Nonequivalent After-Effects of Alternating Current Stimulation on Motor Cortex Oscillation and Inhibition: Simulation and Experimental Study

Abstract

The effects of transcranial alternating current stimulation (tACS) frequency on brain oscillations and cortical excitability are still controversial. Therefore, this study investigated how different tACS frequencies differentially modulate cortical oscillation and inhibition. To do so, we first determined the optimal positioning of tACS electrodes through an electric field simulation constructed from magnetic resonance images. Seven electrode configurations were tested on the electric field of the precentral gyrus (hand motor area). We determined that the Cz-CP1 configuration was optimal, as it resulted in higher electric field values and minimized the intra-individual differences in the electric field. Therefore, tACS was delivered to the hand motor area through this arrangement at a fixed frequency of 10 Hz (alpha-tACS) or 20 Hz (beta-tACS) with a peak-to-peak amplitude of 0.6 mA for 20 min. We found that alpha- and beta-tACS resulted in larger alpha and beta oscillations, respectively, compared with the oscillations observed after sham-tACS. In addition, alpha- and beta-tACS decreased the amplitudes of conditioned motor evoked potentials and increased alpha and beta activity, respectively. Correspondingly, alpha- and beta-tACSs enhanced cortical inhibition. These results show that tACS frequency differentially affects motor cortex oscillation and inhibition.

Keywords: electric field simulation; oscillation; primary motor cortex; spike-timing-dependent plasticity; transcranial alternating current stimulation.

Figure 1

The hypothesized relationship between tACS frequency and neuronal activity. Gray lines denote 10 Hz (top trace) and 20 Hz (bottom trace) tACS, and red lines denote up and down states of neural firing. We hypothesized that (A) 10 Hz oscillations would be synchronized with the peak phase of 10 Hz and 20 Hz tACS, and (B) 20 Hz oscillations would be synchronized with both the peak and trough phases of 10 Hz tACS, as well as the peak phase of 20 Hz tACS.

Figure 2

Electric field simulation. Seven tACS electrode configurations based on the International 10–20 system (C1-Pz, FC1-Pz, FC3-Pz, C3-Pz, Cz-CP1, C1-CPz, and C3-CPz) and anatomical structure from MRI image (A). The electric fields induced by each of the seven tACS montages were averaged on the precentral knob in the precentral area (B).

Figure 3

(A) The experimental design for testing the effects of the alpha- (10 Hz), beta- (20 Hz), and sham-tACS conditions on brain oscillations and cortical inhibition. Testing was performed on three different days. (B) Five EEG electrodes were placed at the FDI muscle hotspot (HS), and 2.5 cm lateral anterior (LA), medial anterior (MA), lateral posterior (LP), and medial posterior (MP) to the hotspot.

Figure 4

Normal component of the electric field (group-level analysis, n = 18) during tACS in seven montages. For practical comparison, the tACS phase depicted here was chosen so that the electric field’s normal component is towards the precentral wall.

Figure 5

Grand-averaged power spectra before and after (A) alpha-, (B) beta-, and (C) sham-tACSs. Dashed and solid lines denote the power spectra before and after tACS, respectively. Shaded areas indicate the standard error of the mean. The power spectrum of alpha-band oscillation was increased after alpha-tACS, whereas beta-band oscillation was increased after beta-tACS. However, the power spectra of alpha- and beta-band oscillations were not changed by sham-tACS.

Figure 6

Normalized power changes in (A) alpha and (B) beta-oscillatory neural activity after alpha-, beta-, and sham- tACSs. Dots and error bars denote the mean and standard error of the mean, respectively. Alpha-tACS resulted in an increase in alpha power oscillations and decreased beta power oscillations, whereas beta-tACS increased beta power oscillations. *: p < 0.05.

Figure 7

Grand-averaged time-series of the (A–C) conditioned and (D–F) unconditioned MEP amplitudes by the LLT model. Dashed and solid lines indicate actual and estimated MEP amplitudes, respectively. Shaded areas indicate the standard error of the mean. The actual MEP amplitudes fluctuated randomly, whereas the fluctuation of estimated MEP amplitudes was reduced by the LLT model.

Figure 8

The normalized (A) conditioned and (B) unconditioned MEP amplitude among alpha-, beta-, and sham-tACS. Dots and error bars denote the mean and standard error of the mean, respectively. Alpha- and beta-tACSs decreased both conditioned and unconditioned MEP amplitudes. *: p < 0.05.