Back

Cortex-Wide Substrates for Body Schema and Action Awareness

Severson, K. S.; Lu, J.; Xiao, W.; Jiang, H.; Lou, T.; Caplan, K. A.; Levy, M. S.; Li, T.; Dunn, T. W.; Wang, F.

2025-11-12 neuroscience
10.1101/2025.09.23.678085 bioRxiv
Show abstract

The concept of the body schema originated from studies of patients with cortical lesions who were unable to localize their body parts in space or perceive passive movements (Head & Holmes, 1911). Current consensus defines the body schema as the brains internal representation of the body configuration in space used for both embodied perception and voluntary action (de Vignemont, 2010; Berlucchi & Aglioti, 2010; Vallar et al., 2025). Despite extensive research, how the cortex efficiently encodes the vast repertoire of body postures to support natural behavior remains poorly understood. Here, we reveal the representational format underlying the body schema by combining large-scale electrophysiological recordings with full-body 3D tracking and joint angle computation in freely behaving mice. Across cortical regions, neurons encoded diverse postural features in distinct reference frames: posterior parietal cortex (PPC) populations preferentially encoded body midline-referenced features, while sensorimotor cortices utilized a gravity-referenced frame. Strikingly, during performance of different stereotyped actions, distinct cortical ensembles were selectively activated. These action ensembles fired in phase-locked sequences that tiled the full action cycle, providing a neural substrate for action awareness. A network model recapitulated these dynamics and identified low-dimensional action subspaces within population activity. Together, these findings reveal that the cortex organizes the body schema around actions, compressing high-dimensional kinematics into an efficient, body-midline-and gravity-anchored, action-aware neural code.

Matching journals

The top 2 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.