sign in sign up
<<
Home , Brenda Milner, Sigmund Freud

The and Donald O. Hebb By rooting behavior in ideas, and ideas in the , Hebb laid the groundwork for modern . His theory prefigured computer models of neural networks

by Peter M. Milner

onald O. Hebb, one of the most they called physiologizing). But many ing care to prevent schoolwork from be- inßuential of his psychologists had grown weary of the ing used as a punishment, instead send- time, began his life in- artiÞcial theories these strictures had ing miscreants out of class to in the tending to be a novelist. Deciding that engendered, and they were captivated school yard. Hebb became absorbed in his calling required an understanding by HebbÕs project and his engaging lit- his educational experiments and seri- of , he embarked on a course erary style. The book became a classic, ously considered remaining in the pro- that led into two decades of re- and Hebb became a household word (at fession. Two developments dissuaded search. His studies culminated in 1949 least in psychologistsÕ households). him. He came down with a tubercular with the publication of The Organiza- Hebb never claimed that his 1949 the- hip that conÞned him to bed for a year tion of Behavior, a keystone of modern ory was Þrmly grounded in physiology. and left him with a slight limp. Then his neuroscience. His model gave workers something to bride of 18 months was killed in an au- The monograph broke new ground by look for, and later, as of the tomobile accident. He therefore decid- positing neural structures, called cell as- brain grew, it became possible to frame ed to leave Montreal. semblies, which were formed through his ideas in more realistic neural terms. While conÞned to bed, Hebb wrote the action of what is now called the None of this subsequent research has a masterÕs thesis that involved him in Hebb synapse. The cell-assembly theory invalidated HebbÕs basic hypothesis. In- the nature-nurture controversy. The the- guided HebbÕs landmark experiments deed, its inßuence appears in many ar- sis attempted to explain spinal reßexes on the inßuence of early environment eas of current research. as the result of Pavlovian conditioning on adult . It foreshadowed in the fetus. He subsequently buried all neural network theory, an active line of ebb was born in Chester, a references to this essay both because research in artiÞcial intelligence. small Þshing and boat-building he changed his mind about its content HebbÕs book came at the right time Htown in Nova Scotia. His parents and because he came to oppose psy- because it ßew in the face of behavior- were physicians, and his two brothers chological research that lacked an ex- ism just as that school was losing its and his sister followed in their parentsÕ perimental foundation. dominance. The behaviorists denounced footsteps. But Donald demonstrated One of his examiners was Boris P. explanations of behavior by his independence early by studying En- Babkin, a physiologist who had worked of ideas (which they called mentalism) glish in preparation for a career as a with Pavlov in St. Petersburg. He recom- and by the action of (which , graduating in 1925 from Dal- mended that Hebb get some experience housie University in Halifax. To earn in the laboratory and arranged for him his living while gestating his Þrst nov- el, he taught school in his hometown. PETER M. MILNER is professor emeri- A year later he set out to see life, going DONALD O. HEBB made his name as tus of psychology at McGill University. He a theoretician but was equally distin- received a B.Sc. in engineering from the west to work an eight-horse team on University of Leeds in the U.K. in 1942 prairie farms. Then, failing to get a job guished as a teacher. Here he appears and was immediately recruited for radar as a deckhand on a freighter to China, in a seminar held in the late 1960s. research by the novelist C. P. Snow. ÒMy he returned east and got a job as a la- Þrst serious encounter with psychology borer in . followed soon afterward,Ó he says, Òwhile In 1927 an aspiring novelist not only working on the interface between radar had to know life but also the works of displays and the human visuomotor sys- Sigmund . This was HebbÕs intro- tem.Ó He later went to Canada to work on nuclear . In 1948 he left physics duction to psychology. He was suÛ- to take a Ph.D. in physiological psychol- ciently intrigued to apply to the psy- ogy at McGill under the supervision of chology department of McGill Univer- Donald O. Hebb. Milner is co-discoverer, sity, where he was accepted in 1928 with the late James Olds, of the reward as a part-time graduate student. Again mechanism in the brain and has written he supported himself by teaching and, extensively on the and its impli- again, what started out as a temporary cations for theories of reinforcement and . He has also written a textbook interest verged on becoming a career. on as well as After one year he was made principal papers on , visual recognition, of an elementary school in a working- and the mind-body problem. class district of Montreal. He was deter- mined to make learning enjoyable, tak-

