Decision dynamics in the jamming avoidance response of weakly electric fish: impact of conspecific motion

Wave-type weakly electric fish sense in dark environments using a quasi-sinusoidal electric organ discharge (EOD). When two fish have similar EOD frequency (EODf), each fish changes its frequency to avoid signal interference. In this 'jamming avoidance response' (JAR), fish must decide whether to increase or decrease EODf using amplitude and phase information from the combined EODs. The JAR has typically been examined in stationary fish, but during natural interactions, swimming produces noisy amplitude modulations (AMs) that could impact JAR decisions. To test this, we delivered artificial signals to Eigenmannia virescens mimicking a dynamically approaching conspecific (2-20 Hz AM noise) with lower EODf (i.e. the correct JAR comprises an EODf increase). Noise intensity did not significantly affect maximum JAR magnitude (p = 0.784) nor the time to a correct JAR (p = 0.883). However, noise improved accuracy when signal amplitude was low (p = 0.037) but reduced accuracy when amplitude was high (p = 0.008), suggesting that motion-generated AMs can benefit or impair JAR encoding depending on signal saliency. We also observed that short-latency decisions were biased slightly towards an incorrect JAR despite the noise benefit, and that the classic JAR algorithm combined with a biased random-walk model for low-amplitude conditions can explain the JAR dynamics produced by motion-generated AMs.