Bifocal tACS over the primary sensorimotor cortices increases interhemispheric inhibition and improves bimanual dexterity

Abstract

Concurrent application of transcranial alternating current stimulation over distant cortical regions has been shown to modulate functional connectivity between stimulated regions; however, the precise mechanisms remain unclear. Here, we investigated how bifocal transcranial alternating current stimulation applied over the bilateral primary sensorimotor cortices modulates connectivity between the left and right primary motor cortices (M1). Using a cross-over sham-controlled triple-blind design, 37 (27 female, age: 18 to 37 yrs) healthy participants received transcranial alternating current stimulation (1.0 mA, 20 Hz, 20 min) over the bilateral sensorimotor cortices. Before and after transcranial alternating current stimulation, functional connectivity between the left and right M1s was assessed using imaginary coherence measured via resting-state electroencephalography and interhemispheric inhibition via dual-site transcranial magnetic stimulation protocol. Additionally, manual dexterity was assessed using the Purdue pegboard task. While imaginary coherence remained unchanged after stimulation, beta (20 Hz) power decreased during the transcranial alternating current stimulation session. Bifocal transcranial alternating current stimulation but not sham strengthened interhemispheric inhibition between the left and right M1s and improved bimanual assembly performance. These results suggest that improvement in bimanual performance may be explained by modulation in interhemispheric inhibition, rather than by coupling in the oscillatory activity. As functional connectivity underlies many clinical symptoms in neurological and psychiatric disorders, these findings are invaluable in developing noninvasive therapeutic interventions that target neural networks to alleviate symptoms.

Introduction

Transcranial alternating current stimulation (tACS) employs weak, alternating currents and is thought to have neuromodulatory effects by inducing phase synchronization between endogenous brain oscillations and the applied tACS frequency (Thut et al. 2012; Herrmann et al. 2013). The simultaneous application of tACS to distant cortical sites has been shown to alter communication between the stimulated regions, indicating a modulatory effect on functional connectivity (Polanía et al. 2012). Furthermore, it is believed that tACS promotes long-term potentiation-like plasticity even after the stimulation ends, suggesting its potential to induce enduring changes in brain connectivity and function (Vossen et al. 2015; Wischnewski et al. 2019). Impaired functional connectivity underlies various psychiatric, neurological, and age-related conditions (O’Reilly et al. 2017; Marzola et al. 2023; Oliveira and Ferreira 2023; Zhang et al. 2023). Therefore, enhancing functional connectivity between cortical sites positions tACS as a powerful tool for research with potential therapeutic applications (Tavakoli and Yun 2017).

Modulating functional connectivity by simultaneously stimulating distant cortical regions by tACS is based on established principles of brain network dynamics. The “communication-by-coherence” hypothesis posits that neuronal communication is optimal when the neural oscillatory phases of the communicating cortices are synchronized (Fries 2005; Fries 2015). Given that motor and cognitive tasks are not executed by a single brain region in isolation but involve complex interactions across multiple brain regions (eg Battleday et al. 2014), it is reasonable to assume that simultaneous stimulation of functionally relevant cortical regions via tACS may result in improved functional communication between the stimulated regions (Violante et al. 2017; Reinhart and Nguyen 2019; Grover et al. 2022; Meng et al. 2023).

Polanía et al. (2012) empirically demonstrated that bifocal tACS modulated functional connectivity between cortical regions during a working memory task. They applied 0° relative phase (in-phase) and 180° relative phase (anti-phase) stimulation to the left prefrontal and parietal cortices, finding that in-phase stimulation enhanced working memory performance, while antiphase stimulation led to a performance decline. These results suggest that tACS likely modified the phase alignment of endogenous oscillatory activity within fronto-parietal networks that are instrumental in working memory performance. Further research by Fujiyama et al. (2023) showed that in-phase beta frequency (20 Hz) bifocal tACS applied over the right inferior frontal gyrus (rIFG) and presupplementary motor area improved response inhibition and increased task-related functional connectivity observed with electroencephalography (EEG). Similarly, other studies have reported beneficial effects of bifocal tACS on various cognitive and motor functions, particularly in the theta and beta frequency ranges (Violante et al. 2017; Reinhart and Nguyen 2019; eg Grover et al. 2022; Meng et al. 2023). While these findings demonstrate that bifocal tACS can enhance functional connectivity and related behaviors, the exact mechanisms underlying these effects remain elusive. To further elucidate these mechanisms, a multimodal approach combining different neurophysiological techniques could be beneficial.

EEG studies suggest that tACS induces oscillation entrainment, synchronizing brain activities with the phase of the applied tACS frequency (Thut et al. 2012; Herrmann et al. 2013). In vivo animal studies indicate that applied sinusoidal currents entrain endogenous oscillations (Reato et al. 2013); however, evidence remains suggestive but unclear in humans. Evidence of oscillatory entrainment primarily focuses on how tACS induces effects during stimulation. However, behavioral changes that persist beyond the stimulation period are thought to result from plasticity mechanisms (Polanía et al. 2012). Specifically, the ongoing synchronization of oscillatory activity by tACS appears to regulate the precise timing of neural spikes, leading to long-term potentiation demonstrably distinct from entrainment-related aftereffects (Vossen et al. 2015). Alternate theories suggest that plasticity mechanisms may better explain how tACS induces oscillatory changes and subsequent behavioral changes (Vossen et al. 2015). While EEG has a high temporal resolution, which is beneficial for characterizing oscillatory coupling induced by bifocal tACS, it provides relatively limited spatial resolution, offering a global measure of neural activity with less spatial specificity in characterizing the bifocal tACS-induced changes in cortical connectivity. In contrast, transcranial magnetic stimulation (TMS) offers greater spatial precision, particularly when investigating connectivity between primary sensorimotor cortices. Dual-site TMS (ds-TMS) with paired electromyography (EMG) protocols can be used to investigate interhemispheric inhibition (IHI) between the primary motor cortices (M1s) (eg Mochizuki et al. 2004; Bäumer et al. 2006; Koch et al. 2006; O’Shea et al. 2007; Ni et al. 2009), facilitating the precise assessment of intracortical M1 circuits and the connected contralateral corticospinal neurons, enabling a more precise understanding of the interactions between cortical regions (Ferbert et al. 1992).

