Locomotion and visual coding

We often take for granted how effortlessly the visual system can process sensory information while the body is in motion. However, most of what we know about neurons in the primary visual cortex (V1), the part of the brain that helps process visual information, comes from experiments in which movement is constrained or absent altogether. Although this work established core properties of visual responses, it remains poorly understood how those responses are modulated as animals actively navigate their environment (Niell & Stryker, 2010)(Lee et al., 2014).

A key discovery from head-fixed preparations is that locomotion substantially changes how V1 encodes visual stimuli. For example, neurons in V1 have stronger responses to visual stimuli during running than during quiescence, and running decreases both response variability and noise correlations across the population (Niell & Stryker, 2010)(Dadarlat & Stryker, 2017)(Bennett et al., 2013). In addition, other work has shown that spontaneous movements can shape activity in visual cortex even outside of a task, suggesting that movement-related signals are represented beyond locomotion alone (Stringer et al., 2019). However, head-fixation substantially limits the range of movements animals can make during natural vision, and it is thus still unclear how locomotion and other movement signals interact with visual encoding when movement, sensation, and behavior are naturally coupled (Odoemene et al., 2018).

To address this gap, I record the activity of many individual neurons in V1 at once using Neuropixels probes, which are high-density electrodes, in freely moving mice performing a visual decision-making task (Odoemene et al., 2018). In this task, mice view simple visual stimuli and make behavioral choices, allowing me to study sensory processing during ongoing behavior rather than under highly constrained conditions. By tracking locomotion continuously throughout each session, I ask how movement state reshapes sensory responses across the population: which neurons are modulated, how strongly, and whether population-level response structure changes with movement. Ultimately, this work aims to establish how locomotion-related signals influence visual coding under naturalistic conditions, and what that implies for how the brain integrates internal state with sensory input.