Neural dynamics of human auditory behavior: computational model-based analysis

Segun Goh, Kyungreem Han, and MooYoung Choi


In the human brain, auditory information is processed in the complex network of auditory cortical neurons; quantitative and qualitative facets of the human auditory information processing have been studied with the help of the indirect measurements of brain activities such as electroencephalography (EEG) and magnetoencephalography (MEG). In particular, the analysis of the auditory evoked potential (AEP), obtained from ongoing EEG/MEG signals, has been considered as a standard method to understand the dynamics of the process. Due to the nonlinearities in the auditory cortical neurons, however, there exist difficulties in interpreting relevant information as to the intrinsic properties of the network, obtained from the event-related potential (ERP) via conventional linear methods. To circumvent these limitations, we in this study, consider various nonlinear analysis methods, by means of which ERP and/or ongoing EEG recordings of human subjects are quantitatively analyzed. Further, we propose a computational model for the human auditory behavior and analyze various facets of the behavior based on the model. Specifically, the structural connectivity of the brain network is constructed from the diffusion spectrum imaging (DTI) data and the resulting network of neural oscillators is described by the Kuramoto-type model. Examining the observed ERP signals and the transient dynamics of the Kuramoto model parameters, we probe the mechanism of the ERP generation and its relation to the resting brain state.