Task-independent movement and engagement

Not all movements have the same relationship to perception or neural activity. Some movements are required to perform a task, such as initiating a trial or reporting a choice, whereas others occur spontaneously throughout a session and may reflect changes in internal state. Recent work has shown that richly varied movements can explain a large fraction of neural activity during behavior (Musall et al., 2019)(Stringer et al., 2019), and work from our lab has shown that task-independent movement (TIM), defined as the component of body motion not directly explained by task events, is negatively associated with perceptual performance (Yin et al., 2025). However, it remains unclear whether TIM simply accompanies disengagement or whether it also marks changes in how sensory information is encoded in V1.

To address this, I use TIM as a trial-by-trial index of movement state and ask how V1 neural responses differ between low- and high-TIM conditions. Using large-scale Neuropixels recordings, which allow me to measure many neurons simultaneously, in freely moving mice performing a visual discrimination task (Odoemene et al., 2018), I test how well the identity of a visual stimulus can be read out from V1 population activity during trials with low versus high task-independent movement. In this way, I can ask whether TIM tracks changes in sensory encoding that are not apparent from task structure alone, and I can place that question alongside broader efforts to separate sensory, motor, and internal-state contributions to population activity (Kobak et al., 2016). Ultimately, this work aims to establish whether task-independent movements provide a useful window into fluctuations in engagement and sensory processing during goal-directed behavior.