Effects of anodal transcranial direct current stimulation on motor evoked potentials variability in humans

Abstract Motor evoked potentials (MEPs) obtained from transcranial magnetic stimulation (TMS) allow corticospinal excitability (CSE) to be measured in the human primary motor cortex (M1). CSE responses to transcranial direct current stimulation (tDCS) protocols are highly variable. Here, we tested the reproducibility and reliability of individual MEPs following a common anodal tDCS protocol. In this study, 32 healthy subjects received anodal tDCS stimulation over the left M1 for three durations (tDCS‐T5, tDCS‐T10, and tDCS‐T20 min) on separate days in a crossover‐randomized order. After the resting motor threshold (RMT) was determined for the contralateral first dorsal interosseous muscle, 15 single pulses 4–8 sec apart at an intensity of 120% RMT were delivered to the left M1 to determine the baseline MEP amplitude at T0, T5, T10, T20, T30, T40, T50, and T60 min after stimulation for each durations. During TMS delivery, 3D images of the participant's cortex and hot spot were visualized for obtaining MEPs from same position. Our findings revealed that there was a significant MEPs improvement at T0 (P = 0.01) after 10 min of anodal stimulation. After the 20‐min stimulation duration, MEPs differed specifically at T0, T5, T30 min (P < 0.05). This indicates that tDCS is a promising tool to improve MEPs. Our observed variability in response to the tDCS protocol is consistent with other noninvasive brain stimulation studies.


Introduction
Early in its development, transcranial direct current stimulation (tDCS) is an ambitious method in clinical and cognitive neuroscience to modulate neuroplasticity. One of the commonly used methods delivers tDCS at an intensity of 1-2 mA (0.029-0.057 mA/cm 2 ) through pad electrodes that are placed on the scalp with a current that flows from the anodal to cathodal electrode (Nitsche and Paulus, 2000;Nitsche and Paulus, 2001;Stagg and Nitsche, 2011). tDCS has significantly developed with more than 1500 research articles published on the topic in the last 10 years (Nitsche and Paulus, 2000;Nitsche and Paulus, 2001;Stagg and Nitsche, 2011;Cappon et al., 2016;Woods et al., 2016). Recent studies have shown that tDCS methods improve or disrupt cognitive functions and help in improving a range of neurological and psychiatric impairments (Kuo and Nitsche, 2012;Fl€ oel, 2014;Kuo et al., 2014;Christians et al., 2016;Shin et al., 2015;Antonenko et al., 2016;Bikson et al., 2016). tDCS has been widely used to modulate motor cortical excitability noninvasively by measuring the amplitude of motor evoked potentials (MEP) induced by transcranial magnetic stimulation (TMS) (Nitsche and Paulus, 2000;Nitsche and Paulus, 2001;Salvador et al., 2012;Moliadze et al., 2014;Tazoe et al., 2014;Vaseghi et al., 2015;Labruna et al., 2016).
The goal of this study was to better understand anodal tDCS-response variability in a crossover design using 1.5 mA for three durations (5, 10, and 20 min) in healthy na€ ıve individuals for brain stimulation methods. In this study, we investigate the effects of anodal tDCS on motor cortical plasticity, as measured by amplitude changes from TMS-induced MEP.

Methods
Thirty-two healthy subjects (age range: "19-63" years old) participated in this study. They were all right-handed, as assessed by the Edinburgh Handedness Inventory (right-handedness 1.97 AE 0.06) (Oldfield, 1971). We used a single-blinded, crossover and counterbalanced design. Subjects participated in three sessions of anodal stimulation for each experimental condition of for three different durations (tDCS T5, tDCS T10, and tDCS T20 min). Each participant attended all sessions, which started at the same time of the day, and were separated by at least 7 days to avoid cumulative increases in cortical excitability. They were naive to tDCS and exhibited normal cognitive status, as indexed by Mini Mental State Examination (MMSE) (Folstein et al., 1975) scores (normal range: 28-30). Furthermore, neurological examination of the subjects revealed no abnormal signs that could suggest any underlying neurological or psychological conditions. None of the participants took any medication known to affect motor cortical excitability at the time of the study and had no contraindications for tDCS or consume caffeine prior to each session (Ferraroni et al., 2007;Fragni, 2008) and TMS (Rossi et al., 2009). Demographic and clinical features of the subjects are shown in Table 1. The investigation was carried out in accordance with the most recent version of the Declaration of Helsinki and was approved by the local review board (King Saud University). All participants gave written informed consent prior to enrollment in the study.

