Investigation of Spatiotemporal Profiles of Single-Pulse TMS-Evoked Potentials with Active Stimulation Compared with a Novel Sham Condition

Abstract

Identifying genuine cortical stimulation-elicited electroencephalography (EEG) is crucial for improving the validity and reliability of neurophysiology using transcranial magnetic stimulation (TMS) combined with EEG. In this study, we evaluated the spatiotemporal profiles of single-pulse TMS-elicited EEG response administered to the left dorsal prefrontal cortex (DLPFC) in 28 healthy participants, employing active and sham stimulation conditions. We hypothesized that the early component of TEP would be activated in active stimulation compared with sham stimulation. We specifically analyzed the (1) stimulus response, (2) frequency modulation, and (3) phase synchronization of TMS-EEG data at the sensor level and the source level. Compared with the sham condition, the active condition induced a significant increase in TMS-elicited EEG power in the 30-60 ms time interval in the stimulation area at the sensor level. Furthermore, in the source-based analysis, the active condition induced significant increases in TMS-elicited response in the 30-60 ms compared with the sham condition. Collectively, we found that the active condition could specifically activate the early component of TEP compared with the sham condition. Thus, the TMS-EEG method that was applied to the DLPFC could detect the genuine neurophysiological cortical responses by properly handling potential confounding factors such as indirect response noises.

Keywords: dorsolateral prefrontal cortex (DLPFC); electroencephalography (EEG); sham coil; source-based analysis; transcranial magnetic stimulation (TMS).

Figure 1

Averaged GMFP for all participants and averaged LMFP at the DLPFC stimulation site. (a) The upper-left panel shows the averaged GMFP waveforms (mean ± S.E.) of the TEP for all participants under active and sham conditions (active stimulation is shown as red waveform and sham stimulation is shown as blue waveform). (b) Likewise, the lower-left panel shows the averaged LMFP waveforms (mean ± S.E.) of the TEP for all participants confined to the DLPFC stimulation site electrodes (F3, F5, F1, F7, and AF3) (active stimulation is shown as red waveform, sham stimulation is shown as blue waveform). The light-blue bars below each graph on the left side indicate the time intervals where the statistical test showed a significant difference between the two conditions (30–60 ms). (c) The figure on the right shows the topoplots corresponding to the time intervals for the early component (30–60 ms) and the late component (100–200 ms) (top: active condition; middle: sham condition; bottom: active condition–sham condition). The electrode sites that showed significant differences between the two conditions were marked with a black asterisk. S.E.: standard error.

Figure 2

Time–frequency analyses (total power and ITPC) in the active and sham conditions. Upper panel: Total power analysis in the DLPFC stimulation electrode sites for each condition averaged over all participants. The panel (a) represents the total power elicited by active stimulation, panel (b) represents the total power elicited by sham stimulation, and panel (c) shows the difference in total power between the active and sham conditions (i.e., (a) minus (b)). Lower panel: ITPC analysis in the DLPFC stimulation electrode sites for each condition averaged over all participants. The panel (a) shows the ITPC map elicited by active stimulation, panel (b) shows the ITPC map elicited by sham stimulation, and panel (c) shows the difference in ITPC between the active and sham conditions (i.e., (a) minus (b)). The inner black line in the graph in panel (c) shows the area of ITPC where significant group differences were found between the active and sham conditions.

Figure 3

Graph theory-based network analyses for the wPLI and the comodulogram of phase–amplitude coupling expressed in the modulation index (MI). The left panel of Figure 3 shows the mean betweenness centrality of all participants in the active condition (left column), the sham condition (middle column), and the statistical map of differences between the active and sham conditions (right column). Each panel shows the results of the theta-band (4–8 Hz), alpha-band (8–13 Hz), beta-band (13–30 Hz), and gamma-band (30–100 Hz) analysis, from top to bottom, respectively. The white circles in the figure indicate the electrode sites that were regarded as hubs in the whole brain network based on the following score criteria: (1) the top 20% with the lowest value of clustering coefficients, (2) the top 20% with the shortest path lengths, (3) the top 20% with the highest degree, and (4) the top 20% with the highest value of betweenness centralities (BC). On the other side, the right panel of Figure 3 represents the comodulogram at the left fronto-central area (i.e., FC5 electrode site) that showed a significant difference in the MI values between the active and sham conditions in the interval 30–200 ms after TMS stimulation. Here, (a–c) are comodulograms showing PAC findings in the active condition, sham condition, and the difference between the two conditions, respectively.

Figure 4

Source-based TMS-elicited response and their time–frequency analyses. (a) The upper panel shows the averaged TMS-elicited response for all participants confined to the source-based region, including the DLPFC stimulation site (the rostral middle frontal gyrus) with active and sham conditions (active stimulation is shown as red waveform and sham stimulation is shown as blue waveform). The intervals at which significant differences (30–60 ms) were found in permutation t-tests are indicated by the light-blue bars. (b) The upper panel shows the total power based on time–frequency analysis with the active condition (upper left) and sham condition (upper right) in the source-based region, including the DLPFC stimulation site. The lower-left panel shows the differences in total power between the active and sham conditions. A two-dimensional cluster-based permutation test did not detect any significant differences between the two conditions (lower right). (c) The upper panel shows the ITPC based on time–frequency analysis with active condition (upper left) and sham condition (upper right) in the source-based region, including the DLPFC stimulation site. The lower-left panel shows the differences in ITPC between the active and sham conditions. The areas showing significant differences between the two conditions in the two-dimensional cluster-based permutation test are marked with a green mask (lower right).