Realigning representational drift in mouse visual cortex by flexible brain-machine interfaces

Abstract The ability to stably decode brain activity is crucial for brain-machine interfaces (BMIs), which are often compromised by recording instability due to immune responses and probe drifting. In addition, many brain regions undergo intrinsic dynamics such as “representational drift”, in which neural activities associated with stable sensation and action continually change over time. In this study, we employed tissue-like flexible electrode arrays for recording visual stimulus-dependent single-unit action potentials in the mouse visual cortex. The flexible electrode array enabled us to record action potentials from the same neurons over extended periods under visual stimuli, allowing us to characterize the representational drift during these stimuli. With this approach, we tested hypotheses about the origins and mechanisms of representational drift, tracked latent dynamics transformation, and modeled these dynamics with affine transformation. Our findings enabled the construction of a single, long-term stable, high-performance visual information decoder that accounts for representational drift, potentiating chronically stable flexible BMIs in brain regions experiencing representational drifts.