Modeling human auditory evoked brain responses to complex sounds: Modeling human auditory evoked brain responses to complex sounds

Hearing deficits are typically linked to the degeneration of specific structures within the cochlea. While pure-tone audiometry represents the standard for hearing assessment, it does not capture all aspects of cochlear pathophysiology. On the other hand, auditory evoked potentials (AEP) provide objective information about the status of various cell populations across the auditory pathway by measuring their synchronized responses to sound stimuli. Clinically, potentials generated in the brainstem are widely used as indicators of cochlear health due to their relatively large amplitudes. However, the relationship between cochlear pathology and brainstem responses is not fully understood, making it challenging to infer cochlear function from non-invasive AEP measurements in humans. This research employed an AEP modeling framework, incorporating a computational auditory nerve (AN) model for cochlear processing and a linear convolution stage, to simulate full-waveform evoked potentials and analyze how cochlear processing influences brainstem responses. An experimental setup was also established to simultaneously record brainstem and peripheral evoked potentials for comparison with the simulated data. This work revealed that brainstem responses to periodic stimuli may be more sensitive to age-related neural degeneration in the cochlea than responses to transients, while remaining more robust to the loss of outer hair cells. Simulated AEPs qualitatively resembled experimental responses to various stimuli across different intensities. Certain disparities with the measured responses led to revisions to the computational AN model which significantly improved the amplitudes and latencies of simulated AEPs across varying stimulus intensities. In addition, the AEP modeling framework was applied to simulate responses from a novel clinical testing paradigm, which led to reduced test times and greater specificity regarding cochlear status. The modeling framework could explain the main effects observed in such experimental brainstem responses and was deemed useful in quantifying and explaining the underlying mechanisms within the cochlea. This thesis demonstrates that the proposed AEP model framework offers important
insights into the influence of cochlear processing on brainstem-evoked potentials. This framework, combined with specific electrophysiological responses in humans, may allow for the improvement of physiologically accurate human auditory models, which are crucial for the development of new and more precise diagnostics.