Current Research

Neural Basis of Implicit Learning and Action Strategies

Much of our normal behavior depends on learning how to perform sequences of acts so smoothly that we can carry them out almost without conscious effort. This type of learning is crucial for maximizing cognitive function. We depend on it to free us to think and to react to new events in the environment. Much evidence suggests that the basal ganglia are centrally involved in this type of learning, variously known as procedural, implicit or habit learning. But what neural process lets us transform behaviors into habits? And how do we break habits once they are formed? The goal of our research is to answer these questions.

The basal ganglia function as part of a distributed network -- the frontal cortex and basal ganglia operate in loop circuits that change during the acquisition and performance of habit behavior. Our lab is interested in identifying the neural plasticity that occurs in these cortico-basal ganglia loops during acquisition, and the neural encoding that mediates subsequent performance the of behavioral routines. We use two experimental approaches to focus on these issues.

Ensemble Recording in the Rodent

We record chronically with multiple tetrodes (four-channel electrodes) from ensembles of neurons in the striatum and cortex of rats and mice as they acquire learning tasks and then as they perform learned tasks. We find large-scale and long-lasting changes in the response properties of striatal neurons during learning.

Ensemble Recording in the Primate

We also record from the frontal cortex and striatum of macaque monkeys as they acquire motor tasks and then perform them "automatically" as habit-like routines. We find specific neural modulation patterns in the cortex and striatum as the monkeys shift cognitive strategies during learning. These patterns allow us to track shifts in behavioral state by monitoring the dynamic changes in ensemble activity in the cortico-striatal networks.

Microphysiology in the Mouse

We seek molecular events that mark the stages of procedural learning and seek to discover links between these molecular markers and the physiology. We map early response genes for initial clues to identify relevant circuits, and have cloned novel striatum-enriched genes to allow transgenic approaches to studying habit learning.

Computer Modeling

Our laboratory collaborates Pennsylvania State University, Professor Sebastian Seung of MIT and KTH in Sweden to develop computational models of cortico-basal ganglia loop function based on recordings carried out in our laboratory using chronic multi-electrode techniques in monkeys and rodents.

Back to Top