Application of the oddball paradigm for investigating predictive coding in the bat auditory system

Predictive coding is a theoretical framework for explaining how the auditory system efficiently codes the auditory scene by focusing on unexpected deviations. It is hypothesized that predictive coding plays a prominent role in the auditory processing of biosonar echo streams. At least two complementary biological mechanisms contribute to predictive coding: stimulus-specific adaptation (SSA), a reduced response to repetitive stimuli that emerges early in the ascending auditory pathway, and mismatch negativity (MMN), an enhanced response to low-probability stimuli mediated by the descending cortical feedback. Here, we utilized the oddball paradigm, an experimental design that measures the brain’s sensitivity to rare (deviant) stimuli, to detect and measure SSA and MMN in the auditory brainstem response (ABR) of anesthetized free-tailed bats. We predicted that SSA would be detectable in the early ABR peaks, and that MMN, if present, should appear in the slower P0 waves reflecting cortical influences on auditory expectations. Stimuli consisted of paired tones, FM sweeps, and other naturalistic stimuli, and we tested the effects of varying stimulus probabilities from 50/50 to 90/10. Results show that both SSA and MMN were detectable and separable even in early peaks of the ABR waveform.