Sleep reveals dynamics integrating and segregating movement and stimulus representations in V1

Area of Science:

• Neuroscience
• Computational Neuroscience
• Systems Neuroscience

Background:

• The primary visual cortex (V1) processes complex visual information, including movement and stimuli.
• Understanding how V1 organizes these representations within neuronal populations is crucial.
• Intrinsic circuit dynamics during sleep (non-REM sleep) may provide insights into V1 organization.

Purpose of the Study:

• To investigate the relationship between movement/stimulus representations and intrinsic circuit dynamics in mouse V1.
• To explore how V1 neuronal activity is structured by low-dimensional dynamics during non-REM sleep.

Main Methods:

• Extracellular recordings were performed in deep layers of mouse V1.
• Neuronal activity was analyzed during non-REM sleep to identify intrinsic dynamics.
• The distribution of movement and stimulus representations relative to these dynamics was examined.

Main Results:

• Non-REM sleep revealed distinct dynamical subspaces: 'on-manifold', 'off-manifold', and 'unstructured'.
• Both movement and stimulus representations were found on-manifold, interacting additively.
• Stimulus representations were also found off-manifold, where they were segregated from movement activity.
• Off-manifold coding involved population-sparse activity in specific neurons, linking dimensionality to sparse coding.

Conclusions:

• Intrinsic dynamics in V1 constrain neuronal activity, creating a structured substrate.
• This structure balances the integration of movement and stimulus information on-manifold.
• Segregation of stimulus representations off-manifold minimizes interference from movement-related activity.