Your message and the articles by Moses and Martin, and Dorsey et al., on the status of neuroscience research in the December 2006 issue of Annals were both interesting and timely.1–3 You have identified a serious mismatch between the expansion of funding for neuroscience research and the generation of new effective treatments for diseases of the nervous system. This mismatch suggests that the National Institutes of Health and the neuroscience research community have failed either to direct research to clinical problems or to translate disease-relevant discoveries into new treatments. The public has backed its faith in medical research with huge sums of money. To a certain degree, we, as stewards of this massive medical research program, have not kept our end of the bargain. In part, this is because of the complexity of the problems we are addressing. On the other hand, I think it is reasonable to question the degree to which the research dollar is actually being spent in the way the public intended. Despite significant positive accomplishments, a large proportion of NIH neuroscience grants support interesting and high-quality research that has no obvious clinical value. The fact is that study sections often ignore the disease relevance of the research proposals they review. As a member of two study sections, I speak from personal experience. The large pharmaceutical companies, although totally invested in the relevance of their research dollar, are handcuffed by attention to market factors rather than disease mechanism. I have spent hours with major pharmaceutical and biotechnology companies trying to explain to them why clinical research must be more than therapeutic trials. National Institutes of Health and the pharmaceutical and biotechnology industry need to invest in the study of disease mechanism in patients. Although academic and commercial entities often use animal models of disease to test their candidate therapies, such models, particularly of nervous system disorders, can be notoriously misleading. I need only mention the failure of animal models of pain, stroke, or brain injury to accurately predict the therapeutic value of new pharmaceutical interventions. This is a problem that calls out for more mechanism-oriented research in patients.

The explosion of new, noninvasive methods to study patients in the early stages of their diseases, for example, using genetic analysis and functional brain imaging, presents an unparalleled opportunity to study human disease mechanisms. Why has progress been so slow? Part of the problem is that clinicians are typically too busy (and too poorly compensated) in their practice to spend enough time with patients. Without detailed examination of patients, it is difficult to make the critical observations that can lead to an understanding of the underlying neurobiology of disease. Clearly, the financial pressures on practicing clinical neurologists are a barrier to a better understanding of disease, and thus to development of new, more effective treatments. One approach to support the study of disease mechanism would be to establish neuroscience laboratories specifically designed for human research and funded so that the clinician-scientist can study patients in detail; test mechanistic hypotheses, for example, using functional imaging; and then, ideally, take this mechanistic information and hypothesis back to animals where more invasive cellular and molecular approaches can be brought to bear. Note that the flow of hypothesis and experiment in this case goes in both directions; there is continuous and dynamic feedback between patient observation and laboratory experiment. But the key is that this research is driven primarily by patient observation. Nowhere is this approach more crucial than for diseases of the nervous system, especially degenerative and functional diseases. The brain is the most unique human organ, and although some diseases can be studied perfectly well in animals (eg, transmissible diseases, toxicology, and so forth), the biggest challenge for most animal work on central nervous system diseases such as epilepsy, pain, movement, and cognitive/behavioral disorders is to show that a particular animal model has any relation to the human disease. The optimal solution for this issue is to study patients with all available tools. Functional imaging will be a major part of that effort, but careful clinical examination of patients is the irreplaceable centerpiece. Unfortunately, there really is no financial mechanism to support this vital class of research. It is typically too preliminary and anecdotal to satisfy a study section composed of laboratory-based scientists, especially if they are unfamiliar with the disease in question. So an RO-1 grant is, in my experience, rarely, if ever, awarded for this type of work. Furthermore, because this early high-risk work typically has no obvious relation to product development, commercial entities are reluctant to support it. University hospitals must be run as businesses, and the public does not understand what is unique about this clinically driven experimental approach. Somehow, we must find a way to get this type of work funded. Without it, true bidirectional translational work will continue to swim in a muddy pond, and the mismatch between spending and effective new treatments will continue or worsen.


  1. Top of page

Howard Fields MD, PhD*, * Department of Neurology and Physiology, University of California, San Francisco, San Francisco, CA.