Decoding of Single Auditory Features Investigated by Mismatch Negativity

• ¹ Centre for Functional and Diagnostic Imaging and Research, Copenhagen University Hospital, Danish Research Centre for Magnetic Resonance, Denmark
• ² Technical University of Denmark, DTU Compute, Denmark
• ³ H. Lundbeck A/S, Synaptic Transmission 2, Denmark
• ⁴ University of Copenhagen , Department of Neurology, Psychiatry and Sensory Sciences, Denmark
• ⁵ Mental Health Centre Sct. Hans Copenhagen University Hospital, Research Institute of Biological Psychiatry, Denmark
• ⁶ Centre for Functional and Diagnostic Imaging and Research, Copenhagen University Hospital, Danish Research Centre for Magnetic Resonance, Denmark
• ⁷ Bispebjerg Hospital, University of Copenhagen, Department of Neurology, Denmark

A fundamental task our brain successfully masters on a daily basis is to detect unexpected changes in our environment. This process can be studied electrophysiologically with electroencephalography (EEG) by measuring the event related response to auditory irregularities in a stream of standard tones. Such irregularities give rise to the so-called mismatch negativity (MMN), an evoked fronto-central response emerging 100 - 150 ms after onset of an oddball stimulus (R. Naatanen et. al 2007). Previous studies used a classical oddball paradigm to test up to five different deviant types within one paradigm, but failed to investigate the memory trace effect for each of the different oddballs and how this leads to the prediction of sensory changes. Here we propose a multiplex MMN roving paradigm, independently introducing pitch and duration deviants, which then become the new standard within their domain, pitch and duration respectively. This paradigm evokes specific responses to distinct auditory changes and allows to characterize the memory trace to single auditory features. We successfully separated the frontal MMN responses to pitch from duration deviants within our multiplex roving paradigm in healthy human subjects. We compared these signals to responses from a pure pitch roving paradigm using a simple parametric comparison between the standards and the deviants. Our results show similar MMN prefrontal responses to pitch in the pure pitch and the multiplex paradigm. This is in line with previous studies on mismatch negativity and predictive coding suggesting prefrontal regions to play a role in integrating bottom-up sensory signals with top-down predictive signals (Rao and Ballard, 1999). Further, our results imply that single auditory features are decoded independently from each other and hence our findings may contribute to the characterization of change detection responses in the human cortex.