P 54 Multivariate pattern analysis reveals location specific aftereffects of 10Hz motor cortex transcranial alternating current stimulation. (May 2022)

• Record Type: Journal Article
• Authors: Tzvi-Minker, E.; Alizadeh, J.; Schubert, C.; Classen, J.
• Is Part Of: Clinical neurophysiology. Volume 137(2022)
• Page Start: e45
• Page End: e46
• Publication Date: 2022-05
• Subjects: Neurophysiology -- Periodicals; Electroencephalography -- Periodicals; Electromyography -- Periodicals; Neurology -- Periodicals; 612.8
• DOI: 10.1016/j.clinph.2022.01.085
• Languages: English
• ISSNs: 1388-2457

Abstract

Abstract : Introduction: Transcranial alternating current stimulation (tACS) affects endogenous oscillations "online", during the stimulation, and "offline", after the stimulation terminated, presumably reflecting neuroplasticity. This aftereffect has been previously explored by applying tACS at 10Hz or the individual alpha frequency (α, 8-12Hz) during rest as well as during task performance. While over parieto-occipital cortex, α power increased following tACS, motor cortex tACS led to a decrease in α power. The spatial specificity of this effect has not been investigated thus far. In this study, we explored 10Hz tACS aftereffects on the motor network when specific nodes of the network are targeted. The results of this work could have important consequences for therapeutic interventions using tACS. Methods: 25 young subjects received 10Hz tACS in 3 sessions: active tACS to left motor cortex (lMC) or right cerebellum (rCB) for 20min, as well as sham (either location, 10sec) (Fig. 1A, C). The order of the sessions was balanced across subjects. EEG was collected before (PRE) or after (POST) tACS (Fig. 1B). Signals were processed to eliminate eye-blink, muscle and other artifacts and multivariate pattern analysis (MVPA) was applied to the clean EEG power at different frequency bands. To this end, a classifier was trained to distinguish short segments of POST-PRE EEG power between the stimulation protocols. The classifier was then tested on other POST-PRE EEG power segments to produce a classification accuracy measure. Notably, this method accounts for the multitude facets of the EEG signal and thus is more sensitive to subtle modulation of the EEG signal compared to standard averaging methods. The analysis was performed in both the electrode and the source-space using beamforming. Non-parametric cluster-based Monte-Carlo permutation testing was used to distinguish classification accuracies between stimulation protocols at group level. Results: EEG signals in both theta (θ, 4-8Hz) and α, as well as during motor task performance and rest, were better classified to lMC-tACS compared to rCB-tACS/sham, specifically at lMC-tACS stimulation locations (electrodes FC3 and CP3, Fig. 2A-C). This effect was associated with a decrease in α power over the motor cortex at rest following tACS (Fig. 2D). In source-space, better classification accuracies at lMC-tACS were found in premotor cortex but not in primary motor cortex as a computational model of the current distribution suggested (Fig. 1C). Discussion: Our results demonstrate a clear broadband (θ, α) aftereffect of 10Hz tACS on the premotor cortex, independent of the state (both task and rest). These results may indicate location-specific neuroplasticity due to tACS and could motivate therapeutic interventions targeting θ/ α oscillations in the motor network. Fig. 1.