Hierarchical, Intermingled and Segmented: The Organization of Neural Activity in Mouse Sensorimotor Cortex

Cortical control of movement is a distributed computation spanning multiple, densely interconnected regions. Although we have rich anatomical atlases and a coarse understanding of how function maps to these areas and subareas, we lack a detailed account of how behaviorally relevant activity is organized across the cortical sheet. Progress requires two key factors: (i) a challenging sensorimotor task with many conditions to engage sensorimotor circuits, and (ii) large-scale, single-neuron-resolution sampling across the full set of involved areas. In this dissertation, I address these gaps with two studies that pair a motorically demanding behavior with dense sampling of mouse sensorimotor cortex. First, we trained mice to perform a challenging reach-to-grasp task while recording neural activity in primary motor cortex (M1) and forelimb somatosensory cortex (S1-fl). Detailed kinematic information was similarly present in both regions; however, movement-related signals emerged earlier and were more persistent in M1, whereas S1-fl responses were more time-varying. These findings support a distributed control framework and are consistent with somatosensory areas receiving and transforming efference-copy-like information. Second, we expanded the task to fifteen target locations and performed cortex-wide, single-neuron sampling across much of sensorimotor cortex. Using statistical models on tuning features, we identified interleaved subpopulations with distinct response profiles. This large and rich dataset reveals a parallel-processing architecture where local specializations sit inside a distributed, flexibly routed network.