EEG Artifact Removal in TMS Studies of Cortical Speech Areas

Karita S-T Salo,Selja M I Vaalto,Tuomas P Mutanen,Risto J Ilmoniemi

doi:10.1007/s10548-019-00724-w

Abstract

The combination of transcranial magnetic stimulation (TMS) and electroencephalography (EEG) is commonly applied for studying the effective connectivity of neuronal circuits. The stimulation excites neurons, and the resulting TMS-evoked potentials (TEPs) are recorded with EEG. A serious obstacle in this method is the generation of large muscle artifacts from scalp muscles, especially when frontolateral and temporoparietal, such as speech, areas are stimulated. Here, TMS–EEG data were processed with the signal-space projection and source-informed reconstruction (SSP–SIR) artifact-removal methods to suppress these artifacts. SSP–SIR suppressed muscle artifacts according to the difference in frequency contents of neuronal signals and muscle activity. The effectiveness of SSP–SIR in rejecting muscle artifacts and the degree of excessive attenuation of brain EEG signals were investigated by comparing the processed versions of the recorded TMS–EEG data with simulated data. The calculated individual lead-field matrix describing how the brain signals spread on the cortex were used as simulated data. We conclude that SSP–SIR was effective in suppressing artifacts also when frontolateral and temporoparietal cortical sites were stimulated, but it may have suppressed also the brain signals near the stimulation site. Effective connectivity originating from the speech-related areas may be studied even when speech areas are stimulated at least on the contralateral hemisphere where the signals were not suppressed that much.

Highlights

Transcranial magnetic stimulation (TMS) excites neurons noninvasively below the stimulation coil (Barker et al 1985); combined with electroencephalography (EEG), TMS can be applied to study effective connectivity, i.e., causal connections between different cortical areas (Ilmoniemi et al 1997; Komssi et al 2002; Massimini et al 2005)
We evaluated the ability of signal-space projection and source-informed reconstruction (SSP–SIR) to suppress muscle artifacts when cortical speech areas are stimulated
We showed that signal-space projection (SSP)–SIR is effective for removing most of the TMS-induced artifacts from EEG data when M1 or frontolateral or temporoparietal cortical areas, with large muscles, are stimulated (Mutanen et al 2013)

Summary

Introduction

Transcranial magnetic stimulation (TMS) excites neurons noninvasively below the stimulation coil (Barker et al 1985); combined with electroencephalography (EEG), TMS can be applied to study effective connectivity, i.e., causal connections between different cortical areas (Ilmoniemi et al 1997; Komssi et al 2002; Massimini et al 2005). Large muscles such as those located in the lateral sides of the head generate EEG artifacts that can be up to 1000 times the size of the neuronal EEG signal and can last tens of milliseconds (Rogasch et al 2013; Mutanen et al 2013) These muscle artifacts can make it hard or impossible to interpret the TMS-evoked potentials (TEPs) recorded with EEG (Nikulin et al 2003; Rosanova et al 2009; Cona et al 2011; Farzan et al 2013), especially when frontolateral and temporoparietal cortical areas are stimulated. This information could be directly used in the planning of therapeutic TMS protocols to treat language impairments, for example, those caused by stroke or other pathophysiologies (Hamilton et al 2011; Carreiras et al 2012; Thiel et al 2013; Heikkinen et al 2019)

Methods

Results

Conclusion