Mechanisms of memory formation in cortical networks during learning of goal-directed behaviors

A central hypothesis in neuroscience is that changes in connectivity patterns between neurons support learning and memory formation. Most methods for examining connectivity between individual neurons rely on ex vivo experiments (e.g., in brain slices).

However, in vivo measurements are required to study how neurons causally influence each other's activity ('causal connectivity') in the living brain, and how these causal interactions change over time.

Thus, while brain networks are among the most studied biological networks, the cellular-level patterns and dynamics of causal connectivity in vivo remain unknown.

Here, I propose to study how causal connectivity between individual neurons and across entire brain areas changes over time during learning of memory-guided behaviors.

To this end, we will use novel causal optical methods to longitudinally map causal connectivity and neural activity at cell resolution in vivo, focusing on the motor cortex and related areas.

We will combine these methods with a novel goal-directed behavior in mice that does not require pretraining, which will serve as a baseline to study learning mechanisms of more complex behaviors that rely on short-term memory.

Specifically, we will map changes in causal connectivity during learning within the motor cortex (Aim 1) and across cortical areas (Aim 2) and relate it to the computational functions of the network.

We will also perturb neurons based on their connectivity and coding properties to identify changes in network mechanisms for short-term memory and action selection at various learning stages (Aim 3).

Finally, we will work towards identifying constraints on memory formation via optogenetic induction of artificial connectivity patterns.

Taken together, this research will enable for the first time to causally study dynamics in network interactions across time on different spatial scales, and to test fundamental mechanisms of memory formation and representation in cortical networks.