An Internal Model of Sensorimotor Context in Freely Swimming Electric Fish

Abstract

Nervous systems are hypothesized to learn and store internal models that predict the sensory consequences of motor actions. However, little is known about the neural mechanisms for generating accurate predictions under real-world conditions in which the sensory consequences of action depend on environmental context. Using novel methods for underwater neural recording in freely swimming electric fish, we demonstrate that complex movement-related input to the active electrosensory system is effectively cancelled, despite being highly-dependent on the nearby environment. Computational modeling and closed-loop electrophysiological experiments indicate that the cerebellum-like circuitry of the electrosensory lobe generates context-specific predictions of self-generated input by combining motor signals with electrosensory feedback. These results provide mechanistic insight into sophisticated internal models supporting natural behavior in freely moving animals. One-Sentence SummaryUnderwater recordings in electric fish reveal neural mechanisms for predicting the sensory consequences of behavior under natural conditions in freely moving animals.