Co-registration across arrays and sessions

The problem of aligning recordings from multiple implants — or from the same implant across days — into a single, consistent coordinate frame. It has an anatomical side (where in the brain each site sits, relative to other arrays and to a reference atlas) and a functional side (matching channels or latent dimensions across sessions despite drift, unit turnover, and electrode movement).

Co-registration is the glue that makes distributed and whole-cortex recording more than a pile of disconnected data: only after alignment can one study signal flow between areas or pool data across arrays and days. Functional alignment methods — matching low-dimensional neural manifolds across sessions — are closely related to the latent-space stabilisation used by drift-robust decoders, and are an active research area rather than routine practice.