Removal of large muscle artifacts from transcranial magnetic stimulation-evoked EEG by independent component analysis

Abstract

We present two techniques utilizing independent component analysis (ICA) to remove large muscle artifacts from transcranial magnetic stimulation (TMS)-evoked EEG signals. The first one is a novel semi-automatic technique, called enhanced deflation method (EDM). EDM is a modification of the deflation mode of the FastICA algorithm; with an enhanced independent component search, EDM is an effective tool for removing the large, spiky muscle artifacts. The second technique, called manual method (MaM) makes use of the symmetric mode of FastICA and the artifactual components are visually selected by the user. In order to evaluate the success of the artifact removal methods, four different quality parameters, based on curve comparison and frequency analysis, were studied. The dorsal premotor cortex (dPMC) and Broca's area (BA) were stimulated with TMS. Both methods removed the very large muscle artifacts recorded after stimulation of these brain areas. However, EDM was more stable, less subjective, and thus also faster to use than MaM. Until now, examining lateral areas of the cortex with TMS-EEG has been restricted because of strong muscle artifacts. The methods described here can remove those muscle artifacts, allowing one to study lateral areas of the human brain, e.g., BA, with TMS-EEG.

The EEG correlates of the TMS-induced EMG silent period in humans

“…Approximately 13.0 ± 15.0% of trials were exclude in each subject, and the average number of clean trials per subject was 43 ± 9 trials. The TMS-related artifacts might have been due to short-lived high voltage TMS-induced artifact (Ives et al, 2006), TMS-induced activation of the peripheral nerves and cranial muscles near the coil (Korhonen et al, 2011, Maki and Ilmoniemi, 2011, Mutanen et al, 2012), the movement of the EEG sensors due to the electromotive forces (Sekiguchi et al, 2011) or the TMS-induced nerve and muscle activation, the TMS-induced accumulation of charges and their slow decay at every interface with capacitive properties such as the skin-electrodes interface (Veniero et al, 2009) or even the interface between several deeper epithelial layers of the skin (Julkunen et al, 2008). The time window contaminated by the large-amplitude TMS-related artifacts (36 ± 10ms) was replaced by interpolating the last artifact-free data point from the pre-stimulus period and the first artifact-free post-stimulus data point using cubic interpolation.…”

Cortical Dynamics of Sensorimotor Integration during Grasp Planning

“…Surviving data were rereferenced to the average signal of all sensors to remove any spatial effects on voltage differences with respect to the localization of the reference electrode. Data then entered into an independent component analysis (Lee et al, 1999;Makeig et al, 2004) to identify and remove residual signals related to eye movements, blinks, muscle tension, and TMS artifacts (Korhonen et al, 2011). After powerline noise was removed (using a discrete Fourier transform notch filter at 50, 100, and 150 Hz), data were bandpass filtered (0.75-150 Hz, sixth-order Butterworth filter) and downsampled to 500 Hz.…”

Smartware electrodes for ECG measurements : Design, evaluation and signal processing

“…To reduce the infliction of EMG noise, independent component analysis have been applied by Korhonen [56] to separate EMG signals from EEG recordings, a technique used also in ECG signal processing.…”