This study aimed to investigate the neurophysiological mechanisms underlying how concurrently applied beta tACS-induced changes in functional connectivity between the bilateral M1s. The choice of beta frequency was guided by the understanding that beta oscillations (13 to 30 Hz) play a crucial part in movement preparation and execution (see review, Kilavik et al. 2013). To date, studies investigating the impact of tACS targeting M1 on functional connectivity measured with EEG have reported mixed results. No modulation of resting-state ImCoh was found by Wischnewski et al. (2019) in which unifocal beta tACS over the left M1 was applied. In contrast, Wansbrough et al. (2024) reported current intensity-dependent modulations in ImCoh following unifocal beta tACS targeting the left M1 with different current intensities. Specifically, sham and 0.5 mA stimulation resulted in an increase in beta ImCoh, while 1.0 and 1.5 mA stimulation decreased beta ImCoh. These mixed findings suggest the need for further investigation to better understand the effects of beta tACS on functional interactions.

A multimodal approach was utilized employing EEG, ds-TMS, and a behavioral measure using the Purdue pegboard to examine the effect of bifocal tACS and sham on neuronal phase synchronization, IHI, and manual dexterity, respectively. We specifically focus on interhemispheric interactions between the left and right M1 to answer respective research questions since the assessment protocols for M1-M1 functional and structural connectivity and its involvement in motor functions are well established (e.g., Bäumer et al. 2006; Ni et al. 2009; Fujiyama et al. 2016). We predict that the simultaneous application of beta-frequency in-phase tACS would enhance functional connectivity between the left and right M1 regions, observed across neurophysiological measures, and would consequently result in improved manual dexterity when compared to sham. Identifying the underlying mechanisms by which tACS improves functional connectivity is crucial, as enhancing functional connectivity holds significant clinical potential for developing future interventions for both health and disease.

Materials and methods

Participants

A power analysis was performed for sample size estimation using the “simr” package (Lenth 2020) in R statistical package, version 4.4.1 (R Core Team 2024), drawing on data from our previous studies investigating the effect of noninvasive brain stimulation on neurophysiological measures (Fujiyama et al. 2016, 2022). With an alpha level set at 0.05 and a target power of 0.85, consistent with recommendations for minimizing the risk of Type II errors in neurophysiological studies (Lakens 2013), we estimated that a sample size of 34 participants would be sufficient to observe a large pre–post change (Cohen’s d = 0.8), as supported by previous studies with similar designs (e.g., Fujiyama et al. 2017, 2022) employing neurophysiological measures. To ensure a conservative estimate and enhance statistical power, we recruited 37 (27 female) participants, which provides a power of approximately 0.9.

Participants were aged 18 to 37 years (M = 24.70 years, SD = 6.18 years) and were recruited through a university research participation portal. Student participants earned testing time-equivalent course credits for both the real-tACS and sham sessions or received an AUD20 gift voucher following completion of the final session. All participants were screened for noninvasive brain stimulation contraindications, and those with psychiatric or neurological conditions or prescription medications that conflicted with stimulation were excluded (Rossi et al. 2021). The Edinburgh Handedness Inventory (Oldfield 1971) was used to assess participant handedness. Only those with a score above 40, indicating right-handedness, were eligible to participate, as left-handedness has been associated with differing motor cortical representations and noninvasive stimulation aftereffects (Nicolini et al. 2019; Fitzgerald et al. 2021). Written informed consent was obtained before participation in the study. This study was approved by the Murdoch University Ethics Committee (2021/240).

Materials

tACS

Bifocal tACS was delivered with a neuroConn DC-STIMULATOR MC (NeuroConn, Ilmenau, Germany) using round rubber electrodes and conductive ten20 paste (Weaver and Company, Aurora, CO, USA) in a 4 × 1 montage (Fig. 1A). A central electrode (2 cm diameter) was positioned over the scalp site targeting the hand representation of M1 in each hemisphere, identified by TMS (see the TMS and EMG Recording section for more details), with four surrounding return electrodes (2 cm diameter) placed in a radius of 3 cm, measured from the center of the central electrode to the center of each return electrode (Villamar et al. 2013), which is the optimal configuration for M1 focal stimulation (Edwards et al. 2013). Electric field modeling demonstrates (Fig. 1A) that both the primary motor and primary sensory areas are stimulated via the employed tACS montage. Alternating currents were applied at 20 Hz at 1.0 mA peak-to-peak amplitude and with zero DC offset to both stimulation sites at 0° phase lag.

Two conditions were employed for 21 min: tACS at 1.0 mA and sham stimulation. During tACS, the intensity ramped up to the target current intensity for 30 s and maintained intensity for 20 min before ramping back down at the last 30 s. In the sham condition, the tACS stimulation ramped up to 1.0 mA and immediately ramped down at the beginning and end of the 20 min for 30 s to simulate stimulation (Woods et al. 2016). The tACS machine was preprogrammed by a research associate with codes representing the two conditions to ensure both participant and researcher blinding.

TMS and EMG recording

TMS assessments were performed using two Magstim 200² units (Magstim Company, Dyfed, UK) with two 50 mm (D50a coils, Magstim Company, Dyfed, UK) figure-of-eight coils. Surface EMG electrodes recorded TMS-induced motor-evoked potentials from participants’ first dorsal interosseous (FDI) muscles. EMG surface electrodes (Ag/AgCl) were positioned in a belly-tendon montage over the FDI. Signals were amplified with a gain of 1000, band pass–filtered (10 to 500 Hz), and sampled at 2000 Hz using a 16-bit AD system (CED1902, Cambridge, UK) for analysis. TMS coils were positioned tangentially, 45° from the sagittal midline, to induce a posterior–anterior current flow, optimal for eliciting motor evoked potentials (MEPs) in the FDI, ie motor hotspot (Jensen et al. 2005). The motor hotspot was identified by systematically repositioning the coil above the estimated target area until the location that provoked the strongest MEP was found. At the beginning of each session, each individual’s resting motor threshold (rMT) was determined as the lowest intensity that evoked MEPs in the FDI of greater than 50 μV in at least three out of five consecutive trials (Carroll et al. 2008) at the marked hotspot.

