![](/wp-content/uploads/2021/03/abstractheadergraphic2014.gif)
Meeting Abstract
P2.27 Sunday, Jan. 5 15:30 Frequency tracking and spatial localization of unconstrained weakly electric fish reveal complex social interactions in natural populations JAYAKUMAR, R P*; MADHAV, M S; STAMPER, S A; FORTUNE, E S; COWAN, N J; Johns Hopkins University; Johns Hopkins University; Johns Hopkins University; New Jersey Institute of Technology; Johns Hopkins University rperurj1@jhu.edu
Weakly electric fish emit an electric organ discharge (EOD) that is used both for social communication, and the localization and characterization of nearby objects. Since the EOD of these fish can be modeled as an oscillating electric dipole, their field amplitudes in the water are extremely sensitive to distance and orientation. The relative movements of individual electric fish have dramatic impacts on social signalling, but also make it more difficult to localize fish using their electrical signatures on a set of electrodes. In this study we sought to create a technique for examining the relationship between social signals associated with both changes in an individual fish’s EOD and those associated with relative movements. We developed a method to localize and track the spatial position and orientation (pose), and EOD frequency of multiple wave-type electric fish using spatially distributed voltage measurements. Each measurement at an electrode records a combination of EODs from multiple individuals. The amplitude of each EOD varies in relation each individual’s pose relative to the electrode. Using this data, we designed a Bayesian filter to estimate the state of each individual: its position, orientation, velocity, and EOD signal parameters. The approach was validated against both simulations, and laboratory trials where multiple individuals of both Eigenmannia and Apteronotus were recorded using both video and a rectangular grid of 8 electrodes. In addition, we made recordings of groups of Eigenmannia along the Napo River in Ecuador using a 1.5 meter square grid with 16 electrodes, and our next step is to apply this technique in the analyses of this field data.