Experimental setup
In order to test the effects of tDCS on MEPs, subjects underwent a structural MRI scan and then received the neuronavigated TMS protocol so that TMS-induced MEP could be obtained before and after tDCS. To test the hypothesis that there is a nonlinear modulatory effect depending on stimulation duration, we assessed tDCS-induced changes in CSE before and after three tDCS sessions varying in duration (tDCS T5, tDCS T10, and tDCS T20 min).
For the TMS protocol, we first determined the individual resting motor threshold (RMT) at each session based on the recommendations from the International Federation for Clinical Neurophysiology. RMT was defined as the lowest stimulator output intensity that produced at least five MEP out of 10 consecutive pulses of at least 50 lV peak-to-peak amplitudes on the EMG. Active electrodes were attached to the skin on top of the right FDI muscle to collect MEP. We then collected MEP from the FDI before and after tDCS delivery in the same way: we delivered 15 single TMS pulses 4-8 sec apart at an intensity of 120% the subject's RMT to the hot spot. During the measurements, the subjects sat in a comfortable recliner and held their hands supine on their laps. They were asked to remain silent during the study to avoid speechinduced modulation of cortical excitability. FDI muscle relaxation was controlled by continuous visual and audio EMG monitoring during all experiments. The TMS system delivered trigger pulses that synchronized the TMS and EMG systems. The EMG signals were filtered (8À500 Hz), amplified, displayed, and stored for off-line analysis. Neuronavigation was used to determine the RMT and a single-pulse TMS protocol was used before and after each tDCS session to ensure that the same brain area in a single subject was targeted across sessions.

tDCS stimulation
Stimulation was delivered through a constant current with two 35 cm 2 (5 9 7 cm) saline-soaked sponges (Soterix Medical 1 9 1 device). The anode was positioned above the motor cortical representational area of the right FDI, as revealed by the neuronavigated TMS. The cathode was placed above the contralateral supraorbital cortex. The constant current flow was monitored by a voltmeter. In separate sessions, tDCS was delivered for 5, 10, or 20 min, which are well within current safety limits (Fragni, 2008).
Before each tDCS session, 15 MEPs were collected as baseline measurements. After each tDCS session, 15 MEPs were collected at eight epochs: T 0 , T 5 , T 10 , T 20 , T 30 , T 40 , T 50 , and T 60 min. The participant's subjective ratings of pain and discomfort were collected at the end of every tDCS session as a safety measure; a protocol was in place if pain or discomfort was reported after each session.

