The Tragicomic Tale of the Chemistry of Memory

McConnell was a psychologist interested in understanding learning. We don't know much about learning, and he decided to use a standard strategy: try to understand learning in a really simple organism. He decided to look at planaria (flatworms).

The first question is, can planaria learn? His lab learned to train planaria to associate a light with a shock (so that they scrunched up when the light went on, before the shock) and they learned to run mazes (that is, a Y: they learned to go left or right).

Why was McConnell doing this? Once he established standard tasks that worms could learn, he hoped to use them (in much the way rats are used) to study learning. The hope was that, since worms are so simple, some basic phenomena of learning would be easier to identify in the worms than in rats.

Along they way, McConnell discovered, almost by accident, that if you fed trained worms to untrained worms, the untrained worms acquired (at least some of) the knowledge of the trained worms. McConnell thought that was really cool, largely because it provides a new way of defining and studying learning.

When McConnell announced the result, the response was almost uniformly hostile. Why? The result went against all established theories of how learning takes place. On the other hand, a lot of labs immediately set out to duplicate McConnell 's results.

Ungar, a biochemist, was one of the people who became interested in McConnell 's work. Ungar, however, was not a psychologist, and so he didn't focus on what McConnell was interested in, that the worms were learning. What he, as a biochemist, was interested in was that there seemed to be a chemical that could transfer the learning. Ungar did experiments suggested by those of McConnell in rats, mice, and goldfish. He trained rats, fed the rat frappes to mice, and the mice acquired what the rats knew. The training he focused on first was fear of the dark. The changes make sense motivated by the desire to extract a chemical rather than the desire to study learning. Once Ungar publishes his results, magically, the criticism of McConnell 's learning model evaporates.

Ungar extracts a chemical from the rats and characterizes chemically, that he calls scotophobin: fear of the dark chemical. It is not accepted, for various reasons. By Ungar's count, far more of the labs that tried to duplicate the work succeeded than failed. However, some of the most reputable scientists (a team at Stanford, a team run by someone who had gotten a Nobel prize) failed to replicate. Those who didn't want to believe the results could use that as adequate reason to dismiss them.

Now, a "paradigm shift" occurs: scientists discover that peptides are ubiquitous and important in the brain. Peptides are really really short proteins. Since they are short, easy to work with, ubiquitous, and medically important, every biochemist working on the brain shifts to them. Ungar's work was difficult and expensive.

The end result: No one works on the chemical transmission of learning any more, after Ungar retires, and there is, today, a consensus that there is no established evidence that chemical transmission of learning is possible.

The choice to take up one avenue of research and drop the other is only sociological and political insofar as it concerns the internal sociology and politics of science. It is not, for example, motivated by a desire of biochemists to maintain control of society. In addition, as we have seen from Kuhn, it is rational, well motivated, by the desire to acquire knowledge: experimentation without a theory to guide it is nowhere near as useful and important as normal experimentation within an established paradigm. Only someone who had a naively empiricist view of science would think otherwise.

-- ShaughanLavine - 21 Apr 2009