1st Leibniz MMS Days - Abstract
May, Patrick
The auditory system of the brain is tasked with analysing the spectral content of incoming sounds and then representing these in the context of preceding events. In our effort to understand these processes, we are incorporating the anatomical and physiological features of the auditory system into a computational model. Specifically, the serial structure of auditory cortex is combined with short-term synaptic plasticity into a model where the dynamical units represent the mean firing rates of local, excitatory and inhibitory neural populations. The result is a coarse description of the auditory system which offers measures of neural activity both on the local (invasive) and global (non-invasive) level. The model is able to account for a large variety of experimentally observed phenomena from auditory neuroscience: forward suppression, forward enhancement, stimulus-specific adaptation, selectivity to complex sounds, auditory streaming, and various MEG and EEG phenomena related to sensory memory. In so doing it bridges the gap between single- and multi-unit measurements in animal models and MEG/EEG experiments in humans. Methodologically, it is difficult to justify a priori the parameter values of the model. To partly overcome this difficulty, we are using linearization and diagonalization of the state equations of the model to gain analytical solutions. We are also planning to look at top-down decision processes occurring in working memory and auditory streaming paradigms and involving the interactions between auditory and frontal areas of cortex.