Bridging the gap from neuroscience to psychiatry using computational methods

Our goal is to identify computational principles underlying learning and decision-making behavior and inform our understanding of mental disorders through the study of variation in their neural realization. We aim to translate our computational findings to develop improved diagnoses and personalized treatments for mental illness.

 

To this end, we bridge computational models that target various levels of analysis, including the algorithms (e.g., reinforcement learning models) and their neural implementations (e.g., biophysical models), as well as models of machine intelligence (e.g., deep convolutional neural networks). We test the models' predictions in our empirical studies with human participants (e.g., neuroimaging), as well as through close collaborations with basic and clinical neuroscience/psychiatry laboratories.

​​

We are affiliated with the Department of Psychiatry, the Center for Theoretical Neuroscience, the Data Science Institute, and the Zuckerman Institute at Columbia. We welcome graduate students from various programs at Columbia, including the Ph.D. Program in Neurobiology and Behavior.  

​