Interview: Nobel Laureate Eric Kandel

Does psychotherapy work? We'd be a lot more certain if we slapped a little science on it.

By Susan Kruglinski|Sunday, April 2, 2006
Eric McNatt

Neurobiologist Eric Kandel's early fascination with how the mind works led him into training as a psychiatrist interested in psychoanalysis. But in the 1960s he traded a therapist's sport jacket for a lab coat and began studying one of the slimiest creatures on Earth. At a time when brain researchers thought nothing could be learned from invertebrates, Kandel stunned the fledgling world of neuroscience by uncovering the mechanisms of memory in sea slugs, which earned him a Nobel Prize in 2000. He also coedited Principles of Neural Science, the book every medical student in America is required to read—all 1,414 pages of it.

In his new book, In Search of Memory: The Emergence of a New Science of Mind, Kandel examines the convergence of four critical fields—behaviorist psychology, cognitive psychology, neuroscience, and molecular biology. Still busy trying to coax more information out of sea slugs, Kandel also serves as a professor of biochemistry, physiology, and psychiatry at Columbia University, where his laboratory is located.

Your most recent lab work involves the potent emotion of fear. What have you learned?

We have identified some of the genes in the mouse that are important for both learned fear and instinctive fear. We've shown that by knocking out the stathmin gene, we can produce a mouse that is relatively fearless.

How does a fearless mouse behave?

Normally, when you put a mouse into an open field, it walks along the edge of the field, where the walls are—the mouse is in an enclosed chamber with walls around it—because it's afraid of being attacked by an intruder, and it makes occasional dashes into the center to make sure it's not missing some food or an interesting sexual partner. If you frighten a mouse, it stays in one corner and doesn't move at all. If you knock out this gene that is important for fear, the animal spends a lot of time in the center. It is no longer afraid.

Does the knockout shut down instinctual fear?

Yes, but it works on learned fear too.

Can you imagine using this research to benefit humans?

This gene could become a new target for antianxiety agents. We've also opened up the biology of happiness, showing that you can produce an animal that is more relaxed than normal, and this recruits activity in the pathways that are involved in positive reinforcement. That might open up a biology of security and comfort.

How do you study happiness?

In order to produce learned fear, you take a neutral stimulus like a tone, and you pair it with an electrical shock. Tone, shock. Tone, shock. So the animal learns that the tone is bad news. But you can also do the opposite—shock it at other times, but never when the tone comes on. Under those circumstances the tone indicates security and safety. We find that the animal acts as if it is content and secure, even more than it does without any shocks whatsoever. When we looked in the brain, we found that not only were the pathways that mediated fear turned off but also that happiness pathways were activated. The caudate nucleus, a part of the brain that mediates the effects of drugs that make you feel good, is lit up by this procedure.

So fear and happiness are part of the same system?

That's right, but I want to be clear that the knockout mice and the happiness procedure are not identical. They are different experiments. We showed that fear comes to a certain neural circuit and that there are genes that control that neural circuit, and you can turn that circuit on and off with specific genes. Period, end of paragraph. In a separate set of studies, we looked to see whether we can behaviorally—without manipulating genes—produce the opposite, and that is happiness. And that is how we got onto this paradigm. They may be related—it may be the same genes that shut off fear and turn on happiness, but we don't know that yet.

How is your research in memory related to fear and happiness?

Let's assume that you have a traumatic experience in childhood. Let's say you were sexually abused. You may or may not remember the cognitive components of it—the molester bothering you. But there are associated with the experience a series of autonomic and emotional changes that are implicit. And that component is learned fear. Seeing a strange person may therefore elicit in you a response that may be completely inappropriate—that strange person may have nothing to do with the molestation event, but somehow there are similarities that you recognize that bring this back. You learn emotional experiences as much as you learn cognitive experiences, except that they are more unconscious. Sometimes one represses the cognitive component of it, but it's often more difficult to repress the emotional component.

You have written that your training in psychotherapy influences your neurobiological work. Isn't there a split between psychotherapists and neurobiologists?

Yes, but I think it's a temporary one and perhaps an unnecessary one. I'm on an advisory board for the Ellison Medical Foundation, which is exploring whether one can use MRI brain imaging to evaluate the outcome of psychotherapy.

Haven't you questioned the value of psychoanalysis?

Early in my career, I was disappointed that psychoanalysis was not becoming more empirical, was not becoming more scientific. It was primarily concerned with individual patients. It wasn't trying to collect data from large groups of people who have been analyzed.

Will the Ellison project address this kind of problem?

What our study group is discussing is whether or not the time is ripe to use brain imaging to evaluate the outcome of psychotherapy. There are now two forms of psychotherapy that have been medically proven to be effective. One is cognitive behavioral therapy, developed by Aaron Beck at the University of Pennsylvania. The other is interpersonal therapy, which was developed by Myrna Weissman here at Columbia. Those are two scientifically validated forms of short-term therapy. In 20 sessions you can see improvements in mildly and moderately depressed patients. And there have been some preliminary studies with obsessive-compulsive neurosis where you can see a metabolic abnormality in the caudate nucleus in imaging. If you treat people with psychotherapy and they get better, that metabolic abnormality is reversed, which is the same thing that happens if you give them pharmacological treatment, like Prozac. So that's encouraging. We want to see whether there is a science here, whether or not the foundation should get interested in it. And people like Tom Insel, head of the National Institute of Mental Health, are going to participate as evaluators.

