Research

The brain must keep acquiring new skills throughout life without erasing what it has already learned. This raises a fundamental question: new learning can interfere with established representations, yet stable behavior demands that prior memories be preserved. We study how the motor cortex resolves this trade-off — how existing representations are maintained across weeks of continual learning, and how the structure and order of training determine which memories interfere and which coexist. Building on our finding that the cortex assigns context-specific coding directions to separate and combine motor memories, we use long-term calcium imaging to track the same neural populations as animals learn, retain, and relearn, aiming to uncover the coding principles that make lifelong learning possible. 

Learning is not merely the strengthening of existing signals; it is the formation of entirely new internal representations. When and how do these representations first appear, and how do they change once formed? We ask whether learned representations emerge during active experience or are consolidated offline, and whether their refinement proceeds gradually or through discrete, stepwise transitions. By imaging large neural populations continuously across the full course of learning — before, during, and after the acquisition of new behavior — we seek to capture the moment representations arise and to characterize the dynamics through which they mature and stabilize. 

Identical behaviors can be produced by many different neural configurations, a property known as degeneracy. We propose that this principle underlies two persistent puzzles: why individuals differ in how their brains solve the same task, and why neural representations drift over time even when behavior remains stable. Leveraging large-scale longitudinal imaging across many animals, we examine how degenerate solutions are selected, how they vary from one individual to the next, and how they reorganize while preserving behavioral output. Our goal is a framework linking individual variability between neural and behavior solutions throught the notion of degeneracy.