124 SCIENTIFIC AMERICAN January 1993 Copyright 1992 Scientific American, Inc. to work with another Russian emigrŽ, not be stored in a single region of the mally. This Þnding indicated that the Leonid Andreyev. Hebb conditioned brain but must be spread throughout. organization of the visual system was dogs and became less impressed with In 1934, when Hebb went to Chicago, innate and independent of environmen- Pavlovian techniques. After much soul- Lashley was concentrating on the study tal cues, a view coinciding with that of searching as to whether he should con- of vision. the Gestalt school, to which Lashley was tinue in psychology, he decided in 1934 sympathetic [see ÒThe Legacy of Ges- to burn his boats, borrow money and go year later Lashley was oÝered a talt Psychology,Ó by Irvin Rock and to Chicago to continue his doctoral re- professorship at Harvard Univer- Stephen Palmer; SCIENTIFIC AMERICAN, search under Karl S. Lashley. A sity and managed to take Hebb December 1990]. What Hebb did not The elder scientist was to exert a pro- along. Hebb had to start his research notice, although the results were includ- found inßuence on HebbÕs approach, from scratch, and having only enough ed in a paper he published at the time, above all in his emphasis on physiolo- money for one more year, he sought an was that the dark-reared rats took much gy. Lashley had never doubted that to experiment that could support a thesis longer than normal rats to learn to dis- understand behavior one must Þrst un- no matter how it came out. He con- tinguish vertical from horizontal lines. derstand the brain. As a lab boy in 1910, trived to adapt his interest in the na- Only many years later, after he had he had salvaged slides of a frog brain ture-nurture question to LashleyÕs vi- again changed his ideas about the rela- from the trash heap and tried to Þnd sion project by investigating the eÝects tive importance of innate and learned in the neural connections some clue to of early experience on the development mechanisms, did he appreciate the sig- frog behavior. Lashley performed exper- of vision in the rat. niÞcance of this result. iments to detect memory traces in the Contrary to the empiricist ideas of his Hebb received his Ph.D. from Har- brain, inventing techniques for making masterÕs thesis, Hebb found that rats vard in the middle of the Depression, brain lesions and measuring their loca- reared in complete darkness could dis- when there were no jobs in physiologi- tion and extent. By around 1930 he had tinguish the size and brightness of pat- cal psychology to be had. He therefore become convinced that could terns as accurately as rats reared nor- stayed on for a year as a teaching assis- tant, a post that enabled him to contin- ue his work with Lashley. In 1937 there was still no improvement in the job market, but HebbÕs luck held out. His sister was taking her Ph.D. in physiolo- gy at McGill and heard that Wilder Pen- Þeld, a surgeon who had just estab- lished the Montreal Neurological Insti- tute there, was looking for someone to study the consequences of brain surgery on the behavior of patients. She passed on the information to her brother, and his application for the two-year fellow- ship was successful. He married again and returned to Montreal. The young man who he could run away from his family and become a novelist found himself one of a medi- cal group pioneering the treatment of neurological disorders. PenÞeldÕs specialty was the treat- ment of focal by surgically re-