IHI was assessed using a ds-TMS paradigm following the methodology outlined by Ferbert et al. (1992). We investigated M1-M1 projections in both directions (left to right hemisphere and vice versa). The first condition of our protocol involved delivering a single-pulse testing stimulus (TS) set at an intensity of 130% of rMT. The following condition employed ds-tMS delivering a conditioning stimulation (CS), intensity set at 110% of rMT, to one M1 followed by the TS to the contralateral M1 at an interstimulus interval (ISI) of 10 ms, and then the third condition repeated this double pulse stimulation protocol with an ISI of 40 ms (Ni et al. 2009; Kroeger et al. 2010; Hinder et al. 2012). Previous research suggested that short ISIs (≤10 ms) are associated with IHI mediated by postsynaptic GABA_A mechanisms, while longer ISIs (≥40 ms) involve GABA_B-ergic circuits (Sanger et al. 2001; Irlbacher et al. 2007).

At each TMS assessment time point, there were two blocks of IHI, testing left M1- right M1 and right M1- left M1 IHI in counterbalanced order across sessions and participants. Each block lasted 3 min and consisted of 60 trials, alternating between 3 conditions: 20 single-pulse stimuli (TS), 20 CS-TS IHI10 (CS followed 10 ms later by TS), and 20 CS-TS IHI40 (CS followed 40 ms later by TS) stimuli. The TMS pulses administered throughout hotspot identification, rMT determination, and IHI were applied every 4 to 6 s to avoid inducing neuroplastic changes in the brain (Rossi et al. 2021).

E‌EG

EEG was recorded using a 128-electrode EEG HydroCel Geodesic Sensor Net (Magstim EGI, Eugene, OR). Net Station (5.4.2) software recorded sensor-level EEG signals from Ag-AgCl scalp electrodes. EEG was amplified using a Net Amps 300 amplifier, low- and high-pass-filtered (0.1 to 500 Hz) with a 1000 Hz sample rate. Signals were referenced to Cz during recording and electrode impedance was kept below 50 kΩ as recommended by the manufacturer (Magstim EGI, Eugene, OR). During resting-state EEG recording, participants viewed a fixation cross presented on a personal computer (PC) screen for 3 min.

Behavioral assessment

The Purdue Pegboard Test assessed gross motor ability and fine motor dexterity (Tiffin and Asher 1948). We employed three conditions: unimanual (left and right) and a bimanual assembly task. Participants performed all conditions seated at a table and completed practice trials before the assessment. In the unimanual task, participants were instructed to place as many pegs as possible with one hand into a pegboard within 30 s. The number of pegs successfully placed was recorded for each hand. For the assembly task, participants were instructed to use both hands to assemble a peg, washer, collar, and washer down the right column of the pegboard. The number of individual pieces assembled on the board within 60 s was recorded.

Control measures

Sleep, caffeine, and alcohol questionnaire

The Sleep, Caffeine, and Alcohol (SCA) questionnaire consists of four questions designed to record various factors that could influence tACS effects and EEG results. It measures sleep quality on a scale from 1 (poor) to 10 (excellent), the number of hours slept, and the amount of caffeine (in milligrams) and alcohol (in units) consumed in the last 12 h. The SCA was administered during both tACS sessions to account for individual variations that might affect outcomes (Valenzuela 1997; Zulkifly et al. 2021).

Transcranial electrical stimulation sensation questionnaire

After each session, participants completed the Transcranial Electrical Stimulation (tES) Sensation Questionnaire. This questionnaire first assessed 13 common sensations, such as burning, tingling, itching, and headaches, experienced during tACS using a Likert scale ranging from 1 (nothing) to 5 (very strong) (Woods et al. 2016). It also included items to determine when these sensations started (beginning, middle, or end of stimulation) and stopped (soon, in the middle, or at the end of stimulation). Additionally, participants rated how these sensations affected their PPT performance using a 5-point scale from “absolutely not” to “very much.” This questionnaire also evaluated the effectiveness of blinding regarding the condition participants received.

Study design

This study employed a cross-over sham-controlled triple-blind within-subject design to investigate the effect of bifocal tACS on M1-M1 functional connectivity and motor function using EEG and TMS. All participants underwent two identical experimental sessions, with assessment measures recorded before and after either real or sham tACS conditions. The order of sessions was counterbalanced across participants, and both sessions occurred at a similar time of day. After the initial session, participants returned for testing a minimum of 7 days later, as the effects of a single session of noninvasive brain stimulation are likely to be diminished by 7 days (eg Fujiyama et al. 2017). Following this minimum of 7 days, testing took place at the same time as the primary session to control for the possible effects of cortisol level fluctuation throughout the day that may affect tES responses (Sale et al. 2008).

Procedure

Figure 1B illustrates a procedural schematic of the testing sessions. For each session, assessments before the stimulation were carried out in the following order: PPT, TMS, and EEG, while, after the stimulation, the same assessments were conducted in reverse order, ie EEG, TMS, and PPT. The order of the assessments was kept constant across participants for a practical reason. Although TMS and EEG are compatible methods, the presence of an EEG net on the participant’s head requires an increase in TMS intensity. Therefore, the EEG net was applied after the TMS pre-assessment and removed following the post-EEG assessment prior to the post-TMS assessment. The EEG nets remained on the participants’ heads during the tACS application, as the tACS electrodes could be applied without the need to remove the nets.

Data processing and statistical analysis

TMS data preprocessing

Corticospinal excitability was determined as an average peak-to-peak MEP in the TS trials of the FDI muscles from 20 to 100 ms post-TMS. The IHI ratio was determined as the mean peak-to-peak MEPs in CS-TS trials relative to the mean amplitude of MEP response to the single TS, ie IHI = MEP_CS-TS/MEP_TS. The IHI ratio is inversely related to interhemispheric inhibition: A decrease in the IHI ratio indicates increased inhibition, whereas an increase in the IHI ratio indicates reduced inhibition. As such, the “IHI ratio” refers to specific numerical values, while “IHI” describes interhemispheric inhibition in the conventional definition. The trials exceeding the root mean square EMG of 10 μV during the 40 ms immediately preceding the TMS pulse were excluded (Carson et al. 2004).

