From the retina to neural representation

From physics to biophysics

Yoon was born in Seoul, Korea, in 1939. He entered Seoul National University in 1958 and graduated with a B.S. in physics, with honours, in 1962. He began graduate study in theoretical physics, but in May 1962 he was conscripted into the Korean Army, where he served until February 1965.

That same spring, he had been offered the chance to study biophysics at the University of California, Berkeley. Military service put the plan on hold, but letters reproduced in his memoir, Tribute to Mentors and Friends, show that the Berkeley scientists Cornelius Tobias and Hardin Jones kept encouraging him through those years. He arrived at Berkeley in 1965, first as a research assistant in Tobias’s bio-medical group at the Lawrence Radiation Laboratory and then as a W. O. Donner Predoctoral Fellow in Biophysics.

At first he explored several directions, including molecular biology and mathematical biophysics. He eventually became fascinated by the nervous system and joined the visual neurophysiology work of Horace Barlow, one of the leading figures in twentieth-century research on neural coding and vision.

His Ph.D. dissertation, completed in 1969, was titled “Neural responses of the cat’s retina to light.” His thesis committee was chaired by Barlow and included the molecular biologist Gunther Stent and the Nobel Prize–winning physicist Owen Chamberlain, a combination that reflected the territory Yoon was working in, between physics, biology and neuroscience.

How does the retina encode the visual world?

Yoon’s Berkeley experiments addressed one of the fundamental problems of sensory neuroscience: how does the visual system stay sensitive across the enormous range of light intensities found in the natural world?

Recording from individual retinal ganglion cells in cats, he studied the relationship between illumination, adaptation and neural response. His early papers included “Reversal of Weber’s law for an extraordinary unit in the cat’s retina” in Vision Research (1970), and work with Barlow and William Levick on ganglion-cell responses to single quanta of light. His principal dissertation paper, published in the Journal of Physiology in 1972, showed how adaptation changes both the sensitivity and the timing of ganglion-cell responses.

The underlying question was already one of neural representation: how does physical light become an informative pattern of neural activity? A move to Caltech in 1969 would take Yoon from the coding performed by single neurons to the physical organization of the brain itself.

Roger Sperry and the problem of neural specificity

After Berkeley, Yoon joined the Division of Biology at the California Institute of Technology as a research fellow in the laboratory of Roger W. Sperry, who would later share the 1981 Nobel Prize in Physiology or Medicine.

Sperry was investigating one of developmental neuroscience’s deepest questions: how do billions of growing nerve fibers find their proper targets and build an orderly nervous system? His chemoaffinity hypothesis proposed that neurons and their targets carry molecular identities that allow the right connections to form.

Yoon began working with the goldfish retinotectal system. When the goldfish optic nerve is cut, the axons of retinal ganglion cells regenerate into the brain and re-establish functional connections with the optic tectum, the main visual centre of the fish brain. An investigator can therefore disrupt the physical organization of the visual system, let it rebuild itself, and then record which map of the visual world has emerged. The question that guided Yoon for years was simple to state:

How does a regenerating optic fiber know where it belongs?

The surprising plasticity of the visual map

Yoon’s first experiments showed that the adult retinotectal system was far more plastic than a fixed wiring diagram would suggest. His 1971 paper in Experimental Neurology demonstrated that retinotectal projections reorganized after surgical alteration of the optic tectum. Later experiments showed that retinal projections could be transposed onto normally foreign tectal territory, and that reorganized maps could readjust again when conditions changed.

One striking form of this reorganization is known as field compression. When part of the optic tectum is removed, the projection from the whole retina can eventually compress itself onto the remaining tissue while preserving an orderly map of visual space.

This posed a problem for a rigid interpretation of neuronal specificity. If each retinal fiber had a single predetermined target, how could an entire visual map reorganize onto a radically altered brain? The question occupied much of Yoon’s research in the 1970s.

Rotating a piece of the brain

One of Yoon’s most distinctive experiments approached the problem by physically changing the geometry of the brain. He removed a rectangular piece of optic tectum from an adult goldfish, rotated it 180 degrees and put it back. After cutting the optic nerve, he let the retinal fibers regenerate and then mapped the restored visual projection electrophysiologically.

There were two possible outcomes. If the fibers connected according to the graft’s new position, an orderly map would appear across the whole tectum. If the transplanted tissue somehow retained information about its original orientation, the fibers would treat it differently from the surrounding tectum.

The second happened. Within the rotated graft, the regenerated visual projection was reversed relative to the surrounding tectum: the transplanted tissue had kept its original spatial polarity. Yoon reported the result in 1973 in the Journal of Physiology as “Retention of the original topographic polarity by the 180° rotated tectal reimplant in young adult goldfish.”

The experiment suggested that the tectum was not a passive surface onto which retinal fibers projected. The tissue itself carried positional properties that shaped the pattern of incoming connections. In Tribute to Mentors and Friends, Yoon describes the tectum as an “active accommodator” of incoming optic fibers, and he called its directional property its “topographic polarity.”

