So the two years that I spent at Harvard in my psychiatric residency, between leaving the NIH [National Institutes of Health] and joining Ladislav Tauc in Paris, were extremely influential on many levels. To begin with, I learned a lot of psychiatry. In particular, I learned about the strengths and weaknesses of psychoanalytical approaches to it. I realized how unempirical psychiatry was and how no attempt was made to relate psychoanalytical thinking to brain science. This also was the period in which we had our first child, so it was really wonderful from that point of view. From a scientific point of view, it gave me a chance to set up my first lab, and to run experiments on the hypothalamic neurons, the neuroendocrine cells. And also it gave me a wonderful opportunity to interact with the Kuffler group.

Now one of the things that struck me about the Kuffler group that knew really an enormous amount about neurobiology, is the wonderful synthesis that Steve was able to achieve by bringing together the anatomy of the nervous system with the biochemistry of the nervous system, with the physiology of the nervous system. But one thing that really intimidated them, if I can use that word, was behavior. They thought that, number one, they didn't understand behavior, number two, they thought that brain science was not ready to tackle problems related to behavior.

And I in my naiveté thought what's so scary about behavior? If you can sit with a patient with schizophrenia, you going to be worried about the behavior of a snail or a fly or a mouse? So I had a somewhat different attitude than they did, really out of naiveté. And one thing that struck me throughout this period is that I made a number of decisions, which in retrospect seemed quite reasonable, based on rather naïve extrapolation of what could work, rather than confident deep knowledge about the subject matter. So I decided I would work on Aplysia because I realized that memory does not reside in single nerve cells, the unique properties of single nerve cells. You needed a neural circuit. So the question is: how do you then go about using Aplysia to do that?