E‌EG data preprocessing

EEG data were preprocessed using the EEGLAB toolbox RELAX (the reduction of electroencephalographic artifacts) (Bailey et al. 2023) through the MATLAB environment (MathWorks, R2020a). RELAX is a data-cleaning pipeline that reduces vascular, ocular, and myogenic-induced artifacts to preserve neural signal readings. Using RELAX, the data were downsampled at 500 Hz. Then, a 50 Hz notch filter and a bandpass filter (1 to 80 Hz) were applied. Second, deficient EEG sensors and noisy time periods were removed. Third, eye blink and muscle movement-induced artifacts were removed using the multiple Wiener filtering method. The wavelet-enhanced independent component analysis removed artifacts not detected in the previous method. Last, once the removed EEG sensors and artifacts were interpolated, the data were divided into 2-s segments for statistical analysis.

For ImCoh calculation, the time series of each electrode were convolved with complex Morlet wavelets for frequencies between 4 and 90 Hz in 1 Hz increments (87 wavelet frequencies in total). The length of the wavelets started at 3 cycles for the lowest frequency and logarithmically increased as the frequencies increased, such that the length was 13 cycles for the highest frequency. This approach balances temporal and frequency precision (Cohen 1988). To minimize the effects of edge artifacts, analytic signals were obtained from time windows of 400 to 1,600 ms (at 20 ms intervals) within each 2,000 ms epoch. ImCoh values were computed for each electrode pair within the C3 (left M1) and C4 (right M1) using the following formula:

Here, |$i$| and |$j$| represented the time series of each electrode. For frequency |$f$|⁠, the cross-spectral density |${S}_{ij}(f)$| was taken from the complex conjugation of the complex Fourier transforms |${x}_i(f)$| and |${x}_j(f)$|⁠. Coherency was extracted by normalizing the cross-spectral density by the square root of the signals’ spectral power, |${S}_{ii}(f)$| and |${S}_{jj}(f)$|⁠. Then, the imaginary component of the resultant complex number was obtained. An estimate of M1-M1 ImCoh was obtained by averaging ImCoh values across all electrode pairs at each time point, frequency, and trial.

Statistical analysis

For a mixed-effects model, for each dependent variable, within-subject fixed factors of SESSION (tACS/sham) and TIME (prestimulation/poststimulation) were included in the model with by-subject intercept as random effects. For PPT data, an additional fixed factor of CONDITION (left, right, bimanual) was included in the model, while, for TMS IHI data, additional factors of ISI (10/40 ms) and DIRECTION (LR/RL) were considered in the model. For EEG power and TMS rMT and MEP_TS analyses, HEMISPHERE (L/R) was used instead of DIRECTION. A linear mixed-effect model (LMM) was used when the data followed a normal distribution curve, while a generalized linear mixed-effect model (GLMM) was used when the data deviated from a normal distribution. The assumptions for the G/LMMs, including linearity, homogeneity of variance, and normal distribution of the model’s residuals, were evaluated using the “DHARMa” package (Hartig 2024), which employs a simulation-based method to examine residuals for fitted G/LMMs. Null hypothesis significance testing for main and interaction effects was conducted with Wald chi-square tests for the GLMM analyses and F-tests for LMM analyses, and significant main and interaction effects were further investigated with Bonferroni-corrected contrasts. Standardized effect sizes were not reported for each fixed term in the G/LMM, following recommendations by Pek and Flora (2018). The partitioning of variance within the G/LMM makes obtaining a standardized effect size for each model term difficult. Instead, to facilitate the interpretation of G/LMM results, Cohen’s ds for follow-up contrasts were provided alongside significance levels. The critical P-value was set at 0.05.

For the control measures, the SCA and tES Sensation Questionnaire, Shapiro–Wilk tests were conducted to assess the normality of the data. For normally distributed variables, paired-sample t-tests were conducted. For non-normally distributed variables, Wilcoxon signed-rank tests were conducted.

Statistical analyses and visual illustrations were performed using R statistical package, version 4.4.1 (R Core Team 2024) with an integrated environment, RStudio version 2024.04.2 + 764 (RStudio Team, 2015) for Windows using “lmerTest” v3.1-3 (Kuznetsova et al. 2020) to fit the LMM and GLMM, “DHARMa” package (Hartig 2024) for LMM assumptions of linearity, homogeneity of variance and normal distribution of residuals “emmeans” v1.10.3 (Lenth 2024) for follow-up contrasts, “ggplot2” 3.5.1 package (Wickham et al. 2016) for graphical plots, “dplyr” v1.1.4 (Wickham et al. 2023) for data manipulation, “janitor” v2.2.0 (Firke 2023) for cleaning data, “here” v1.01 (Muller and Bryan 2020) for declaring paths, “knitr” v1.47 (Xie 2024) for report generation, “reader” v1.0.6 (Cooper 2017) for reading data files, “reshape2” v0.8.1 (Wickham 2007) for reshaping data, “skimr” v2.1.5 (Waring et al. 2022) for data summaries, and “stringr” v1.5.1 (Wickham 2023) for string operation wrappers. All data and codes are publicly available on the Open Science Framework: https://osf.io/exf54/.

Results

During the study, all participants tolerated the interventions without reporting discomfort, and no adverse events were experienced during experimental procedures. As shown in Table 1, paired-sampled t-tests indicated that participants were likely blinded to the conditions, as there were no significant differences in the tES sensation questionnaire responses across sessions. Similarly, no significant differences were observed across sessions in the components of the SCA questionnaire.

Behavioral measure

Bimanual coordination improves following bifocal tACS

Manual dexterity changes following tACS were assessed using three conditions in the PPT: left-hand only, right-hand only, and bimanual assembly. A GLMM revealed a significant main effect of CONDITION, χ²(2, n = 37) = 3939.27, P < 0.001, and a significant interaction of SESSION × TIME × CONDITION, χ²(2, n = 37) = 3.34, P = 0.04. Follow-up contrasts on the higher-order interaction revealed that bimanual dexterity improved significantly by 10.10% (4.32 parts more, ie more than one assembly) from pre- to poststimulation in the tACS session, z = 3.54, P < 0.01, d = 0.90, which is illustrated in Fig. 2, whereas this pattern was absent in the sham session, z = 0.05, P = 0.96, d = 0.01. Importantly, bimanual assembly performance at poststimulation was significantly greater in the tACS session than in the sham session, z = 2.18, P = 0.03, d = 0.56. In contrast, unimanual conditions showed no significant changes across time points for both tACS and sham sessions, |zs| = 0.92, ps > 0.36, |ds| < 0.23.