Dalhousie and “topographic addresses”

Yoon joined the Department of Psychology at Dalhousie University in Halifax as an associate professor in November 1972. From 1974 to 1978 he was an I. W. Killam Research Professor, a position that gave him unusual freedom for research, and in 1978 he became Professor of Psychology. His laboratory in Halifax became the centre of his retinotectal research.

The rotated-graft experiments led him to a more ambitious idea. Directional polarity alone could not explain how retinal fibers found precise locations within the tectum, so he proposed that tectal neurons might carry what he called “topographic addresses”: positional information distributed systematically across the tissue.

Yoon was careful to present this as a description of an unexplained biological phenomenon, not a known molecular mechanism. As he writes in his memoir, “The real physical nature of the hypothesized ‘topographic address’ is entirely unknown.” That distinction shaped the direction his career took next.

Specificity and plasticity

Yoon’s experiments increasingly confronted an apparent paradox. The rotated grafts suggested stability: tectal tissue kept its original positional character despite being cut out and turned around. Field compression demonstrated plasticity: when the size and boundaries of the tectum changed, the neural map could reorganize dramatically. How could both be true?

He proposed a “neuronal re-specification hypothesis”: neurons normally carry stable topographic addresses, but those addresses can themselves change systematically when the overall boundaries of the tectum are altered.

He then designed experiments that could have disproved the idea. If reorganization appeared almost immediately after surgery, biological re-specification would seem implausible; if it developed over weeks, an underlying change in the tissue would be more credible. The results showed that orderly compression developed slowly. When regenerating fibers entered a half-tectum about 40 days or more after surgery, the remaining tissue could accommodate fibers from the entire retina in an orderly, compressed map.

In further experiments, Yoon combined field compression with rotated grafts. The same piece of tissue could keep its original topographic polarity while taking part in the plastic reorganization of the larger map. In his memoir he describes this as evidence for the compatibility between the rigid retention of original topographic polarity and the plastic readjustment of the adult visual system. The relationship between specificity and plasticity was arguably the central scientific problem of his work in the 1970s.

From phenomenology to mechanism

Yoon was aware of the limits of these experiments. They showed that positional information existed and behaved in particular ways, but not what that information physically was.

A memorable encounter in 1977 brought the problem into focus. Stephen Heinemann invited Yoon to present his research at the Salk Institute. During the seminar, someone in the audience called out “bravo” at one of his results; Heinemann later told him it was Francis Crick. Yoon met Crick privately the next morning and explained that his experiments characterized phenomena such as topographic polarity and topographic addresses, but that he did not understand their physical basis. As he recalls in his memoir:

“My work was merely a phenomenological formulation of scientific questions with experiments.”

He told Crick that he wanted to pursue the problem at the molecular level. Crick encouraged him, writing shortly afterward that a collaboration between Yoon and Heinemann “could be very fruitful.” In the summer of 1978, Yoon returned to the Salk Institute as a visiting research scientist sponsored by Heinemann, and his research turned decisively toward cellular and molecular mechanisms.

Axons, glia and the biology of regeneration

A series of visiting appointments widened Yoon’s experimental range: the National Institute for Medical Research in London with R. M. Gaze (1976), the University of Michigan with Bernard Agranoff (1977), the Salk Institute with Heinemann (1978), McLean Hospital and Harvard Medical School with Larry Benowitz (1978–79), and Berkeley again with Gunther Stent (1979).

With Pamela Johns and Agranoff, he studied the directed outgrowth of regenerating optic fibers in culture, published in Nature in 1978.

With James Stevenson at Dalhousie, he asked a different question: what happens to cells in the target tissue when optic fibers regenerate into it? They found enhanced cell proliferation in the optic tectum during optic-nerve regeneration, and in 1981 identified an important population of the dividing cells as radial glia. Their experiments suggested that regenerating optic fibers could induce glial cells in the adult brain to divide.

Regeneration, in other words, was not simply an injured axon growing back toward an inert target. The target tissue itself responded.

Proteins of regeneration

Yoon’s collaboration with Larry Benowitz approached the same problem at the molecular level. In 1981, Benowitz, Victor Shashoua and Yoon showed that optic-nerve regeneration is accompanied by specific changes in the proteins that retinal neurons make and rapidly transport along their axons.

In 1983, in a paper in Science, Benowitz, Yoon and E. R. Lewis showed that some of these molecular changes are regulated by interactions between the regenerating optic fibers and the optic tectum: the molecular state of a regenerating neuron depends partly on its contact with its target. Yoon, Benowitz and F. A. Baker extended the finding in Brain Research in 1986.

Over a decade, the question had moved from where regenerating axons connect, to what properties of the target determine those connections, to what cellular and molecular interactions take place between regenerating neurons and their targets. The molecular work was less a departure from his earlier research than an attempt to reach one level deeper into the same problem.

A wider view of regeneration

Through the 1980s, Yoon and his collaborators studied other parts of the regenerating system. With Stevenson he characterized radial glia, ependymal cells and periventricular neurons in the goldfish tectum. With J. D. Radel he studied the ultrastructural changes in regenerated optic-fiber terminals. With D. B. Henken he investigated the proliferation of radial glia and found that optic-nerve injury also alters the proliferation of rod precursor cells in the retina.

