The Auditory Mismatch Negativity as an EEG-derived Measure of Predictive Coding (en)

* Presenting author
Day / Time: 07.03.2023, 14:20-14:40
Room: Saal X 7-8
Typ: Regulärer Vortrag
Abstract: The predictive coding framework of neuronal processing hypothesizes that the layers in the processing hierarchy propagate prediction errors rather than abstractions of sensory input. In the auditory system a detuned tone in the sequence of a well-known melody would represent a source of such a prediction error. Instead of the tone being processed and sequentially integrated with the rest of the sequence (feedforward), it is compared to the expected next tone in the sequence, predicted from memory (feedback). Previous research has shown that deviant stimuli in a sequence of standards elicit the mismatch negativity (MMN) response. The MMN is a component of the scalp-recorded electroencephalogram (EEG) that is suggested to carry the prediction error. Within the predictive coding framework, it is thus expected that deviant stimuli elicit different MMNs in response to various deviating features of the stimulus. In this work, the aim is a systematical analysis of differences between MMNs to stimuli which deviate in one, or a combination of two features.