Neurophysiological measures

TMS parameters

Table 2 shows the rMT across sessions for the left and right hemispheres. An LMM revealed that there was a significant main effect of HEMISPHERE, χ² (1, n = 37) = 78.43, P < 0.001. The mean rMT for the right M1 was significantly higher than the left M1 in both sessions. These differences are likely explained by the coils used for each hemisphere. We used the same coil for each M1 across sessions and participants. Due to the stimulators’ position relative to the participant, a longer corded coil was required consistently for the farther side. In addition to the dominant left M1 generally having a lower motor threshold in right-handed participants, the coil used for right M1 stimulation appeared to produce a lower field intensity, leading to a higher rMT than that of the left M1. Since the stimulation intensities for TMS assessments, ie single-pulse TMS and IHI, were adjusted based on each individual’s rMT, discrepancies in rMT should not impact the interpretation of TMS results. The main effect of SESSION and the interaction of SESSION × HEMISPHERE, χ²(1, n = 37) = 0.01, P = 0.91, were not significant, suggesting the differences in rMT between hemispheres were constant across tACS and sham sessions.

Corticospinal excitability decreases following bifocal tACS

We investigated how corticospinal excitability changed following bifocal tACS to the bilateral M1s. As time-related changes, ie pre-tACS vs post-tACS, in corticospinal excitability were of primary interest, only those main effects and interactions involving TIME as a factor are described in detail.

A GLMM revealed significant main effects of TIME, χ²(1, n = 37) = 6.04, P = 0.01 and HEMISPHERE, χ²(1, n = 37) = 25.47, P < 0.01 and significant interactions of TIME × HEMISPHERE, χ²(1, n = 37) = 6.67, P < 0.01, and SESSION × HEMISPHERE, χ²(1, n = 37) = 6.90, P < 0.01. The follow-up contrasts for the significant interaction of TIME × HEMISPHERE revealed that at baseline, the left M1 (1.79 ± 1.38) showed significantly greater excitability than the right M1 (1.54 ± 1.17), z = −5.48, P < 0.001, d = −0.246. However, at post, the difference between the left M1 (1.56 ± 1.41) and right M1 (1.44 ± 1.15) became nonsignificant, z = −1.72, P = 0.89, d = −0.080. This change was primarily driven by a significant decline in excitability from pre- to poststimulation in the left M1, z = −3.18, P = 0.002, d = −0.302, while the right M1 showed no significant change, z = −1.44, P = 0.150, d = −0.136. These results suggested that, regardless of tACS or sham, the left M1 excitability showed a distinct change over time relative to the right M1 excitability. The follow-up contrast for the significant interaction between SESSION × HEMISPHERE revealed that, during sham, the left M1 exhibited significantly greater corticospinal excitability than the right M1, z = −5.52, P < 0.001, d = −0.247. During tACS, the left M1 and right M1 corticospinal excitability were not significantly different, z = −1.69, P = 0.09, d = −0.078. These results indicate that bifocal tACS had a more pronounced effect on the left M1, reducing its excitability to a level comparable with the right M1, which remained relatively stable throughout the intervention. The higher-order interactions, including TIME × SESSION, were not significant, χ²s < 2.92, ps > 0.06. This indicates that the corticospinal excitability change pre to post was comparable between tACS and sham sessions; this is illustrated in Fig. 3.

IHI is augmented by bifocal tACS

A GLMM assessing the IHI ratio revealed a significant main effect of DIRECTION, χ²(1, n = 37) = 11.10, P < 0.01. The IHI ratio was significantly higher for the left to right direction (0.79 ± 0.17) than right to left direction (0.72 ± 0.21), indicating that, at rest, the inhibitory projection from the left M1 to the right M1 was less prominent than from the right M1 to the left M1. This difference is likely due to the fact that in right-handed individuals, the left M1 is more active during ipsilateral hand movements and exhibits stronger inhibitory control over the right M1 rather than the right M1 over the left M1 (Ziemann and Hallett 2001; Shin et al. 2009).

The GLMM also revealed a significant interaction of SESSION × TIME, χ²(1, n = 37) = 5.53, P = 0.02. As shown in Fig. 4, follow-up contrasts revealed that during the tACS session, the IHI ratio significantly decreased (19.3% reduction) from prestimulation to poststimulation, z = −2.40, P = 0.02, d = −0.92, indicating increased interhemispheric inhibition between the left and right M1, while, in the sham session, IHI ratio did not change significantly, z = 0.19, P = 0.85, d = 0.04. In addition, at poststimulation, the IHI ratio was significantly lower in the tACS session compared to the sham session, z = −2.00, P = 0.045, d = −.81, while baseline IHI ratios were comparable between the tACS and sham session, z = 0.71, P = 0.48, d = 0.15). There was no significant interaction of SESSION × DIRECTION, χ²(1, n = 37) = 0.347, P = 0.55, suggesting that the effect of tACS on IHI was comparable for both directional projections.

E‌EG reveals beta power decreases following bifocal tACS

For the spectral power analysis of beta-power (20 Hz) changes at electrodes C3 (left M1) and C4 (right M1) following bifocal tACS, a GLMM revealed a significant interaction of SESSION × TIME, ×HEMISPHERE χ²(1, n = 37) = 4.41, P = 0.04. Follow-up contrasts revealed a decrease in beta power at poststimulation irrespective of the sessions in the left, z = −2.71, P < 0.01, d = −0.39, and right M1, z = −5.64, P < 0.01, d = −0.77, except for the right M1 after sham, which showed an increase from pre- to poststimulation, z = 0.51, P < 0.01, d = 0.21. During both tACS and sham sessions, the left M1 showed a decrease in beta power. While both tACS and sham sessions showed significant beta power decrease in the left M1, follow-up contrasts revealed that the magnitudes of decrease in beta power were significantly greater in the tACS session relative to the sham session irrespective of the hemispheres, z = −2.71, P < 0.01, d = −0.37.

