We are interested in how we learn and make decisions in a dynamically changing environment. In such conditions, how does one decide to explore vs. exploit, learn from positive and negative feedback, demonstrate cognitive flexibility to update their beliefs, collaborate, or compete to maximize their payoffs. Understanding inter-individual differences in behavior and the underlying mechanisms may provide us insights into the impact of neuropsychiatric disorders, such as anxiety, stress, schizophrenia, and depression, as well as factors like momentary affect and sleep. 

We utilize computational models as a framework for understanding latent cognitive processes underlying learning and decision making. For this we use reinforcement learning models, drift diffusion models, and other such computational process models. By using these models we parameterize individual differences in positive/negative feedback based learning, sensitivity to feedback, motivation and urgency related to decision making etc. These parameters help us monitor the subtle changes in learning and decision making in mental health conditions. Overall, we aim to identify potential biomarkers that can serve as early indicators of anxiety disorders and depression in vulnerable populations.

We are also interested in understanding the neural mechanisms that support computations associated with learning and decision-making. For this, we record the neural activity from brain regions associated with these behaviors. The initial focus is on brain regions like the anterior cingulate cortex and the dorsolateral and dorsomedial prefrontal cortex. We also build our own flexible probes for neural recording; for more on this, check out the startup Eywa Neuro. These are ultrathin film polyimide electrodes that are soft, cause less tissue damage, and can be customized for the brain region of interest. We are already testing these probes in mice, rats, and, in the near future, nonhuman primates. In an ongoing project, we are using these electrodes to understand epileptic seizures. 

After determining underlying mechanisms, to establish causal underpinning, we use various methods ranging from neuropharmacological approaches, electrical stimulation, and we also intend to do chemogenetic manipulations. For example, to investigate the effects of pharmacological stress, we administer hydrocortisone, which changes cortisol levels, and/or yohimbine, which alters norepinephrine levels. We also use the neural probes to directly stimulate brain regions of interest. 

DBT Wellcome Trust India Alliance