Welcome to the inaugural post of our interview series. We are starting off the series with an interview with Dr. Charles Legéndy
, a scientist whose research and career we thought would be particularly interesting to our members. We asked him to discuss his beginnings in neuroscience, his perspective on the future of the field, and give advice for researchers starting their careers today. Our sincere thanks to Charles for going first!
What are the major influences towards your study of neuroscience?
When I was young, and fairly crazy, and was still midway through writing my PhD thesis in physics, I decided to quit physics and go into brain research. Life-changing decisions are sometimes made on the basis of chance encounters and freak accidents; and in retrospect this one of mine was a pretty good example.
A physicist named Brian Pippard, fairly famous at the time in solid-state physics, gave a speech at the festive opening of a new IBM research laboratory at Yorktown Heights, NY, where he said, in effect, that the field of physics was already pretty well cleaned up. Of course this turned out to be completely wrong; but, all the same, at the time it did appear to us graduate students that relatively many physicists were chasing relatively few unsolved problems.
In contrast, it was widely agreed that in brain research the exact opposite was the case. In physics, Newton Heisenberg and Schroedinger had already come and gone -- in brain research, they hadn't come yet. With a little bit of luck, I could be one of those! And of course what young science student could resist that kind of a challenge!
How has the field changed since you developed your career in the 60’s and 70’s?
Experimental techniques have made spectacular advances, reflecting the technological revolution of the last decades. But when it comes to theoretical insight, the field has remained, more or less, unchanged. The most elementary questions of the mind-body problem still stand unanswered. How do our mental images look in terms of neurons? How do the spikes emitted by neurons make up typical events of macroscopic processing in the brain?
One big difference between then and now is that the problem of insufficient theoretical understanding is now more widely recognized. Theoretical brain research, which was essentially without funding during the years when I started, now receives some amount of government money; although much of the money is reserved for computer science departments. The part explicitly earmarked for brain research is to a great extent applied toward the field known as "computational neuroscience" -- but that field is somewhat timid and appears to be kept in the shadow of experimental projects. It is mainly used as a tool of improved interpretation of experimental points, utilizing the tools of computer simulations and (typically) based on simple and often-used models.
However, computer simulation has never been a substitute for straight theory which, when well formulated, transparently obtains results from assumptions.
What are some of the highlights of your research in neuroscience?
There were three papers that stand out in my mind: a theory paper from 1967 on cell assemblies, an experimental paper from 1985 on the spontaneous firing of cells in the visual cortex, and a theory paper in 1970 on the role of low-probability firing events in the brain.
The 1967 paper showed that cell groups, if the conditions are right, can be made capable of a chain-reaction-like mode of firing called "ignition," and act as threshold devices more reliable than their individual elements. The ignitions of such groups can carry the individual signature of the groups, and can in that way send out signals recognizable to the neurons and networks receiving them.
The 1985 paper, written with the neurosurgeon Mike Salcman, was inspired by an experiment I did at the Max Planck Institute in Goettingen, where I tried to see whether pattern vision caused any measurable difference in the firing of neurons in the visual cortex. I recorded from cells in unanesthetized cats freely looking around in the lab, then from the same cells when the eyes were covered with translucent tissue paper, and found that there was not much difference except that the tissue paper somewhat reduced the average spike rate.
Then I ran the same spike trains through a computer program which looked for epochs of elevated spike rate that were "surprising" in the sense that they strongly contradicted the assumption that the firing was a random (Poisson) process. The result was that when the cat was looking around the spike trains contained many super-suprising events, and when its eyes were covered they didn't. When other laboratories began doing "Poisson surprise runs" in other preparations, they found out that the validity of the finding extended well beyond the visual cortex. Since then Poisson surprise runs have become a useful diagnostic tool, sometimes able to detect diminished functionality in certain brain regions.
The 1970 paper was, in retrospect, written much too early for what it undertook to do, but it did at least succeed in introducing some of the concepts underlying the work I was finally able to complete in my recent book, Circuits in the Brain
How does your new book, Circuits in the Brain, relate to the rest of your work?
It is much more ambitious than any of my previous work, because it attempts to spell out a whole new way in which brain function can be analyzed. The book only deals with primary visual cortex, because it is the best-understood part of the brain, but I expect that its methods can be applied equally well in other parts.
In many places, the book makes assumptions where data are unavailable, in an effort to present a more coherent picture; and, naturally, it is unlikely that all of the stated assumptions will prove correct. Accordingly, I envision the main contribution of my book to be along the lines of (for instance) the Hamiltonian formulation in physics, which is an analytic tool and can be used equally well to write good theories and bad ones. But in either case the formulation guards against certain mistakes which are quite easy to make without it.
What do you see happening in the field ten years from now?
I am optimistic. I think ten years from now we will be down to a much finer level of functional detail. One may hope that people will begin to explain the way certain DNA subsequences guide nerve growth and other details, and obtain information about the brain through them. Once a solid method of theoretical analysis becomes available (and once computerized methods are developed for reproducing some of the complex thought processes), progress will be very fast.
What words of advice do you have for students entering the field of neuroscience?
As a student, I think, you should keep your eyes open for new analytic methods, not only for new experimental methods. The field of neuroscience is at this moment almost drowning in unexplained data. This means that the "art of explanation," in other words the art of fitting the many pieces together, will be the key to the next great discoveries.
--Interview by Melissa Higgs
with Dr. Charles Legéndy, a senior member of the research faculty, Columbia Psychology Department
To suggest names/themes for future interviews, please contact Melissa.