AIR FORCE RESEARCH ON LIVING PROTOTYPES

Harvey E. Savely

Air Force Office of Scientific Research

As a biologist in the Air Force, it is indeed gratifying to see so many physical scientists motivated to join with biologists in discussing common interests in the life sciences. Such a gathering as this could hardly have been possible, even 10 years ago particularly in a military setting. The rapid progress in all the sciences, including biology, has given us this prospect of finding technological implications in the study of living things.

Biology has been a descriptive science for most of its history. And even today the great diversity among living things is still being identified and described. But we are witnessing today a rapid "coming of age" of biology as an analytical science. This is due in part to developments in the physical and chemical sciences which are now being applied to the study of the fundamental patterns in living systems. The Air Force, along with other military services, has recently shown an increasing interest in biology as a source of principles applicable to engineering. The reason clearly is that our technology is faced with problems of increasing complexity. In living things problems of organized complexity have been solved with a success that invites our wonder and admiration. It is natural, therefore, that we look to these successful inventions in nature for clues, as well as inspirations, for new classes of man-made machines with greatly increased capabilities. I look for this interest on the part of the military to increase, as biology takes on this new significance to military research planning.

It would be impossible in a short time to treat adequately the great diversity of animals, not to mention plants, which may show behaviors that could serve as models for engineering application. All of the wonderful and unusual adaptations of animals and plants are a part of our folk lore and common knowledge. To recite many of them would, I am afraid, turn this into a nature lovers' lecture. I must ask you to fill in from your imagination the great variety and richness of species, each of which stands at the end of a long chain of change and innovation, stretching back for two billion years.

I will talk briefly about three interrelated aspects of the nervous system of animals which have been occupying much of our attention in the Air Force. These are:

(1) The sensory receptors of animals

The integrative action of their nervous systems (3) The storage and retrieval of information.

The sensory receptors are the transducers by which animals stay attuned to significant events in the world around them, as well as in the machinery of their own bodies. We commonly think of them as the basis of our five senses.

But the receptors are much more varied than we might infer from our human sensations. In our own body, for example, sensitive mechanoreceptors monitor pressure, volume, position, displacement, tension, acceleration, and vibration, usually without our being aware of it. In addition, sensitive chemical receptors within our body monitor the important aspects of body regulator mechanisms.

We have recently come to appreciate the great range of physical events in the environment to which animals are sensitive and new types of senses are still being discovered and described.

For example, only a few years ago Dr. Bulloch and his associates at UCLA described the sensitive infrared sensing organ in the rattlesnakes. This organ, which is located in the pit between the nostril and the eye, is so sensitive that it responds to a change in temperature at its surface of 0.001° C. The frequency of nerve impulses flowing in the nerves leading from this organ changes so greatly, with small changes of temperature, that if you were to compute the change in the terms conventionally used to describe the dependency of chemical reactions on temperature you would come out with a Q10 of 1030. To quote Bulloch, "this figure offers considerable room for speculation about high amplification with preservation of reasonable stability."

Another form of energy, which most animals cannot sense except as a shock, is electrical energy. Yet there are at least three families of fish existing in the tropical areas around the world that can detect small changes in the electrical field of their surroundings. Some, like the electric eel, produce high voltages--up to several hundred volts--which are use for warning, defense, or attack on prey. But even more interesting are the several groups which emit pulses of low voltage in the order of 1 volt-in either bursts or continuously throughout their lives. The frequency and pulse form is characteristic of each species, and ranges from 50 to 1,600 cps. By listening to an audio signal produced by these frequencies, a biologist with a good sense of pitch can identify the species, as well as guide himself to them in a stream. Little is known of what these animals do with this highly developed sensory system. In one species that has been studied by Prof. Lissman of Cambridge University it was found that the fish was sensitive to a change in the field in the water of 0.003 microvolts/mm. For the benefit of electronic engineers who may be shocked by this figure, I have checked my reference carefully. I am talking about 3 x 10-9 volts per millimeter. This corresponds to a current through the fish of 2 x 10-5 microamperes per cm2. This fish could discriminate between two covered glass rods in its tank differing in diameter by only 2 mm. It is likely therefore that many electric fish have a highly developed form of object location for identification and orientation.