By the end of the decade, the program had grown from the geometry of neural maps to the responses of axons, synapses, glia, precursor cells and proteins to injury and regeneration. Yoon summarized much of this work in two chapters written for volumes honouring his principal mentors: “Neural reconnection between the eye and the brain in goldfish” (1989), in Brain Circuits and Functions of the Mind, a Festschrift for Roger Sperry, and “On reformation of visual projection: cellular and molecular aspects” (1990), in Vision: Coding and Efficiency, a Festschrift for Horace Barlow.

From neural maps to mental representations

By the late 1980s, Yoon’s interests had begun to move in another direction. After a sabbatical at the University of Cambridge in 1987, again with Horace Barlow, he began cognitive experiments on how young children develop mental representations using linguistic signs and other abstract symbols. A 1990 presentation at the Boston University Conference on Language Development, “Development of narrative representation by children,” came out of this work.

It was a considerable methodological departure, but there is a conceptual thread. The early Yoon studied how locations in visual space come to be represented in an ordered physical map in the brain. The later Yoon asked how more abstract information, such as objects, words, sentences and narratives, comes to be represented in the mind.

Into the human brain

Yoon retired from Dalhousie University in 2003. Two years later, he and the Dalhousie radiologist Larry Gates published “Distinct and shared cortical regions of the human brain activated by pictorial depictions versus verbal descriptions: an fMRI study” in NeuroImage.

Gates and Yoon compared brain activity while subjects processed objects and pictures, which they called analogical representation, with activity while they processed words and sentences, or symbolic representation. They found both distinct and extensively overlapping cortical regions. About 96% of the relevant cortical volume overlapped for single objects versus single words, falling to about 86% for the more complex comparison of arranged objects versus sentences. Pictorial and linguistic representation, the study suggested, draw heavily on shared neural structures and become more differentiated as they grow more complex.

Published after his retirement, it was a fitting last contribution to a career that had begun nearly four decades earlier by recording single retinal neurons responding to photons.

A scientific life across levels of explanation

Seen as a whole, Yoon’s career crossed an unusually wide range of neuroscience. It began with sensory neurophysiology: how does the retina encode light? At Caltech and Dalhousie it turned to neural organization: how do retinal fibers build an ordered map of visual space in the brain? His transplantation experiments exposed the relationship between specificity and plasticity: how can neural tissue keep its positional identity while reorganizing after radical disruption? His cellular and molecular research asked what allows damaged axons to regenerate, recognize their targets and rebuild synaptic connections. His late work asked how more abstract representations are organized in the human mind and brain.

Across these phases, his published work shows a recurring fascination with representation, organization and reconstruction in the nervous system.

Mentors and a scientific tradition

Three senior neuroscientists were especially important to Yoon’s formation. Horace Barlow supervised his doctoral work and introduced him to quantitative visual neurophysiology. Roger Sperry provided the intellectual and experimental setting in which he began the retinotectal research that became his principal independent contribution. Stephen Kuffler, though never his formal supervisor, became an important informal mentor as Yoon moved toward scientific independence. Kuffler followed his research, encouraged his presentations, introduced him to scientists and opportunities, advised him on appointments and research directions, and helped bring him into the emerging community of neurobiology. In Tribute to Mentors and Friends, Yoon recalls Kuffler treating a “humble foreign stranger as if I were one of his students.”

From Barlow he learned to investigate the neural coding of sensory information; from Sperry, the problem of neuronal specificity. Yoon then made the tension between specificity and plasticity experimentally concrete, showing that neural tissue could hold persistent positional properties while also displaying a remarkable capacity for reorganization, and went on to seek cellular and molecular explanations for phenomena first found through electrophysiology and microsurgery.

In his memoir, Yoon describes this path with characteristic modesty, presenting himself as a student fortunate to meet exceptional mentors. The published record also shows an independent experimentalist willing to devise unusually direct, and sometimes technically audacious, tests of difficult biological ideas. Removing pieces of an adult brain, rotating or transplanting them, letting its nerve fibers rebuild their connections, and then asking the rebuilt nervous system where it believed the visual world to be was an unusually physical approach to neuroscience, perhaps a reflection of the physicist he had been at the start.

Epilogue

Yoon’s scientific career began with a fundamental problem of sensory biology: the transformation of light into neural activity. It ended with another problem of representation: the relationship between pictorial and linguistic information in the human brain.

Between them lay his principal experimental contribution, showing how the visual system of the goldfish can regenerate and reorganize while still preserving an ordered relationship between the eye and the brain. His experiments helped illuminate a basic property of nervous systems: their organization is neither simply fixed nor infinitely malleable. Neural tissue can preserve remarkably specific information about its identity while keeping an equally remarkable capacity to rebuild itself after injury. Understanding how specificity and plasticity coexist was at the heart of Myong G. Yoon’s scientific work.

In his memoirs: Tribute to Mentors and Friends