Evolutionary expansion of the corticospinal system is linked to dexterity in Peromyscus mice

Animals have evolved behavioral variation to adapt to distinct environmental features. The expansion of neuron number is a potential neural mechanism underlying this behavioral adaptation. Corticospinal neurons (CSNs) are a classic example: an expansion in the corticospinal system in the primate lineage has been hypothesized to underlie their exceptional dexterity. However, the role of CSN number in behavior has been difficult to assess due to the large evolutionary distance between primates and less dexterous taxa with fewer CSNs. Here, we use deer mice (Peromyscus maniculatus) to overcome this challenge. We compared two closely related subspecies of deer mice: forest mice, which evolved dexterous climbing to support a semi-arboreal lifestyle, and prairie mice, which are less dexterous. We find that forest mice have about two-fold larger corticospinal tracts (CSTs) driven by an increase in CSN number in secondary motor and sensory cortical areas (M2 and S2). Furthermore, in a reach-to-grasp test of dexterity, forest mice display higher success and greater grasping flexibility, using multiple grasp types. In contrast, prairie mice use a stereotyped scooping motion, consistent with the idea that an increase in CSN number supports more dexterous movement. High-throughput neural recordings during this task revealed a difference in the timing of neural activity between forest and prairie mice in M2, but not in primary motor cortex (M1): in forest mice, the peak of activity was shifted towards the time of grasp. Forest mice also outperform their prairie counterparts on an ecologically relevant climbing task, where they spend more time upright crossing a thin rod, move faster, and right themselves more quickly when they fall, suggesting a general difference in motor dexterity not restricted to hand use. To assess whether the increase in CSN number contributes to observed behavioral adaptations, we generated forest-prairie F2 hybrid animals with shuffled genomes, neural features, and behavior. We find that the F2 hybrids with larger CSTs perform better on the rod crossing task, suggesting that expansion of the CS system likely supports the adaptive increase in climbing dexterity in forest mice. Together, our work establishes the forest-prairie deer mouse system as a novel model to investigate the role of neuron number expansion, and CSNs in particular, in dexterous movement.