Stable task information from an unstable neural population.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Systems Neuroscience

Background:

• Neural activity in the sensorimotor cortex reconfigures over time during learned tasks, challenging the idea of stable neural engrams underlying stable behavior.
• The systematic nature of this neural drift and its implications for information processing in downstream circuits remain unclear.

Purpose of the Study:

• To investigate whether neural drift in the posterior parietal cortex is systematic and facilitates consistent information readout.
• To determine if plasticity mechanisms can compensate for neural drift to maintain behavioral decoding over time.

Main Methods:

• Analysis of long-term calcium imaging recordings in mice (Mus musculus).
• Investigated neural activity representing position and movement during learned tasks.
• Calculated the plasticity required to compensate for neural drift.

Main Results:

• Neural drift in the posterior parietal cortex is systematically constrained, supporting a linear readout of behavioral variables.
• A component of drift degrades fixed readouts, indicating it does not exist in a null coding space.
• The required plasticity to compensate for drift is physiologically achievable.

Conclusions:

• Systematic neural drift can be leveraged for consistent behavioral decoding.
• Biologically plausible local learning rules can compensate for neural drift, maintaining accurate decoding over extended periods.
• Findings challenge classical theories and offer insights into neural plasticity and information processing in the brain.