TMS Induced EEG Artifacts Analysis Based on the Partial Cross-Correlations

Combining Transcranial magnetic stimulation (TMS) and Electroencephaloraphy (EEG) offered insights into neural interaction during cognition which allows the investigation on the causal role of specific brain areas in behavior and the interactive connection between the brain areas. The TMS pulse in the TMS-EEG combining study introduces artifacts in the EEG electrodes which may mask the underlying neural activity lasting about 5 through even hundreds milliseconds after the pulse. In this study a statistical method on the removal of TMS induced EEG artifacts is presented based on partial cross-correlations. The method yields estimates of linear correlations between components obtained by Independent Component Analysis (ICA) and TMS signal that are not affected by linear correlations with other artifacts. We used the EEG data obtained from four healthy subjects without any neurological disorder, who were receiving single pulse TMS-EEG and sham-EEG stimulus on the left Broca’s area. ICA filters trained on the reduced version of 60 channel EEG data collected during single pulse 1426 Unjoo Lee and Woo-Kyoung Yo TMS-EEG and sham-EEG recordings. It identified the reduced number of statistically independent source channels, in which the ICA components originating from the TMS-induced artifact are classified by estimating the partial cross-correlation coefficients between ICA components of single pulse TMS-EEG and sham-EEG stimulus after ICA decomposition. The results showed which ICA components are related to TMS induced EEG artifacts, suggesting the efficiency and the reliability of the method developed in this study.