TREBLE: a generalizable framework for high-throughput behavioral analysis

Abstract

Uncovering relationships between neural activity and behavior represents a critical challenge, one that would benefit from facile tools that can capture complex structures within large datasets. Here we demonstrate a generalizable strategy for capturing such structures across diverse behaviors: Time-REsolved BehavioraL Embedding (TREBLE). Using data from synthetic trajectories, adult and larval Drosophila, and mice we show how TREBLE captures both continuous and discrete behavioral dynamics, can uncover variation across individuals, detect the effects of optogenetic perturbation in unbiased fashion, and reveal structure in pose estimation data. By applying TREBLE to moving mice, and medial entorhinal cortex (MEC) recordings, we show that nearly all MEC neurons encode information relevant to specific movement patterns, expanding our understanding of how navigation is related to the execution of locomotion. Thus, TREBLE provides a flexible framework for describing the structure of complex behaviors and their relationships to neural activity.