Compass cells in the brain of an insect are sensitive to novel events in the visual world

Abstract

The central complex of the insect brain comprises a group of neuropils involved in spatial orientation and memory. In fruit flies it mediates place learning based on visual landmarks and houses neurons that encode the orientation for goal-directed locomotion, based on landmarks and self-motion cues for angular path-integration. In desert locusts, the central complex holds a compass-like representation of head directions, based on the polarization pattern of skylight. Through intracellular recordings from immobilized locusts, we investigated whether sky compass neurons of the central complex also represent the position or any salient feature of possible landmarks, in analogy to the observations in flies. Neurons showed strongest responses to the novel appearance of a small moving square, but we found no evidence for a topographic representation of object positions. Responses to an individual square were independent of direction of motion and trajectory, but showed rapid adaptation to successive stimulation, unaffected by changing the direction of motion. Responses reappeared, however, if the moving object changed its trajectory or if it suddenly reversed moving direction against the movement of similar objects that make up a coherent background-flow as induced by ego-motion. Response amplitudes co-varied with the precedent state of dynamic background activity, a phenomenon that has been related to attention- dependent saliency coding in neurons of the mammalian primary visual cortex. The data show that neurons of the central complex of the locust brain are visually bimodal, signaling sky compass direction and the novelty character of moving objects. These response properties might serve to attune compass-aided locomotor control to unexpected events in the environment. The difference to data obtained in fruit flies might relate to differences in the lifestyle of landmark learners (fly) and compass navigators (locust), point to the existence of parallel networks for the two orientation strategies, or reflect differences in experimental conditions.