No one has ever conducted studies like this?

It has never been done systematically. We would like to consider doing it on a larger scale, maybe having several universities studying patients with different diagnostic categories to see how a controlled psychotherapy trial produces physical brain changes as a result of treatment.

Why has no one used imaging techniques to study psychotherapy before?

The field is very young. It's only recently that people got confident that psychotherapy under these circumstances works. The imaging methodology is relatively new, and the resolution is not that great yet. We need biological markers for each mental illness in order to see whether or not they can be reversed. So there are lots of technical problems. You know, our understanding of the mind is at a very early stage. These are the most difficult problems in all of biology.

Which illnesses might be considered for study?

Primarily obsessive-compulsive neurosis and anxiety disorders, such as post-traumatic stress disorders.

When did you first begin questioning psychotherapy?

I guess when I was at the National Institutes of Health, and when I came back into my psychiatric residency, which was in the 1960s. I was 30 years old. I was training as a psychiatrist. I had a personal analysis as part of my training. But it was not my personal analysis that caused me to question it. I actually benefited greatly from it. It was the fact that psychoanalysis as a discipline was not becoming scientific.

What do you think of psychotherapy these days?

Well, it is a little chaotic because there are lots of competing therapies out there. We might want to compare modes of therapy. I mean, who knows which is best? It's possible that you might benefit from one, and I might benefit from another. The different kinds of patients, different kinds of disorders, might be selectively treated by one kind of psychotherapy versus another. But that, I think, requires an independent standard for evaluation. And I think imaging might be one of the tools that provides that.

Do you think that some of the therapies are problematic?

I have no way of knowing that. I think it's problematic that people have not gone to the trouble to study this.

Has the psychology-neurobiology split hurt patients?

I think that psychopharmacological treatments have revolutionized psychiatry. On the other hand, I think anyone practicing psychiatry realizes that drug therapy is very effective but not perfect, that there are some patients who don't benefit from it, that there are some patients who benefit from drugs together with psychotherapy, and that in some cases psychotherapy by itself works. There was a time when psychoanalysts wanted nothing to do with biology, and biologists didn't want to touch psychoanalysis. That's changing. Now we need to have a systematic approach to the psychotherapeutic component of treatment, just as we have to the psychopharmacological aspects of treatment. That has not been done. What the Ellison Foundation and I are hoping to encourage is a more holistic approach to psychiatry, in which psychotherapy is put on as rigorous a level as psychopharmacology.

What does a psychotherapist need to know?

First, I think that people working in psychiatry should have a background in neuroscience, because psychiatry is in fact a form of clinical neuroscience. Second, that insofar as there are psychotherapies that are made available to patients, they should be shown to be effective. I am proposing a demanding criterion: that you be able to detect abnormalities in patients beforehand by such brain-imaging techniques as functional MRI [which measures blood flow in the brain], and then use imaging to see whether or not there is a change in those markers for the disease as the therapy progresses. Therapists need not necessarily use the insights of biology in their therapy, but they should be aware of what the indices are, have their patients imaged, and be able to follow the outcome.

Has any recent research in neuroscience surprised you?

Oh, my gosh, there are lots of things. For example, the study of decision making in the brain by people like Paul Glimcher is very interesting. William Newsome's work on the importance of value in decision making. Giacomo Rizzolatti's work on empathy, on how monkeys copy one another, is of profound significance. The work on smell by Richard Axel and Linda Buck is extremely interesting. I think the work on the molecular basis of sociological phenomena by Tom Insel and Cori Bargmann is very interesting. And Tom Jessell's work with neuronal networks is extraordinarily important. Sten Grillner's work on the workings of complex neurocircuitry is also of extreme importance. I would say the progress in understanding motor systems, the cognitive role in motor systems, is a brilliant advance and has revolutionized our understanding of how the nervous system is wired.

Does the research trickle down to therapists?

Yes. I would think that residents in psychiatry should be trained in neuroscience, like residents in neurology are. And this is happening. In fact, I see them as being sort of interrelated disciplines. I mean, the modes of therapy are different, and the character structure of the practitioners is different. But they are both treating the brain as an organ that is the target of disease.

But that isn't really happening, is it?

I don't think that's fair to say. Fields move slowly. It's happening. Nothing happens as fast as I would like it to. Even Discover magazine doesn't publish things as rapidly as I would like it to! [laughs]

What are the big unanswered questions in neuroscience?

I think we need to understand how sensory information is translated into action. We need to understand how unconscious mental processes develop. Where do they occur? What are the processing steps? What is the nature of decision making? Of free will? Can we get a vantage point on consciousness?

How close are we to understanding consciousness?

I think we have not made much empirical progress. But I think we have made a fair amount of conceptual progress. The work of Gerald Edelman and Antonio Damasio and of Christof Koch and Francis Crick has been influential in getting people to think about these problems in a useful way.

What do you think researchers will find consciousness to be?

Oh, my gosh. I have no guesses. I think it's a very deep problem, and I don't really have any original ideas about that.

Comment on this article