THE NEURAL SOCIAL CODE

This project maps brain-wide neural activity at a cellular resolution during real-time social interactions in zebrafish, capturing dynamic and distributed activity patterns. We identify a robust neural signature that predicts an upcoming approach toward a conspecific, distinct from the dynamics preceding non-social movements. This signature was found specific and necessary for natural social behavior. Our approach enables the study of the developmental trajectory of social coding and its disruption in models of social impairment. The movie displays raw two-photon recordings from eight selected imaging planes during social interaction with a conspecific.

NEURAL AND TEMPORAL ADAPTATION IN VARYING ENVIRONMENTAL FACTORS

We show that larval zebrafish preserve precise hunting performance across a 10°C temperature range, despite temperature-dependent compression of behavioral timescales. Spatial movement parameters remain stable through coordinated adjustments in tail dynamics, specifically increased tail beat frequency and reduced bout duration. Brain-wide calcium imaging revealed parallel temporal scaling in neural activity, and a simple rate model demonstrated that changes in a single neural parameter, the time constant, can account for the observed compensation. These findings suggest that neural temporal scaling enables behavioral stability under global thermal fluctuations without requiring active regulation.

THE REPERTOIRE OF ACTIONS DURING HUNTING AS A MANIFOLD

Like moves in a board game, each movement of a hunting zebrafish follows rules that shape their strategy. We mapped the fish's movement repertoire and discovered a hidden principle: each movement constrains where the fish can be and how it can face next. Position and heading are coupled, carving a low-dimensional set of future options. By modeling this manifold, we reveal how simple movement rules enable strategic prey pursuit.

MOVEMENT CONTROL USING A DYNAMICAL SYSTEM

We developed a unified dynamical model that reconstructs the full spatio-temporal structure of zebrafish tail movements from a low-dimensional control space. By feeding low-dimensional controls, the model captures the continuous nature of movement space. This sparse control framework allows precise reconstruction of entire tail movement and offers a window into the neural structure of motor commands, linking movement generation to neural dynamics.