planning to make before it made them (see Figure 10-1).

These studies by Nicolelis and Andersen raise the very real possibility of helping people who have lost limbs or been paralyzed. Paralysis is typically due to damage to the spinal cord or to the motor neurons that control muscles. The brain systems that control movements and intentions to move are still intact and functional. It may someday be possible to record the neuronal activity that codes skilled learned movements, perhaps in the cerebellum, for such patients so that they can yet again “play” the piano.

FIGURE 10-1 Neuronal activity in a region of the monkey brain code the animal’s intention to make movements before it makes them. By recording from the neurons, that is by reading the thought, the movement can be made before the monkey makes it.

As more and more is learned about the detailed circuitries of the human brain and how they generate their extraordinary achievements, from consciousness to science, music, art, and literature, it will be increasingly possible to re-create these circuits in computers and even in hard-wired transistor circuits. The brain, after all, is like a computer, granted an extremely complex one with both hardware (neurons) and wetware (chemicals) and with feelings like pain and joy that computers have yet to achieve. Present-day computers cannot yet approach the complexity of the brain (each of the millions of Purkinje neurons in the cerebellum receives synaptic contacts from 200,000 or so different granule neurons, for example). But perhaps it is only a matter of time until sufficiently complex computers will become available.

Along these lines, Theodore Berger and his associates at the University of Southern California have analyzed the information-processing capabilities of the mammalian hippocampus. They simplified the problem by using an engineering approach, treating the hippocampus as a “black box.” They sent a wide range of electrical signals into a hippocampal slice (from a rat) and recorded the output. Knowing the basic neural circuitry of the hippocampus, they were able to construct a computational model of the circuit in a computer, at least in terms of information processing.

Berger and his colleagues then built a physical model of the hippocampal computer circuit using electronic chips (see Figures 10-2 and 10-3). This device, like the computer circuit it simulated, processed information like the hippocampus. Amazingly, this electronic hippocampus turned out to be a language recognition system. In fact, it outperformed all commercially available English-language recognition devices. This sort of unexpected outcome is a particularly clear example of how basic research on the brain can lead to extremely useful applications in society.

The long-term goal of this project is to develop electronic chips that simulate brain circuits and that can actually serve as replacements for damaged circuits in the human brain, particularly circuits like the hippocampus that are critical for memory storage. The project involves marrying silicon chips to neurons so

they can communicate with one another, which has already been accomplished on a small scale. Roberta Brinton, at the University of Southern California, has succeeded in growing neurons in a cell culture on electronic chips that can intercommunicate.

The idea that nth-generation computer chips can serve as replacements for damaged brain regions sounds very much like science fiction—the creation of “cyborgs.” But if this can be achieved, the human brain can be “plugged in” to nth-generation computers, vastly expanding the memory and information-processing capabilities of the human brain. This may well be in our future.