Modeling human auditory evoked brainstem responses based on nonlinear cochlear processing

The aim of this study was to accurately simulate auditory evoked potentials (AEPs) from various classical stimuli such as clicks and tones, often used in research and clinical diagnostics. In an approach similar to Dau (2003), a model was developed for the generation of auditory brainstem responses (ABR) to transient sounds and frequency following responses (FFR) to tones.

The model includes important cochlear processing stages (Zilany and Bruce, 2006) such as basilar-membrane (BM) tuning and compression, inner hair-cell (IHC) transduction, and IHC auditory-nerve (AN) synapse adaptation. To generate AEPs recorded at remote locations, a convolution was made on an empirically obtained elementary unit waveform with the instantaneous discharge rate function for the corresponding AN unit.

AEPs to click-trains, as well as to tone pulses at various frequencies, were both modelled and recorded at different stimulation levels and repetition rates. The observed nonlinearities in the recorded potential patterns, with respect to ABR wave V latencies and amplitudes, could be largely accounted for by level-dependent BM processing as well as effects of short-term neural adaptation.

The present study provides further evidence for the importance of cochlear tuning and AN adaptation on AEP patterns, and provides a useful basis for the study of more complex stimuli including speech.