Behavioral and neurophysiological signatures of the modulation filterbank in an animal model

Abstract

Fluctuations in the temporal envelope of sound, called amplitude modulation (AM), are an important information-bearing feature of many complex communication signals including speech. Human listeners show diminished sensitivity to AM signals in the presence of competing modulations of similar frequency, known as modulation masking. Modulation masking is not explainable by classic power-spectrum models, but instead suggests a “modulation filterbank” processing strategy that separates concurrent sounds (e.g., AM signals from noise) that have different modulation frequencies. The modulation filterbank is an exciting theoretical development because in addition to explaining modulation masking, the model predicts differences in speech perception across noise environments with different envelope statistics. However, the physiological underpinnings of the modulation filterbank remain uncertain due to limited nonhuman animal models. New behavioral and neurophysiological studies of the modulation filterbank are presented in budgerigars, a parakeet species with human-like behavioral sensitivity to many simple and complex sounds. Behavioral modulation-masking results show compelling evidence of the modulation filterbank in this animal model. Neural recordings from the inferior colliculus (midbrain level) further suggest that rate-based AM sensitivity can explain behavioral modulation-masking results. This new animal model of the modulation filterbank can be used to explore hearing mechanisms in real-world noisy listening environments.