Motivated Learning & Memory Laboratory

MLML Research

Motivated Learning & Memory Laboratory

Director: Dan Dillon, PhD

The Motivated Learning & Memory Lab uses methods from cognitive and affective neuroscience to provide new insight into psychiatric illness. Currently, we are focused on understanding how depression affects the psychological processes and neural systems that support long-term memory and reinforcement learning. Our ultimate goal is to generate new targets for prevention and treatment interventions.

Reward Dysfunction as a Cause of Memory Deficits in Depression

Poor memory is a common and troubling issue in depression. Unfortunately, the relevant neural mechanisms are not well-understood. To address this issue, we are pursuing a hypothesis that links memory problems in depression to abnormal responses in brain reward circuits. We are currently using functional magnetic resonance imaging (fMRI) and computational modeling to test the prediction that firing in brain reward circuits triggers long-term memory formation in healthy individuals. Next, we aim to show that responses in reward networks are blunted in depressed adults, which compromises memory formation and ultimately leads to memory failures. By integrating neuroscience research on reward processing and memory, this work promises to provide new insights into the cognitive correlates of depression.

Examining Effects of Depression on Reinforcement Learning

Unipolar depression is a heterogeneous condition that involves low motivation, poor problem-solving, diminished interest in goals, and rumination on negative thoughts. Addressing this diverse set of problems in piecemeal fashion is a daunting task that has consumed researchers’ time for several decades. To facilitate rapid progress, we are approaching depression through the lens of reinforcement learning. Reinforcement learning specifies algorithms that an agent can use to achieve desired goals. For example, researchers in this field have detailed the computations necessary to assign value to different stimuli, choose among them, and then remember the consequences of each choice in order to inform future decisions.

This work has supported fundamental advances in artificial intelligence and robotics, and neuroscience research has demonstrated that reinforcement learning is supported by cortico-striatal circuits that receive projections from the dopaminergic midbrain. Based on work conducted at McLean Hospital and elsewhere, we hypothesize that unipolar depression involves structural and functional changes in these neural networks. Thus, we are using fMRI and computational modeling to test the hypothesis that reinforcement learning is impaired in depressed adults. This work offers an integrated understanding of the diverse problems that characterize unipolar depression, which may ultimately lead to new avenues for prevention and treatment.

 

Contact our This email address is being protected from spambots. You need JavaScript enabled to view it. with any web-related questions or feedback
Website credit: Michael Attisha © 2012