Copyright 1992 Scientific American, Inc. moving scarred areas of the cerebral er than the corresponding areas in less was widely used for the next quarter cortex. He was acutely aware that he intelligent animals. Yet Hebb was not century. But Hebb was proudest of a was operating on the organ of the mind able to detect intellectual loss in pa- theoretical paper in which he proposed and that a false move could deprive his tients whose frontal lobes had been that adult intelligence was crucially in- patient of , intelligent behavior destroyed by accident or surgical ne- ßuenced by experience during infancy, or even . Although Pen- cessity. This seeming lack of eÝect im- basing his argument on the results of Þeld was not a , his work pressed Hebb deeply and inspired his his research at the Montreal Neurolog- exposed him to the relation between quest for a theory of the brain and in- ical Institute. The paper was virtually the mind and the nervous system. This telligent behavior. ignored at the time, although it is now experience no inßuenced his de- accepted almost as a commonplace, cision to appoint psychologists to his lthough his set him having been embodied in such pre- team and explained the close interest oÝ on fruitful lines of inquiry, school enrichment programs as Head he took in their Þndings. A later work showed that Hebb Start. But the concept was too advanced HebbÕs main responsibility was to had relied too heavily on standard in- for its time: in 1940 most psychologists study the nature and extent of any intel- telligence tests. , one of practically deÞned intelligence as an in- lectual changes in patients consequent his students, who continued the work nate characteristic. to cortical excisions. Such research was he had begun on PenÞeldÕs patients, To reconcile his studies of childhood not new: it began after with found that frontal-lobe lesions often inßuences with the apparent harmless- the psychometric testing of soldiers make it diÛcult for the patient to relin- ness of frontal-lobe lesions, Hebb hy- who had suÝered penetrating head quish a behavior that has ceased to be pothesized that the regionÕs main func- wounds and continued later in patients appropriate. Although they may not be tion was not to think but rather to fa- with brain tumors. In many cases, the detected by intelligence tests, personal- cilitate the tremendous acquisition of lesions produced signiÞcant intellectu- ity changes after frontal-lobe damage knowledge during the Þrst few years of al loss, but their locus and extent were can profoundly aÝect the patientÕs life. life. Experiments to determine the rela- diÛcult to determine. In contrast, sur- At the end of his fellowship at the tive eÝects of early and late brain le- gical removals are more precisely de- neurological institute, Hebb Þnally sions did not always support this idea, Þned, and epileptic scars do not cause found a permanent job at QueenÕs Uni- but it provided a stepping-stone to the widespread damage that bullets or versity in Kingston, Ontario. There, de- HebbÕs later theories. tumors do. spite his heavy teaching load, he kept In 1942 Lashley became the director Hebb soon faced a peculiar problem. up work on the problem of intelligence. of the Yerkes Laboratories of Primate Psychologists then regarded the frontal Together with a student, Kenneth Wil- Biology in Florida, and he invited Hebb lobes of the cerebral cortex as the seat liams, he developed a variable-path rat to join his research team to study chim- of , on the grounds maze as an analogue to human intelli- panzee behavior. Hebb jumped at the that this region is relatively much larg- gence tests. The Hebb-Williams maze chance of doing full-time research with

LIGHT

DIFFUSER TARGET 

SLIDING TUBES



LENS POST MOUNTED ON CONTACT LENS



INACTIVE CONTACT FIRING NEURON LENS

FIRING NEURON, PART OF CELL ASSEMBLY HYPOTHETICAL CELL ASSEMBLY begins with parallel Þbers RETINAL FATIGUE supports the cell-assembly theory by caus- connecting input from the retina to corresponding points in ing images to fade in a peculiar fashion. The apparatus Þxes the primary visual cortex. These neurons, in turn, connect to an image on receptors until their signal decays. Then lines the ÒassociationÓ cortex. Converging input Þres cells and acti- drop out, one or two at a time, until the Þgure is gone. Hebb vates closed loops (dark red ). Synaptic changes ensue that en- argued that each line was represented by a neuronal feed- able the loop to Þre with little input, producing output that rep- back loop. When the retinal signal falls below the critical val- resents to the brain what the eye has seen. ue, the loop stops oscillating, and the line disappears.

