Detection and localization of conspecifics in ghost knifefish are influenced by the relationship between the spatial organization of receptors and signals

Abstract

The detection and localization of signals relies on arrays of receptors and their spatial organization plays a key role in setting the accuracy of the system. Electrosensory signals in weakly electric ghost knifefish are captured by an array of receptors covering their body. While we know that spatial resolution for small objects, such as prey, is enhanced near the head due to a high receptor density, it is not clear how receptor organization influences the processing of global and diffuse signals from conspecifics. We investigated the detection and localization accuracy for conspecific signals and determined how they are influenced by the organization of receptors. To do so we modeled the signal, its spatial pattern as it reaches the sensory array, and the responses of the heterogeneous population of receptors. Our analysis provides a conservative estimate of the accuracy of detection and localization (specifically azimuth discrimination) of a conspecific signal. We show that beyond 20 cm the conspecific signal is less than a few percent the strength of the baseline self-generated signal. As a result, detection and localization accuracy decreases quickly for more distant sources. Detection accuracy at distances above 40 cm decreases rapidly and detection at the edge of behaviorally observed ranges might require attending to the signal for several seconds. Angular resolution starts to decrease at even shorter distances (30 cm) and distant signals might require behavioral or neural coding mechanisms that have not been considered here. Most importantly, we show that the higher density of receptors rostrally enhances detection accuracy for signal sources in front of the fish, but contributes little to the localization accuracy of these conspecific signals. We discuss parallels with other sensory systems and suggest that our results highlight a general principle. High receptor convergence in systems with spatially diffuse signals contributes to detection capacities, whereas in systems with spatially delineated signals, receptor density is associated with better spatial resolution.