FEATURE-DETECTING AUDITORY NEURONS IN THE BRAIN OF A SOUND-PRODUCING FISH

The mormyrid fish Pollimyrus adspersus has auditory specializations fo r sound pressure detection and uses acoustic displays in its natural s ocial behavior. In this paper it is shown that auditory neurons in the mesencephalon (torus semicircularis) are activated selectively by tem poral features of complex sounds. Single neurons were recorded while p resenting sounds to fish underwater. The stimuli were acoustic click t rains, 400 ms in duration, and were synthesized with differing inter-c lick-intervals (ICIs). The natural sounds of this species are composed similarly and the range of ICIs synthesized overlapped with the natur al range (5-40 ms). One-third of the neurons studied were strongly sel ective for a narrow range of ICIs, increasing spike rate by ten fold o r more at the best ICI compared to the minimum response observed. The best ICI for interval selective neurons remained stable when the sound pressure of the stimulus was changed. Neurons that were selective gav e phasic responses to tone bursts, and most had non-monotonic rate lev el functions. The origin of interval selectivity is discussed and a ti me-based computational mechanism is proposed.