126 SCIENTIFIC AMERICAN January 1993 Copyright 1992 Scientific American, Inc. Lashley again, although he was not at Þrst very enthusiastic about work- ing with chimpanzees. LashleyÕs inten- tion was to develop tests of learning and for the animals, while Hebb would study their person- alities and emotional characteristics. Then they would start a program to de- termine how brain lesions aÝected a range of variables. The chimpanzees proved more diÛ- cult to train than Lashley had imagined. The delays meant that no brain opera- tions were carried out during HebbÕs tenure at Yerkes. Nevertheless, he was fascinated by his observations of chim- panzees and said he learned more about EXPERIMENT carried the study of sensory deprivation beyond the human in his Þve years of realm of individual cell assemblies. CuÝs prevented touch, a plastic shield disrupt- watching chimpanzees than at any oth- ed pattern vision and a U-shaped foam cushion attenuated sounds not masked er time since his own Þrst Þve years by the air conditioner in the ceiling. EEG electrodes recorded the subjectÕs brain of life. The apes manifested distinct waves, and a microphone enabled him to report his experiences. The volunteersÕ and a sense of fun that ability to think deteriorated, and some of them even started to hallucinate. tended toward slapstick. Hebb and the other members of the staÝ derived a more cerebral amusement from the ver- who had discovered neural loops, or Repeated activation of any given bal contortions of orthodox behaviorist feedback paths, in the brain. Up to that loop might then strengthen that loop visitors as they attempted to describe point, all psychological theories, wheth- in the following way. If the axon of an the animalsÕ practical jokes and broad er physiological or not, assumed that in- ÒinputÓ neuron is near enough to excite clowning without resorting to Òmental- formation passed through the organism a target neuron, and if it persistently isticÓ . along a one-way track, like food through takes part in Þring the target neuron, the digestive system. Hebb recognized some growth process takes place in one ebbÕs long and close that LorenteÕs looping paths were just or both cells to increase the eÛciency of the many chimpanzees in the what he needed to develop a more real- of the input neuronÕs stimulation. Syn- H primate laboratory taught him istic theory of the mind. apses that behave according to this that experience was not the only fac- Feedback was not entirely new in postulate became known as Hebb syn- tor in the development of personality, learning theory. Almost all models as- apsesÑsomewhat to HebbÕs amuse- including pathological manifestations sumed that the output of the organism ment, it may be said, because this pos- such as phobias. He showed, for exam- inßuences the input in some way, for tulate is one of the few aspects of the ple, that young chimpanzees, born in instance, by enabling the animal to re- theory he did not consider completely the laboratory and known never to have ceive a reinforcing . Unfortu- original. Something like it had been seen a snake before, are frightened the nately, feedback proceeding in this way, proposed by many psychologists, in- Þrst time they are shown one. Chim- through a single path, would operate cluding Freud in his early years as a panzees are also frightened of models slowly and often unreliably. But with neurobiologist. of chimpanzee or human heads or oth- millions of internally connected feed- Nevertheless, HebbÕs postulate was er isolated body parts or of familiar back paths, it would clearly be possible the most clear and formal statement, caretakers wearing unusual clothing. to establish internal models of the en- although in 1949 it was pure specula- Moreover, Hebb was one of the Þrst to vironment that might predict the effects tion. Since then, however, studies of observe the social behavior of captive of possible responses without having to single neurons have conÞrmed that porpoises and to suggest that it im- move a muscle. synaptic strengths do change in some plied a level of intelligence comparable HebbÕs specialization in vision led neurons in accordance with the postu- to that of the apes. His observations him to concentrate his early neural the- late. Hebb may also have been correct may have inßuenced his later conclu- ories on that system. Knowing that the about the mechanism of permanent sion that level of play provides a good point-to-point projection from the reti- change. A former student of his, Aryeh index of intelligence. na to the cortex does not extend be- Routtenberg of Northwestern Universi- LashleyÕs interest in the ways the yond the primary visual cortex, he as- ty, has recently pointed out that a pro- brain categorizes into sumed that the neural relays projected tein associated with neuronal growth is knowledge about the world rekindled into the surrounding cortex in random produced when neurons are stimulated HebbÕs curiosity about concepts and directions, thus scrambling the retinal in ways that increase synaptic strength. thinking. The problem can be rephrased pattern [see ÒThe Visual Image in Mind Hebb assumed that most of the syn- as a question: How does the brain learn and Brain,Ó by Semir Zeki; SCIENTIFIC apses in the cortical lattice are initially to lump one triangle, car or dog with AMERICAN, September 1992]. Such an too weak to Þre spontaneously. To Þre, another even though no two triangles, arrangement could recombine signals they would require the converging of cars or dogs produce the same pattern from diÝerent parts of the imageÑthat stimulation from a number of active of stimulation on sensory receptors? is, they could converge on the same neurons. Some neurons in the lattice The turning point came when Hebb target neuron, causing it to Þre. The re- receive converging inputs and thus Þre read about the work of Rafael Lorente sulting impulses could then return to when a particular pattern of neurons in de N—, a neurophysiologist at the Rocke- the earlier neurons in the path, closing the sensory cortex is Þred by a stimulus. feller Institute for Medical Research, the feedback loops. Some of the activated neurons have

Copyright 1992 Scientific American, Inc. SCIENTIFIC AMERICAN January 1993 127 HEBBÕS INFLUENCE propagated as much through his disciples as through his publi- cations. Here, in a graduate seminar from the early 1950s, Hebb appears at the far right, the author in the foreground. The participants went on to pioneer the new Þeld of physiological psychology.

