Hippocampal-prefrontal mechanisms of learning to learn

Project summary The mammalian brain is remarkably ecient at rapidly learning from experience. We can quickly form generalizations about our environment based on previously acquired knowledge, and we are able to establish and maintain detailed memories after just a single exposure to a stimulus. This rapid, experience-dependent learning is critically dependent

on schemas, or learning sets, which are learned cognitive structures that organize knowledge and shape how incoming sensory information is interpreted. A related phenomenon is learning to learn, which is observed when experience with a task accelerates the learning of novel tasks or problems with a similar structure. Decades of experimental work have

emphasized the essential role of the hippocampus and prefrontal cortex|and their interactions|in associative learning and memory consolidation. However, the neural mechanisms underlying learning to learn remain largely unknown. Learning eciently from prior experience is crucial for successful everyday life, and this skill is impaired in patients

with a number of neurological conditions, such as Alzheimer's disease, attention de cit disorder, temporal lobe epilepsy, depression, and schizophrenia, which collectively a ect tens of millions of Americans. A deeper understanding of the brain mechanisms underlying learning to learn can provide important insights into the neurological basis of learning-

related de cits in these patients, and can be used to develop e ective interventions targeted at memory loss and learning disruption. To investigate the neural mechanisms of learning to learn, we will record neuronal population activity in hippocampus and prefrontal cortex as animals acquire a learning set for a visual association task. In Aim 1, we will

determine how neuronal activity patterns within hippocampus and prefrontal transform during learning set formation to support rapid learning. In Aim 2, we will characterize the changes in coordinated activity between hippocampus and prefrontal cortex that re ect learning set formation. We hypothesize that learning set acquisition is accompanied by a

progressive emergence of stable, abstract neural representations of task-related information that can be re-used from problem to problem, thereby constraining the neural exploration space during learning and leading to faster learning through experience. In addition, we expect to observe a gradual shift from hippocampus-dependent to cortex-dependent

representations throughout the process of learning to learn.