Spatial coding of conspecifics in the electrosensory system: Early segregation of information streams.

Abstract

Localizing the source of a signal requires sophisticated neural mechanisms, and we are still uncovering the coding principles that support accurate spatial processing. Weakly electric fish can detect and localize distant conspecifics, but the way this spatial information is encoded is unclear. Here, we investigate the spatial representation of conspecific signals in the hindbrain to determine how the properties of the heterogeneous population of pyramidal cells affect the spatial coding accuracy of conspecific signals. We hypothesize that specific subsets of cells provide more accurate spatial information about conspecific location. We stimulated the fish with an artificial signal that replicates both the spatial and temporal structure of conspecific signals. We recorded from cells with various receptive field positions covering the entire body surface and analyzed the spike train with spike-train distance metrics to determine how accurately the location of the stimulus is encoded. We found that some pyramidal cells (such as ON-type and those within the deep layer) encode the spatial information more accurately, while other subgroups (OFF-type and superficial layer) provide less accurate information. Our results help us understand how conspecific location is represented across the heterogeneous population of cells and suggest that segregation of the spatial information stream begins early in the sensory pathway.