Functional connectivity indexed by ImCoh did not change following bifocal tACS

We investigated whether functional connectivity changed by assessing endogenous oscillatory phase synchronization at 20 Hz following bifocal tACS to the bilateral M1 regions during resting-state EEG. Figure 5 illustrates the ImCoh values at pre- and poststimulation for tACS and sham sessions. For resting-state ImCoh at 20 Hz, there was a significant main effect of TIME χ²(1, n = 37) = 5.54, P = 0.02, indicating that the ImCoh value decreased from pre- to poststimulation in both sessions. The main effect of SESSION, χ²(1, n = 37) = 1.48, P = 0.22 and the interaction SESSION × TIME were not significant, χ²(1, n = 37) = 2.86, P = 0.09. In sum, irrespective of the session, interhemispheric synchrony decreased following stimulation.

Discussion

The present study sought to elucidate the neural mechanisms that underlie functional connectivity modulation induced by bifocal tACS to the bilateral M1s. We found that bifocal tACS delivered at 20 Hz for 20 min to the bilateral M1s induced changes in functional connectivity and behavior, observed through increased IHI and improved bimanual dexterity. This effect contrasts with the lack of modulation observed in oscillatory-based connectivity measures, highlighting the distinct mechanisms underlying the poststimulation effects of bifocal tACS. The lack of functional connectivity changes in oscillatory measure, ie ImCoh, suggests that bifocal tACS applied to the bilateral M1s has no impact on poststimulation measures of oscillatory-based functional connectivity. The observed manual dexterity improvements also indicate that applying bifocal tACS over the bilateral M1s might have enhanced the functional connectivity of the circuits underlying bimanual coordination rather than enhancing global motor function, as we did not observe significant changes in the unimanual tasks. The observed increase in IHI indicated that bifocal tACS applied over the bilateral M1s specifically enhanced the interhemispheric pathways connecting these distant cortical regions, potentially facilitating the improvement of bimanual coordination.

Bifocal tACS improved interhemispheric inhibition and corticospinal excitability declined

Our findings showed that IHI increased following beta-tACS to the bilateral M1s, without a corresponding increase in corticospinal excitability but rather a decrease. Previous studies investigating IHI during or following beta frequency tACS application to unifocal in sensorimotor areas found no change in inhibition (Rjosk et al. 2016; Lafleur et al. 2020). Our findings align with previous research, which found that bifocal tACS induced changes and enhanced functional connectivity (Violante et al. 2017; Reinhart and Nguyen 2019; Grover et al. 2022; Meng et al. 2023). Beta bifocal tACS has also been linked to improvements in response inhibition tasks. Fujiyama et al. (2023) found that bifocal tACS over the rIFG and presupplementary motor area enhanced beta-band functional connectivity. Thus, in addition to the previous studies finding modulations of EEG-based functional connectivity, our study demonstrated that bifocal tACS applied over the bilateral M1 can improve interhemispheric communication, demonstrating connectivity was enhanced between the left and right M1.

It is important to note that, in the current study, no increases in corticospinal excitability were observed, indicating that our observation of IHI modulation was driven by the specific effect of bifocal tACS on interhemispheric pathways rather than an increase in M1 excitability. A meta-analysis by Wischnewski et al. (2019) found that unifocal beta-frequency tACS consistently increased corticospinal excitability when applied at a current intensity above 1.0 mA. In the present study, bifocal application diverges from this consensus. The IHI value can decrease independently of changes in interhemispheric interactions between the left and right M1 if corticospinal excitability increases, as the IHI ratio is calculated as MEP_CS-TS/MEP_TS. However, we observed a decrease in the IHI ratio despite the overall decline in corticospinal excitability following both tACS and sham sessions. Therefore, the observed decrease in the IHI ratio reflects an increase in interhemispheric inhibition and suggests genuine augmentation in IHI between the left and right M1. Previous research indicates that interhemispheric inhibition can be modulated independently of corticospinal excitability, suggesting distinct neural pathways are involved (Ferbert et al. 1992). Similarly in the context of tACS-induced modulations in the motor system, Nowak et al. (2017) found no changes in corticospinal excitability following unifocal tACS applied over M1 yet observed a decrease in intracortical inhibition. Evidence suggests a competitive relationship between IHI and intracortical inhibition, where an increase in one may overshadow the other (Daskalakis et al. 2002). Therefore, the increase in IHI measured with TMS in the current study may relate to the decrease in intracortical inhibition noted by Nowak et al. (2017). Notably, bifocal tACS may specifically enhance interhemispheric pathways, potentially accounting for changes in IHI without affecting overall corticospinal output. Alternatively, the observed changes in IHI, along with the absence of tACS-specific effects on corticospinal excitability, may be attributed to the current intensity used in the current study, ie 1 mA. Wischnewski et al. (2019) suggested that tACS induces robust changes in corticospinal excitability only when the current intensity exceeds 1.0 mA. Future studies should aim to clarify these mechanisms by including IHI, corticospinal excitability, and intracortical inhibition measurements in a single study.

In our investigation of how bifocal tACS may modulate different interhemispheric circuits through stimulation, we found no significant differences in modulations between SIHI and LIHI. Interhemispheric inhibitory interactions are mediated through transcallosal pathways and local inhibitory neurons (Daskalakis et al. 2002). It has been suggested that SIHI and LIHI are modulated by GABA_A and GABA_B-ergic circuits, respectively (Irlbacher et al. 2007). Therefore, this lack of significant difference in modulations between SIHI and LIHI implies that while bifocal tACS over the left and right M1 engages interhemispheric pathways, it does not appear to selectively influence either GABA_A- and GABA_B-ergic circuits. Previous studies examining the effect of unifocal tACS over M1 on intracortical inhibition found gamma frequency tACS induced greater significant changes in GABA_A activity than beta-tACS both during stimulation (Nowak et al. 2017) and following stimulation (Guerra et al. 2018). Nowak et al. (2017) proposed that although beta-frequencies are the predominant oscillations in the sensorimotor cortex, inhibitory neurons may differ and are intrinsically responsive to gamma-band synchrony. Thus, beta-frequency bifocal tACS might not be selective in modulating inhibitory pathways mediated by GABA_A and GABA_B-ergic, instead producing a broader or generalized effect across the sensorimotor circuits.

