Physiological studies of timbre encoding in the auditory midbrain

Abstract

Perceived timbre, especially brightness, is influenced by the spectral peaks of sounds. We hypothesize that spectral peaks are reliably encoded in the inferior colliculus (IC) due to capture and off-characteristic frequency (CF) inhibition. Capture occurs when a component near CF saturates the inner-hair-cell response, reducing the amplitude of neural fluctuations, the low-frequency changes in auditory-nerve firing rate that arise from the interaction of components within complex sounds. IC neurons are sensitive to amplitude modulation and respond to neural fluctuations with enhanced or suppressed rates. Extracellular single-unit recordings were made in the IC of awake rabbits in response to synthetic harmonic stimuli with triangular-shaped spectral magnitudes and zero-phase. Stimuli were presented at 40–80 dB SPL to test for robust encoding over a range of suprathreshold levels. Responses were compared to the output of an auditory-nerve model that exhibits capture followed by a modulation-tuned midbrain model. Results were generally consistent with model predictions, especially for cells suppressed by modulations. Dependent on the modulation transfer function type, responses show dips or peaks in average rate over many levels when the peak harmonic frequency matches CF. This work fills a gap in neural timbre research by investigating spectral peak encoding in the auditory midbrain.