synaptic connections with one another, which are also strengthened whenever the stimulus is presented. Eventually the connections between the simulta- neously Þring neurons in the lattice be- come strong enough for them to con- tinue Þring one another in the absence of input from the stimulus, creating an internal representation of the stimulus, called a Òcell assemblyÓ by Hebb.

he concept of the cell assembly, in my view, was HebbÕs greatest T contribution to psychological the- ory, not to mention philosophy. It re- vived the 19th-century psychologistsÕ attempt to explain behavior in terms of the association of ideas, a project that the behaviorists had derailed by argu- ing that ÒideasÓ were no more real than the notion of little men inside the head. By so arguing, the behaviorists main- BEN DOAN HELEN MAHUT PETER MILNER JIM OLDS tained that ideas, and thus mentalism, Dalhousie University Northeastern University McGill University Caltech had no place in scientiÞc psychology. (DECEASED) Unfortunately, few seemed to notice that the behaviorists replaced ideas ously, making the system useless. This was related to his cell-assembly theo- with equally insubstantial constructs eÝect was observed in computer mod- ry. Experiments to obtain direct phys- with misleading names, such as Òstim- els of the cell assembly, called concep- iological evidence for the theory were uliÓ and Òresponses.Ó These were not tors, constructed in the 1950s by Na- far beyond the scope of contempo- real events or chains of events but at- thaniel Rochester and his colleagues at rary methodology. (They still are.) In- tributes that became associated with the IBM research laboratory in Pough- stead he tested behavioral predictions one another in some imaginary black keepsie, N.Y. Hebb himself seems never of the theory. He tried, for instance, to box that scientists were forbidden to re- to have set Þnger to a computer to strengthen his earlier conclusions on fer to as the brain. Hebb put a stop to his idea that random nerve nets could the inßuence of rearing on adult intelli- this charade by showing, in principle at organize themselves to store and re- gence. Most of the results supported least, that ideas could have just as Þrm trieve information. But such so-called his theory that animals raised in an en- a physical basis as muscle movements. neural nets later inspired many comput- riched, or more , environment They could consist of learned patterns er models, from the perceptron to par- would, in later life, outperform animals of neuronal Þring in the brain, initially allel distributed processing, and have raised in bare cages. driven by sensory input but eventually even found applications in industry. There was one embarrassing excep- acquiring autonomous status. tion. Litters of pure-bred Scotties were In its original form the neural theory y the time The Organization of split, and half the pups were reared as was undoubtedly too simple to have Behavior reached publication, pets in the homes of members of the worked. A major problem was that the B Hebb was back in Montreal as staÝ and half were reared in cages in cell assembly did not incorporate inhi- chairman of McGillÕs psychology depart- the laboratory. Hebb was not fortunate bition, because contemporary science ment. Ten years later, when he stepped in the choice of his puppy, Henry. It did not recognize it. Sir John C. Eccles, a down as chairman, he had forged one of was congenitally incapable of Þnding very inßuential neurophysiologist at the the strongest departments in North its way around, invariably got lost as Australian National University in Can- America. He found it easier to build soon as it was out of sight of the house berra, was still vigorously denying the what he wanted because the depart- and had to be recovered from the dog existence of inhibitory synapses. More- ment was almost nonexistent when he pound on several occasions. Naturally, over, many important connections of began, and he turned out to be adept Henry turned out to be near the bot- the neocortex had not yet been discov- at campus politics and soon discovered tom of the class when, as a full-grown ered, and the functional signiÞcance of how to use his growing reputation to dog, it was tested in a maze. the diversity of cortical neurons was apply pressure where it would do the In a related series of experiments, only hinted at. most good. It is perhaps signiÞcant that Hebb investigated the eÝect of impov- Without inhibiting factors, however, he was also one of the best chess play- erished sensory input on the behavior learning would strengthen synaptic con- ers at the university. of , including human volunteers nections until all neurons Þred continu- Most of HebbÕs research at McGill [see ÒThe Pathology of Boredom,Ó by

