CEREBELLAR POPULATION DYNAMICS DURING LOCOMOTION ACROSS VARIABLE WALKING CONDITIONS

Locomotion is a fundamental behavior in both animals and humans, requiring continuous adaptation to environmental features such as surface structure and unexpected obstacles. This form of motor control depends on several brain regions, notably the cerebellum that processes sensorimotor information to adapt movement patterns. We are interested in the contributions of the cerebellum to locomotor adaptation, including its role in learning rapid, online adjustments as well as in their long-term consolidation into motor memory.
To investigate these processes, we developed a challenging locomotor task in which mice encounter varying walking conditions, including traversing rungs, overcoming obstacles, and adapting to changes in speed. This task, LocoReach, consists of a motorized treadmill made of rungs with two movable rungs that can create obstacles. Successful performance requires dynamic adjustment of paw trajectories and interlimb coordination in response to wheel speed, rung spacing, and obstacle location. During task execution, behavior was video-recorded while cerebellar activity in lobule simplex was measured using chronically implanted Neuropixels probes. Machine-learning–based approaches were used to extract paw trajectories and quantify distinct obstacle crossing strategies. Over several days of learning, mice became increasingly proficient at LocoReach, so that they made fewer, longer strides with faster swings and fewer missteps.
In parallel, cell-type classification enabled us to dissect how different cerebellar neuron populations are modulated at specific points of the step cycle and across walking conditions/patterns.
Together, this work aims to elucidate the role of the cerebellum in the neural mechanisms underlying paw placement during adaptive locomotion.