By Joshua A. Gordon
Since I assumed leadership of the National Institute of Mental Health (NIMH) at the National Institutes of Health (NIH), pushing the boundaries of the possible in computational psychiatry has been among my top priorities.
The science of psychiatry is a study in contrasts. On the one hand, advances in genetics, basic neuroscience, neuroimaging, and integrative computational approaches promise to dramatically improve knowledge of and treatments for mental illnesses. On the other hand, progress in the mental health translational pipeline has been frustratingly slow, leaving us with few reliable biomarkers, indefinite and subjective diagnostic categories, and partially effective treatments.
Why the slow progress? A principal factor involves understanding the complexity of the brain and its influence on behavior. It is incredibly challenging to connect knowledge gained at genetic, circuit, systems, and behavioral levels. For example, how do the genes that predispose schizophrenia alter function in the circuits that govern the cognitive processes in patients with the disorder? And how might computational approaches help us understand and represent circuit dysfunction in information processing that is visible through neuroimaging? How can we comprehend the incredible heterogeneity in affected patients — even those who share the same genes?
I believe that SIAM and its members can help us answer some of these questions through the application of computational science. Computational approaches allow us to describe and test the way in which complex, high-level phenomena emerge from smaller-scale interactions. Computational models of neural circuits that account for differences in genetic makeup can put testable hypotheses pertaining to gene alterations’ potential effect on circuit function into explicit mathematical terms. Similarly, computational models of circuit dysfunction can test such dysfunction’s ability to create a progressive, chronic disorder by impacting neural development and plasticity. They can also quantify and explain findings in neural systems’ dysfunctions—captured through neuroimaging—and link them to behavioral manifestations. Finally, computational approaches can take advantage of large data sets, categorizing brain dysfunction in a way that may help predict treatment response and lead to better diagnoses and improved biomarkers.
NIMH’s Computational Psychiatry Program is the primary engine for this type of research. Through this program, we hope to bring mathematics, biology, and behavioral science into the pathology and physiology of mental illness via the use of theoretical modeling and machine learning. Employing these tools to investigate the underpinnings of neural activity allows us to make discoveries that aid in our fundamental understanding of the way in which mental states develop and change over time.
In addition, data-driven approaches can assist in the evaluation and testing of certain drugs, neuromodulations, and cognitive interventions. This is key because not all patients with similar disorders will react identically to the same treatments. There is increasing interest in the potential of artificial intelligence to decide the most appropriate drugs for treating common mental illnesses such as depression, a disorder for which there is a wide range of pharmaceutical options. Personalized medicine for more widespread disorders like depression can increase the likelihood that sufferers will seek help and be better equipped to stick with their treatment plans.
Vast potential for advancement exists in these areas, but it will not be possible without robust participation from applied mathematicians and computational scientists. Therefore, I urge computational psychiatrists to explore collaborations with the mental health research community and pursue funding opportunities through the Computational Psychiatry Program. NIMH is constantly seeking chances to leverage expertise in computational psychiatry to make new discoveries in mental health and improve training of researchers that are fluent in data and modeling approaches; these efforts will advance the understanding and treatment of mental disorders. In addition, NIMH is always interested in hearing from researchers who may wish to serve on our study sections and review proposals. Doing so can help the organization hone its approach to computational neuroscience while simultaneously exposing researchers to new funding prospects.
Joshua A. Gordon, M.D., Ph.D., is the director of the National Institute of Mental Health at the National Institutes of Health.