Approaches to mid-level auditory processing

The auditory system features a multi-stage processing hierarchy, comprising numerous brainstem nuclei between the inner ear and the cortex. These “mid-level” stages not only extract acoustic features but also support perception shaped by cognitive and behavioral context. This talk introduces several studies by the author and colleagues, focusing on such intermediate processing. One topic concerns how neural selectivity in the brainstem for features such as amplitude modulation and interaural cues might emerge from ecological demands. Using deep neural networks trained on natural sound recognition tasks, we showed that such tuning properties can arise through behaviorally relevant optimization in naturalistic environments. Another topic concerns the relationship between auditory attention and brainstem-related physiological responses. Attention plays a vital role in selecting relevant information from rich sensory input, enabling efficient processing with limited cognitive resources. Through measures such as eye movements, pupil diameter, otoacoustic emissions, and EEG, we investigated how attentional states are reflected at subcortical levels. Findings from various perspectives, including voluntary and involuntary attention, stimulus-driven modulation, and individual variability, highlight interplay between attention and brainstem activity. These studies, through diverse methodological approaches, help elucidate the functional significance of mid-level auditory processing in perception and adaptive behavior.