Abstract
Transcranial alternating-current stimulation (tACS) in the frequency range of 1–100 Hz has come to be used routinely in electroencephalogram (EEG) studies of brain function through entrainment of neuronal oscillations. It turned out, however, to be highly non-trivial to remove the strong stimulation signal, including its harmonic and non-harmonic distortions, as well as various induced higher-order artifacts from the EEG data recorded during the stimulation. In this paper, we discuss some of the problems encountered and present methodological approaches aimed at overcoming them. To illustrate the mechanisms of artifact induction and the proposed removal strategies, we use data obtained with the help of a schematic demonstrator setup as well as human-subject data.
Highlights
Low-current electrical stimulation of the human brain is a powerful technique developed in applied and experimental neuroscience (Herrmann et al, 2013; Paulus et al, 2013)
Demonstrator data were recorded with a basic montage consisting of four EEG leads connected to the phantom (Fp1, Fp2, Fpz as ground, and Cz as reference) as well as derivations of the transcranial alternating current stimulation (tACS) voltage and current signals fed into two of the bipolar AUX channels of the recorder; all signals were sampled at a rate of 1 kHz
The parameters t and m represent small adjustments of t and m, respectively ( t/t, m/m ≪ 1). They are needed to achieve an optimal subtraction of the artifact: the time adjustment t corrects for possible small phase differences between Vsig(t) and Vtacs(t), caused by the hardware or the analysis, whereas adjustments of the modulation index m correct for slow drifts of the modulation depth during averaging over a number of time spans
Summary
Low-current electrical stimulation of the human brain is a powerful technique developed in applied and experimental neuroscience (Herrmann et al, 2013; Paulus et al, 2013). Note that combinations of several methods have been used as well; for example, Fehér et al (2017) have subtracted a moving-average template followed by a principal component analysis (PCA) While most of these artifact-removal techniques are quite well-established, recent investigations by Noury et al (2016) and Noury and Siegel (2017) have revealed that unavoidable quasi-periodic physiological processes, like heartbeat and respiration, can induce additional, non-linear effects in the EEG through a rhythmic modulation of the main tACS frequency. We focus on stimulation frequencies applied in the EEG low and high gamma bands (30– 140 Hz) where modulation artifacts are most liable to impact the overall descending intrinsic EEG power spectrum. By applying the operator F to Equation (4), one finds
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
