Back

Posterior parietal cortex guides sensorimotor associative learning by linking sensation to distal action

Zhu, Q.; Wang, Y.; Huang, Y.; Yang, X.; Guo, Z. V.

2026-04-20 neuroscience
10.64898/2026.04.19.719425 bioRxiv
Show abstract

Sensorimotor associative learning enables animals to adaptively link sensory cues with motor actions, a process that is critical for survival and everyday behavior. While Hebbian mechanisms explain associations formed through temporally overlapping neural activity, a fundamental challenge arises when sensory stimuli and motor responses are separated by a delay, because sensory and motor neurons are rarely coactive. Here, we identify the rostro-lateral posterior parietal cortex (PPC-rl) as a cortical hub that bridges tactile stimuli and temporally delayed licking actions during sensorimotor associative learning. Using cortex-wide calcium imaging with single-cell resolution to track [~]16,000 neurons simultaneously across sensory, motor, and association cortices, we find that PPC-rl uniquely exhibits sustained neural activity during the temporal delay early in learning, a signature that diminishes with expertise. Optogenetic silencing of this activity slows learning without impairing sensorimotor execution in expert mice. Learning strengthens the coupling of population dynamics within and between somatosensory and motor cortices. PPC-rl mediates this process by amplifying a low-dimensional communication subspace that synchronizes co-fluctuations across the somatosensory and motor cortices to facilitate linking. This PPC-rl dependent co-fluctuation dissolves post learning, underscoring PPC-rls role in bridging sensation to distal action. A biologically plausible network indicates that Hebbian plasticity with an eligibility trace gated by reward, PPC-rl persistent activity and PPC-rl dependent sensorimotor subspace communication synergize to support delayed association. Together, our findings uncover a PPC-rl based circuit mechanism that maintains temporal continuity to guide associative learning when sensory and motor events are separated in time.

Matching journals

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