EEG-like signals can be synthesized from surface representations of single motor units of facial muscles

Gizem Yilmaz, Pekcan Ungan, Kemal S. Türker

Electrodes for recording electroencephalogram (EEG) are placed on or around cranial muscles; hence, their electrical activity may contaminate the EEG signal even at rest conditions. Due to its role in maintaining mandibular posture, tonic activity of temporalis muscle interferes with the EEG signal particularly at fronto-temporal locations at single motor unit (SMU) level. By obtaining surface representation of a motor unit, we can evaluate its interference in EEG and if we could sum surface representations of several tonically active motor units, we could estimate the overall myogenic contamination in EEG. Therefore, in this study, we followed re-composition (RC) approach and generated EEG-like artefact model using surface representations of single motor units (RC). Furthermore, we compared signal characteristics of RC signals with simultaneously recorded EEG signal at different locations in terms of power spectral density and coherence. First, we found that RC signal represented the power spectral distribution of an EMG signal. Second, RC signal reflected the discharge rate of a SMU giving the greatest surface representation amplitude and strongest interference appeared as distinguishable frequency peak on RC power spectra. Moreover, for strong interferences, RC also contaminated the EEG at F7 and other EEG electrodes. These findings are important to illustrate the susceptibility of EEG signal to myogenic artefacts even at rest and the research using EEG coherence comparisons should consider muscle activity while drawing conclusions about neuronal interactions and oscillations.