128 SCIENTIFIC AMERICAN January 1993 Copyright 1992 Scientific American, Inc. of his theory. These essays are refresh- ingly forthright. On , for example, Hebb wrote: ÒWe still need an Ajax to stand up and defy the lightning and ask, What is the evidence? when some authority informs the public that believing in Santa Claus is bad for chil- dren, that comic books lead to psycho- logical degeneracy, that asthma is due to a hidden mental illness.Ó Hebb built his department and his Þeld by capturing the interest and imag- ination of the best students at an early stage. He taught the introductory course himself, making it immensely popularÑ at one point it numbered 1,500 stu- dents, about half the yearly undergradu- ate enrollment. Many future professors of psychology found their calling in these lectures. Like most of what Hebb did, his course was unique; no textbook at the time came close to including the material and ideas he dealt with, so he wrote his own. The Þrst edition of A Textbook of Psychology appeared in 1958. In contrast to the majority of in- troductory texts of the day, it had more ideas than pictures. Hebb also gave a graduate seminar that was attended by every psychology G. ROLFE MORRISON UNIDENTIFIED SETH SHARPLESS DONALD O. HEBB graduate student at McGill over a peri- McMaster University University of Colorado, Boulder McGill University od of 30 years. It was famous not only (DECEASED) for its stimulating discourse but also for HebbÕs ever-present stopwatch and Woodburn Heron; SCIENTIFIC AMERICAN, fell on the same place. As the receptor the slips of paper on which he noted January 1957]. Students were paid gen- cells become fatigued, the image fades incorrect pronunciations and other er- erously to undergo severe sensory dep- and disappears, but not all at once. rors of presentation. It was HebbÕs am- rivation for as long as they could stand Usually entire lines disappear sudden- bition never to have a McGill student it (none lasted even a week). Their ability ly, one or two at a time, until the entire overrun his or her allotted time at a to think began to deteriorate, and some Þgure is gone. Hebb explained the phe- meeting, and on the whole he was suc- of them even started to hallucinate. The nomenon by saying that each line is cessful. McGill honored Hebb in 1970 Korean War was then in progress, and represented by neuronal activity cir- by naming him chancellor; he became many workers attempted to use such culating in a closed loop. The activity, the only faculty member ever appoint- isolation experiments to understand once started, continues even after the ed to that position. and combat the ÒbrainwashingÓ tech- input from the retina has decayed to a In 1977 Hebb retired to his birthplace niques employed by the Chinese. low value because of feedback around in Nova Scotia, where he completed his Hebb also pursued his old idea that the loop. But at some critical value the last book, Essay on Mind. He was ap- early brain injury should be more dam- reverberation stops abruptly, and the pointed an honorary professor of psy- aging than injury in an adult. But the line disappears. These experiments do chology at his alma mater, Dalhousie, results were rendered uncertain by sev- not provide conclusive evidence for the and regularly participated in colloquia eral factors, the most important being cell assembly as Hebb envisaged it. Yet there until his death, at 81, in 1985. the capacity of the young brain to reor- even if HebbÕs version should turn out ganize itself. For example, if an infant to be incorrect, it would not diminish sustains an injury in an area of the left the value of his idea that some neural hemisphere that is important for speech activity continues to symbolize an ob- FURTHER READING in the adult, the right hemisphere takes ject even after the object has stopped THE ORGANIZATION OF BEHAVIOR: A NEU- ROPSYCHOLOGICAL THEORY. D. O. Hebb. over this function, and speech is not se- stimulating the sense organs. John Wiley, 1949. riously impaired. But if an adult sustains ESSAY ON MIND. D. O. Hebb. Lawrence damage in the same area, the result may ad The Organization of Behavior Erlbaum Associates, 1980. be a permanent loss of language skills. consisted only of the chapters PARALLEL LEARNING IN AND MA- Because of such problems with the H in which Hebb criticizes current CHINES. G. Ferry in New Scientist, Vol. 109, study of , Hebb came to be- approaches and elaborates his cell-as- No. 1499, pages 36Ð38; March 13, 1986. lieve that the best evidence for the cell sembly theory, it is likely that few peo- TEXTBOOK OF PSYCHOLOGY. Fourth edi- assembly came from experiments on ple would have read it. The bookÕs ap- tion. D. O. Hebb. Lawrence Erlbaum As- sociates, 1987. retinal fading. Images of simple Þgures peal lies in its second half, in which MIND AND BRAIN. Special Issue of Scien- were projected onto the eye by a very Hebb discusses , , tific American, Vol. 267, No. 3; Septem- small lens system attached to a contact mental illness and the intelligence of ber 1992. lens, ensuring that the image always humans and other species in the light

Copyright 1992 Scientific American, Inc. SCIENTIFIC AMERICAN January 1993 129