One of our senses we often take for granted is vision. The ability to make pattern recognition and detect motion has arisen independently in three of the great groups of animals. The vertebrates, to which we belong; the arthropods, which includes the insects, spiders, lobsters, and other animals with exterior skeleton; and in the mollusks, which include the squids, octopus, clams, and mussels. We are encouraging work in all these forms in the belief

that analysis of these separate developments in nature could have important implications for pattern recognition and the science of automata.

Studies on insect vision have already shown the kind of unexpected payoff that can come from an analytical approach to what might seem at first to be a trivial problem. Dr. Hassenstein and Dr. Reichardt at the MaxPlanck-Institut in Tubingen, Germany, have spent several years studying the response of a beetle to moving light patterns. This team consists of a zoologist, a physicist, an electrical engineer, and a mathematician; and the skills of each of these disciplines were required to formulate and carry out the series of experiments that explained the beetles' behavior. When the results were expressed in the language of control systems' theory it appeared that the beetle could derive velocity information from a moving randomly shaded background. The special mathematical formulation for the required autocorrelation had to be derived before the investigators could be convinced of their theory. The pay-off is that these workers have initiated the design of a ground-speed indicator for airplanes which works like the beetle eye and is based directly on the function of just two of the hundreds of facets that make up the compound eye of this insect. Other insects have more highly developed eyes and appear to have pattern recognition and color vision as well.

We have been interested in the work of Professor J. Z. Young and his associates in London and Naples, on vision in the octopus. This animal's eye looks superficially like our familiar vertebrate eye, but you must remember that it evolved in the mollusk quite independently from the course followed by the vertebrates. It is not surprising, therefore, that the octopus solves its pattern recognition problem in a way quite different from vertebrates. Just how is still not well understood any more than is the problem of vertebrate vision.

Perhaps many of you read the article in the IRE Journal by Lettvin, Maturana, McCullock, and Pitts from Massachusetts Institute of Technology entitled, "What the Frog's Eye Tells the Frog's Brain." This marks a significant advance in knowledge of vision in a lower vertebrate, and is a type of study that can lead to principles of engineering significance. Jerry Lettvin is now in Naples with Professor Young trying to apply the same techniques to vision in the octopus.

The chemical sensing organs are another class of receptors that reach a high degree of development in many classes of animals. It is highly developed in man and some of the other vertebrates, and perhaps is more sensitive in some of the insects. Detection of a variety of odors and of chemical substances is possible in amounts so dilute that it is estimated that one molecule may be sufficient to fire off a receptor. In some cases the receptor is sensitive to specific chemicals and even to separate chemical isomers. This ability makes it possible for a male moth to "home in" on the odor from a female from great distances.

On the problems of hearing, the group here at Wright Air Development Division under Dr. von Gierke is making important contributions to understanding the bioacoustics of this class of mechanoreceptors. The principles at work here could contribute just as importantly to problems of pattern recognition as studies on the eye. As in the case of other human senses we can point to some of our animal relatives that have far exceeded man in the

specialization in hearing. You are probably all aware of the echo location ability in bats, porpoise, and other animals that have a kind of sonar. The acute hearing of the owl enables it to be guided accurately in the dark to high pitched noises of mice. Perhaps you have not heard of the moth that has developed a highly effective hearing organ which can detect the ultrasonic cries of a bat, and which enables it to take evasive action just before the bat makes a meal on it. This hearing organ is composed of just two cells, yet it covers the frequency range of the bat. And incidentally, by far the best microphone for picking up bat cries can be made by attaching electrodes to the nerve leading from this moth's ear.

The story on sensory receptors is a long one, and I can do no more than mention a few well known examples. New types of sensory modalities are being reported continually, such as polarized light detectors, wind speed detectors, gyroscopic deflection sensors, hydrostatic pressure sensing in a gas-free system, and combination mechanical and chemical receptors. The list could be extended. Anatomists have described many sense organs for which no function has yet been assigned.

