Posing Hypotheses Responsibly in Psychiatry

It is easy to say that the analysis by Kendler and Schaffner of the status of the dopamine hypothesis of schizophrenia (DHS) is, at the very least, a scholarly read. It includes an exhaustive review of the DHS literature accompanied by a demanding critique. The authors' bar for hypothesis verification is high, and their conclusion is negative—that scientific support is insufficient to retain the hypothesis as such. They proceed to evaluate the reasons they see for both (1) the extensive testing of the hypothesis extending over decades and (2) the failure of the field to falsify in a timely fashion. In a very interesting section, they review philosophical positions about how and in what context scientific advances occur in medicine. Their approach is an exemplary model for evaluating the significance of hypotheses in the field of brain science. I am a scientist who worked directly in this field during all the years of DHS testing that Kendler and Schaffner review. In the face of this critique, I wonder to myself why we were so resistant to "facing facts," so blind? Of course, it deserves being said that for many years, the DHS represented every shred of neurobiological knowledge that we possessed about schizophrenia; it organized our approach to the illness. Thus, giving up on the DHS involved admitting both our broad lack of understanding about the disease and even our poor factual basis for antipsychotic treatments. There has never been any doubt that dopamine receptor antagonism provides the optimal target for antipsychotic efficacy (Carlsson and Lindquist 1963), but even now we do not know the key molecular or cellular determinants of psychosis that would enable us to develop a rational disease target. Nonetheless, the conclusion of these authors regarding the lack of evidence that the DHS accounts for schizophrenia pathophysiology (albeit, done in retrospect) is clear, encompassing, and logical, and challenges the lack of rigor of our scientific approach during these years.

We can ask ourselves why it happened that such a hypothetical set of postulates was articulated for so many years in the face of so little factual support. One perspective on accounting for this outcome, one that the authors refer to obliquely, is that this was never exactly a hypothesis—it was more like an idea. Because we had not previously had such specific hypotheses to explain psychiatric illnesses, our criteria for developing one were weak and fluid; perhaps our excitement about the idea led us to label it 'hypothesis' much too quickly. Today, we might call this a speculation or a model, until confirmatory evidence collected. Also, because in the 1950s fundamental neuroscience was a field where knowledge was sparse and the complexity of neural function was barely recognized, there was very little factual background on which to play out such a hypothesis. All of this happened at a time in medical research overall when fundamental knowledge and its translation to other medical diseases was productive and fascinating. How could we say that we had no idea at all about the disease pathophysiology of something as florid as schizophrenia? Having the DHS supplied a need to know; in the face of extraordinary advances in other areas of medicine, doctors had a bit of real knowledge to share with their patients and families. Moreover, even though the evidence of non-support was apparent, no substitute explanation was available, no "new paradigm" took its place. In fact, as the years passed, schizophrenia scientists themselves largely sidelined the DHS for practical purposes, but without ever specifying any reason or articulating the kind of hypothesis rejection that Kendler and Schaffner have provided in their article.

This line of speculation makes one wonder about engaging in translational research into pathophysiology of complex brain diseases without fundamental neuroscience knowledge. Could we say that our intentions were premature? The context of brain research was productive at the time; hopes were fueled by successes in other brain diseases, like the discovery of dopamine depletion in Parkinson's disease (Hornykiewicz 1966), which itself indicated a rational treatment. A perspective from current neuroscience discovery might suggest that posing hypotheses...