Dynamic population activity in the striatum during associative behavior

The basal ganglia are a set of subcortical nuclei that are thought to play important roles in motor control, action selection, goal-directed actions, motivation, and non-declarative learning. The striatum is the main input area to the basal ganglia, receiving diverse excitatory input from nearly every cortical area, thalamus, amygdala, and other subcortical structures. The striatum is an important site of plasticity in the basal ganglia, and also contains a complex local microcircuitry. Together these three elements are thought to interact to generate striatal output signals that modulate the rest of the basal ganglia nuclei. Despite decades of study, both the roles that these computations play in generating behavior, and how these dynamics arise, are not well understood. Here, I describe my work using large-scale in vivo recordings from populations of striatal neurons in mice performing a Pavlovian learning task. In chapter 1, I study the structure of spontaneous spiking activity to show that striatal populations that are involved in specific behavior are more likely to show significant correlated activity. This suggests that these neurons share specific inputs, most likely as a result of plasticity at glutamatergic synapses. In chapter 2, I focus on task-related population dynamics in striatal activity to address how these dynamics may be involved in animals’ ability to time their actions. I use machine learning techniques to make direct comparisons between the striatum and one of its input areas, the orbitofrontal cortex, showing that the striatum out-performs the OFC in telling time. These results suggest that the representation of time is not uniform in the brain, and that the striatum may have a privileged role in time representation. Lastly, in chapter 3, I use optogenetic inhibition to causally test the necessity of corticostriatal input for generating striatal population dynamics during anticipatory behaviors in our task. Here, I find that suppressing inputs reduces firing rates in the striatum, but does not eliminate the striatum’s dynamic properties. These suggest that local network interactions may still play an important role in shaping striatal activity, and that striatal output is driven by a balance of excitation and local microcircuit activity