There are undoubtedly undiscovered sensory transducers. Birds perform great feats of orientation and navigation by quite unknown sensory systems. There is growing evidence that some of the night migrators may be using some aspect of night sky, perhaps stars for navigation. These, like some other classes of problems in biology, are still in the phenomenological and descriptive stage, which must precede the identification of sensory mechanisms and analytical studies of their function. This whole area of biological transducers is one that deserves considerably more attention from both biological and physical scientists.

Let us turn now to another aspect of nervous systems--what we may call the integrative action by which the inflow of sensory information is filtered and mixed with other inputs to arrive at the pattern of behavior we see. After the transduction of the physical energy into a nerve impulse having a characteristic digital code, the central nervous system of the animal must carry out digital to analogue conversions of the signal to analyze and mix with other disturbances in the nervous system. Here we come up against organized complexity carried to unimaginable extremes. The evolutionary culmination of this process is, of course, the human brain. In the words of Sherrington, the great physiologist of the nervous system, man has become "nature conscious of itself."

The problem of human complexity may seem overwhelming, but fortunately we do not have to solve this problem first. We have before us in nature millions of examples of experiments in information handling, and in all degrees of complexity. It is no great stretch of the imagination to suppose that simple brains or small functional aggregates of nerve cells can come under physical analysis. And we have every right to expect in this aspect of science, as in others, that we will be aided in gaining understanding of the more complex systems by first understanding their simpler elements.

The Air Force has a considerable interest in the extensive research on brain mechanisms of higher vertebrates which is being performed around the world. This work involves chemical, electrophysiological, and behavioral approaches to the study of nervous activity. Much of this work has been

motivated by the need to know more about the brain, as a basis for understanding medical, behavioral, and performance problems, and this is an area where empirical knowledge can be useful. But to the engineer some of these studies may seem diffuse, and to offer few opportunities to synthesize useful laws of action from the facts. It is true that we lack fundamental knowledge of the laws that govern the connections and functions in this complex vertebrate brain, or, indeed, even in simpler brains. But the kinds of behavior we want most to imitate are performed by these complex brains. The task is a difficult one, but will certainly yield rich dividends for technology when a more analytical approach can provide the data for useful syntheses. To understand this aspect of the nervous system new concepts in physics and mathematics may well be needed. It is here, too, that the application of computer simulation techniques may play an important role. For we still do not have the methods for tackling problems of organized complexity which is characteristic of living systems. Whatever new concepts are forthcoming, they will almost certainly require new uses of computer techniques to handle all the variables that are involved.

There is some evidence that a simulation approach can be a real aid in gaining new knowledge about the nervous system and not just confirmatory knowledge as has often been the case in the past. But at the moment we have far too few biologists who can use the language of the communication sciences, and too few physical scientists who appreciate the extent, as well as the limitation, in present biological knowledge. It would seem to be an appropriate time for a science of automata to develop common ground with biology, to the mutual benefit of both.

A third aspect of biological systems that should ultimately have profound implications for engineering is the ability of living things to store and retrieve information. This is a characteristic phenomenon of all living things, at the cellular level, where metabolic activity is influenced by preceding events that may have occurred hours or days or even generations before. In the nucleus of the cell is stored the pattern on which an additional copy of the organism can be built. We know nothing about the code by which this pattern is stored or expressed. We have only to reflect for a moment on the mass of detailed information which must be needed to guide the development of an animal, such as ourselves, to be awed by this problem. But a beginning is being made on the problem by the study of the structure and characteristics of the large molecules that carry out the transmission of genetic information. The field has made great strides in the last 10 years and promises to have far reaching importance for all the sciences and the technologies that rely on them.

Storage and retrieval of information at the level of the central nervous system underlies both the short and long span memory in animals, and the conscious behavior in man. Unfortunately, we can say very little biologically about this problem. Its solution may well be linked to understanding the fundamental aspects of the molecular systems that are at work in all cells. It stands as another aspect of the great challenge of the biological sciences. Its implications are so great that it must be a part of any military program of basic research.