Bifocal tACS had no impact on oscillatory synchrony

Previous research suggests that bifocal tACS is capable of synchronizing oscillatory activity across distant cortical regions, as the oscillatory synchronization of the neural populations in distinct cortical regions is considered a key mechanism for the modulation of functional connectivity (Polanía et al. 2012; Thut et al. 2012). Previous studies found bifocal tACS coupled oscillatory activity at two cortical sites poststimulation at theta frequencies (Polanía et al. 2012; Violante et al. 2017; Reinhart and Nguyen 2019), alpha frequencies (Zaehle et al. 2010; Helfrich et al. 2014), and beta frequencies (Fujiyama et al. 2023; Meng et al. 2023). However, the present results diverged from these observations, as no poststimulation increases were found in the oscillatory functional connectivity measure (ImCoh). Interestingly, when tACS was applied during task performance (online stimulation), oscillatory increases using the same EEG measure, ie ImCoh, were observed (Helfrich et al. 2014; Fujiyama et al. 2023), suggesting that entrainment may primarily underlie online stimulation effects. These findings collectively indicate that beta-frequency bifocal tACS applied at rest may induce neuroplastic changes in motor cortical areas through mechanisms distinct from direct oscillatory synchronization. Another perspective is that the effects of bifocal tACS are region-specific and that the heterogeneity of the underlying circuits within sensorimotor areas needs to be considered. Previous evidence indicated phase synchrony was induced by beta-tACs between the rIFG-presupplementary motor area (Meng et al. 2023) and rIFG-M1 (Fujiyama et al. 2023). The rIFG is implicated in response inhibition studies as it is a key component of the inhibitory control network (Aron et al. 2014). Considering beta oscillations are critical for response inhibition (Swann et al. 2009; Wagner et al. 2015), the rIFG may be more susceptible to beta synchronization than other cortical areas due to its role in response inhibition.

While we found no modulation in ImCoh following bifocal tACS, we observed a significant decrease in beta power at poststimulation, suggesting that bifocal tACS interacts with ongoing neural activity. A decrease in beta power has been associated with the activation of the sensorimotor network, particularly in preparation for motor movement (eg Kilavik et al. 2013). By integrating these findings with the observed improvement in bimanual dexterity, we may infer that beta-frequency tACS may prime cortical areas in a preparatory state for movement as suboptimal preparation likely causes inappropriate execution (Hallett 2000).

Possible explanations as to why beta power showed a decrease following beta-tACS are that the stimulation may have disrupted existing beta oscillations through phase-coupling effects, leading to a reduction in beta power. Additionally, beta-tACS could have modulated excitability in the targeted brain regions, altering the balance between excitatory and inhibitory activity and possibly contributing to the decrease in beta power. The level of beta power can also depend on the difficulty of the task at hand. Bootsma et al. (2021) discovered that beta power decreased during challenging motor tasks compared to medium- and low-difficulty tasks, indicating a compensatory mechanism to meet increased motor demands. Similarly, Engel and Fries (2010) proposed that in situations where external, bottom–up factors drive behavioral responses, beta power tends to decrease. These findings emphasize the intricacy of relationships between oscillatory activity and task demands. The decrease in beta power following bifocal tACS may indicate a shift towards a more adaptable cortical state that supports movement preparation.

We also noted that the decrease in beta power mirrored the decline in corticospinal excitability following tACS. This may suggest a potential coupling between these two neural processes, where reductions in beta oscillations are associated with diminished corticospinal output. A previous study (Hussain et al. 2019) indicated that beta rhythm events (15 to 30 Hz) in the sensorimotor cortex are positively correlated with corticospinal excitability; specifically, higher beta power is linked to larger MEPs. Moreover, fluctuations in EEG oscillations can selectively influence corticospinal excitability, with some studies reporting correlations between prestimulation beta power and MEP amplitudes (Hussain et al. 2019). However, Wischnewski et al. (2019) found that while the phase of beta oscillations differentially modulates corticospinal excitability, beta power itself does not exhibit the same relationship, which is consistent with the findings by Schilberg et al. (2021). Overall, these insights underscore the complex interplay between beta oscillations and corticospinal excitability, suggesting that both power and phase may play distinct roles in motor control.

While the present results are inconsistent with the expectation that beta power would increase following tACS (Thut et al. 2012), a more nuanced perspective emerges when considering alternative contexts. For example, in people with Parkinson’s disease (PD), reductions in the synchronized activity of beta frequencies during rest were associated with clinical improvements in motor functioning (George et al. 2013). The same study demonstrated that higher beta power in PD patients was associated with more severe motor dysfunction. Similarly, Brittain et al. (2014) found that increased beta power in the motor regions of the basal ganglia hindered motor processing in individuals with PD, leading to typical motor impairments such as slowed movements and rigidity. Clinically, reducing this heightened beta activity has been shown to improve symptoms like akinesia and rigidity while lessening tremors (Tinkhauser et al. 2017). However, the present results were derived from a nonclinical population, so we may only speculate on the generalizability of bifocal tACS-related power modulations in therapeutic settings.

Bifocal tACS improved bimanual coordination

We found that bimanual dexterity significantly improved following beta frequency bifocal tACS over bilateral M1. The current study is the first to assess manual dexterity in relation to beta bifocal tACS. The observed improvement in behavior following bifocal tACS aligns largely with previous studies. For example, Meng et al. (2023) found that bifocal tACS at beta frequencies targeting the rIFG and M1 enhanced response inhibition performance. The authors suggested that this improvement in behaviors reflected the neuroplastic reorganization of brain networks, indicating that, as suggested by the effect of bifocal tACS on functional connectivity, it can be broadly transmitted throughout the brain to modulate behavior.

Improvement in dexterity was only observed in the bimanual assembly task, indicating that improvement was unrelated to the shared properties involved in both unimanual and bimanual assembly, including individual hand movement speed or acuity, but rather a performance improvement in the coordination between both hands. The complexity of bimanual tasks, which necessitate effective communication and coordination between brain regions, may be susceptible to the effect of tACS. These higher-level tasks rely heavily on efficient neural communication and regional coordination (Schoenfeld et al. 2021), suggesting that tACS—which has been demonstrated to enhance connectivity between brain regions—would be more beneficial for such tasks (Schoenfeld et al. 2021). Furthermore, bimanual tasks engage the left and right M1 more intensively than unimanual tasks, potentially fostering greater neural plasticity due to the increased effort required for movement coordination (Feurra et al. 2011). The behavioral task implemented in the current study was not a learning task, as demonstrated by Schoenfeld et al. (2021); as such we can only suggest that greater plastic effects may arise from the interhemispheric connectivity changes induced by bifocal tACS over bilateral M1s. Consequently, the substantial improvement in bimanual motor performance observed in the current study highlights the potential of M1 beta frequency bifocal tACS to enhance bimanual coordination by improving interhemispheric connectivity.

