Listening in the dark: Behavioral and neural mechanisms of acoustic recognition in a singing fish

Abstract

Vocalizing species of teleost fish have a simple repertoire of species-typical signals central to their social and reproductive behavior. Male midshipman fish (‘‘the California singing fish’’) acoustically court females using ‘‘hums’’ that are long duration (secs 1h) and spectrally simple. Individual hums exhibit an essentially flat envelope consisting of a low fundamental frequency (90–100 Hz) and several harmonics. By contrast, naturally occurring acoustic beats show envelope modulations: the hums of neighboring males overlap to produce a beat with a difference frequency (0–8 Hz) in which amplitude and phase modulations result from the continual shifting of the two fundamental frequencies. Behavioral playback (phonotaxis) studies show that individual midshipman can use multiple signal parameters including hum duration, gap duration between successive hums, fundamental frequency, and intensity to distinguish hums from each other, while the effectiveness of beats as playback signals depends upon both beat modulation rates and depth of modulation. Neurophysiological evidence shows that auditory neurons in the midshipmans VIIIth nerve and midbrain encode the physical attributes of these signals mainly using spike rate and spike synchronization, mechanisms that we propose contribute to the neural templates adapted for both hum and beat recognition. [Research supported by NIH.]