Limitations and future perspectives

While this study has offered valuable insights into the mechanism underlying bifocal tACS, certain limitations must be considered. Individual variability is crucial in discussions of cortical stimulation as it is scientifically unsound to assume that all brains are identical in function and structure (Wansbrough et al. 2024). As such, it may explain the lack of changes in ImCoh following tACS. To minimize individual variability, employing a closed-loop approach is an emerging trend in tES, which involves dynamically adjusting stimulation parameters based on real-time brain activity rather than predetermining parameters (Stecher et al. 2021; Wansbrough et al. 2024). For example, when determining electrode placement or stimulation voltage, structural or functional magnetic resonance imaging (MRI) can be incorporated to locate the optimal electrode placement for the target cortical area (Jog et al. 2021). Individualizing tES frequency is another way to tailor brain stimulation. Frequency ranges are well established as default within specific cortical regions, and, when stimulation is adjusted to an individual peak neural oscillatory frequency, neurophysiological changes have been observed (Salamanca-Giron et al. 2021; Aktürk et al. 2022; Kudo et al. 2022; Riddle et al. 2022). Tailoring stimulation creates individualized tES protocols that are proposed as the most appropriate method for ensuring consistency across age, genetics, and existing neural networks and increases the reliability of tES interventions (Wansbrough et al. 2024). In the current study, we employed e-field modeling using a standard head model, which indicated that the protocol used effectively stimulated the left and right M1 with an adequate electrical field (refer to Fig. 1A). Future research would benefit from developing individualized protocols that utilize individual participants’ structural MRI scans to fully harness the potential of bifocal tACS.

The temporal constraints of our assessment methodology have limited the scope of our findings, specifically the timing and sequencing of assessments before and after stimulation, as shown in Fig. 1B. Post-stimulation assessments were conducted between 10 and 20 min of reverse sequencing from the prestimulation assessments. Conducting testing this way limited the potential to observe oscillatory changes that may take longer to manifest, but we also lacked insights into the lasting implications of bifocal tACS. We included a 7-day washout period to account for lasting effects; however, no follow-up assessments were conducted to determine how long the effects of bifocal tACS induced lasted. Research has shown that the effects of tES persist beyond application, with alpha-tACS effects lasting up to 70 min (Kasten and Herrmann 2017). Understanding the lasting effects of tACS will inform the longevity of the induced neuroplastic or behavioral changes and provide valuable insight into how tACS is therapeutically beneficial in clinical settings.

In the current study, we examined the effect of tACS applied over bilateral primary motor cortices with 0° in-phase stimulation. However, the lack of a control condition using a 180° antiphase condition limits definitive conclusions regarding the phase-specific effects of the stimulation. Future studies would benefit from including both 0° in-phase and 180° antiphase conditions in a single study, as this would allow clearer conclusions regarding the synchronization effects between targeted brain regions and implications for physiological and behavioral outcomes.

Another limitation of the current study is the potential confounding effects of retinal (Fiene et al. 2022) or somatosensory (Asamoah et al. 2019) costimulation by the tACS protocol used in the study. Given there was no control condition in the current study, future research is warranted to incorporate active control conditions that isolate stimulation from either the retinal or somatosensory system to better assess their specific contributions.

Lastly, while the current study was sufficiently powered to detect pre- and poststimulation changes in neurophysiological measures, future studies are warranted to increase the number of participants. Our exploratory analysis examining the relationship between changes in the IHI ratio and changes in bimanual performance in the PPT revealed a small, nonsignificant positive correlation (Pearson’s r = 0.292, 95% CI [−0.08, 0.59]). Although this correlation was not statistically significant, it does not rule out the possibility of a relationship between modulation in IHI and improvements in bimanual performance in the PPT. Future studies could facilitate correlational analyses to investigate further whether changes in bimanual performance are associated with alterations in interhemispheric inhibition following bifocal tACS.

Conclusion

The current study investigated the neurophysiological mechanisms by which concurrently applied tACS induces changes in functional connectivity between the bilateral M1s. Our findings indicated that tACS enhanced functional connectivity between the left and right M1 through IHI, likely through neuroplasticity mechanisms rather than lasting entrainment effects. Further, we observed improvements in bimanual coordination, while unimanual dexterity remained unchanged. These findings highlighted tACS as a potential therapeutic tool for neurological conditions marked by dysfunctional cortical connectivity, extending its application beyond the motor system to other cortical regions. Future research may benefit from exploring individualized tACS protocols to fully optimize the effectiveness of bifocal tACS. Such effort would facilitate the development of intervention protocols for clinical populations such as PD.

Author contributions

Brooke Lebihan (Data curation, Formal analysis, Investigation, Writing—original draft), Lauren Mobers (Data curation, Formal analysis, Investigation, Writing—review & editing), Shannae Daley (Data curation, Formal analysis, Investigation, Writing—review & editing), Ruth Battle (Data curation, Investigation, Writing—original draft), Natasia Leclercq (Data curation, Formal analysis, Investigation, Writing—review & editing), Katherine Misic (Data curation, Formal analysis, Investigation, Writing—review & editing), Kym Wansbrough (Formal analysis, Methodology, Visualization, Writing—review & editing), Ann-Maree Vallence (Supervision, Writing—review & editing), Alex Tang (Writing—review & editing), Michael Nitsche (Conceptualization, Writing—review & editing), and Hakuei Fujiyama (Conceptualization, Formal analysis, Methodology, Project administration, Supervision, Visualization, Writing—original draft, Writing—review & editing).

Funding

This work was supported by the Neurotrauma Research Program (DoH20193370) awarded to H.F. and A.D.T. A.D.T. was supported by a Sarich Family Research Fellowship.

Conflict of interest statement: None declared.

References