Data analysis
In order to assess if cortical thickness could explain the differences in plasticity measured by MEPs we performed a correlation analysis between subjects' MEPs and the values of cortical thickness of a region of interest (ROI) (Fig. 1). This particular ROI was drawn to encompass the motor strip of both hemispheres of all subjects. The ROI was done first in standard space-brain and then was mapped for each subject in order to measure this region specific for each individual.
For MEP data, a continuous EMG signal was recorded and sampled for 350 msec epochs, 50 msec before, and 300 msec after each TMS pulse. CSE was assessed by measuring the peak-to-peak amplitude of MEPs recorded from the contralateral FDI muscle in response to single TMS pulses applied over M1. To minimize the variability of TMS-induced single-pulse responses, the largest and the smallest MEP amplitude responses from each recording were excluded from analysis. For all subsequent analyses, we thus computed changes in MEP amplitude (%Δ) from the baseline to post-tDCS measures for each epoch.
Data were analyzed using SPSS (IBM Corp. Released 2012. IBM SPSS Statistics for Windows, Version 21.0. Armonk, NY: IBM Corp.). Cortical excitability changes were expressed as increase or decrease in mean MEP amplitudes before and after stimulation. To test the time course of plasticity changes, a repeated measure-ANOVA (9 9 3) with the factors "time course" (baseline, T 0 , T 5 , T 10 , T 20 , T 30 , T 40 , T 50 , and T 60 min) and "stimulation duration" (tDCS-T5, tDCS-T10, and tDCS-T20 min). When appropriate, that is, significant interactions in the repeated measure-ANOVA, Student's ttests (paired, two-tailed) were performed to determine more specifically whether MEP amplitudes differed before and after plasticity induction within and between conditions. In cases of lacking interactions, no further t-tests were conducted. In the linear models, sphericity was tested with Mauchly's test and, if necessary (Mauchly's test < 0.05), the Greenhouse-Geisser correction was used. Data in tables are presented as mean AE standard deviation. Furthermore, grand average analysis was also conducted to examine the percentage of "responders" (favorable MEP increase after anodal tDCS stimulation) and "non-responders" using the mean grand average poststimulation criterion. Subjects with grand averages > 1 were classified as "non-responders" and subjects with grand averages < 1 were classified as "responders". In all figures, error bars refer to the standard error and graphs show untransformed data.

Results
Subjects tolerated all of the experimental stimulation protocols well. Some subjects reported an itching/tingling sensation during the beginning of tDCS, but this faded away after a few minutes. In a few subjects, we observed reddening of the skin under the scalp electrodes; however, this did not persist for longer than 60 min. No other side effects were reported. There were no differences in gender, EHI handedness scores, or MMSE scores among subjects (Table 1).
At baseline, moreover, the amplitude of motor evoked potentials (MEP) elicited from the right hand was not significant for all condition (P = 0.32, Table 1).
Furthermore, grand average analysis was also conducted to examine the percentage of "responders" as shown in Figure 6; majority of the subjects (60%) were responders, whereas 40% are the nonresponders tDCS-T20 min.

Correlational analyses
Spearman's correlation coefficients were used to examine the relationship between relative baseline values (cortical thickness measured as precentral gyrus, surface area, gray volume and average gray surface area for both left and right hemisphere Table 2). These analyses revealed a positive trend-level correlation between tDCS-T10 group and Right Gray volume (r = 0.335, P = 0.046), Right Thickness average (r = 0.360, P = 0.043), and relative mean poststimulation MEPs. In addition, a negative trend-level correlation is revealed between tDCS-T20 group and Left Surface area (r = À0.387, P = 0.028) and relative mean poststimulation MEPs. Concerning all other variables including the tDCS-T5 group, no significant correlations were observed in subjects (P > 0.0.05).

Discussion
The main goal of this work was to assess the reproducibility of tDCS-induced effects on MEPs to measure CSE. We reproduced at the group level the classical anodal DCS (tDCS-T20 min) effect represented by an increase of MEP amplitude after applying the stimulation with current intensity of 1.5 mA for the T 0 , T 5 , T 30 min compared to prevalues. Our findings demonstrate that CSE responses from the short duration of stimulation (tDCS-T5 and tDCS-T10) are highly variable. At least 60% of the CSE responses were nonetheless excitatory following the stimulation.
These findings are consistent with Chew et al. (2015), who reported that between two sessions, there is a low testretest reliability of tDCS on TMS-induced MEP amplitude. Specifically, the first session revealed a main increase of MEP amplitude compared to baseline, whereas the second session did not show significant changes in amplitude (Chew et al., 2015). However, those results differ from L opez-Alonso et al. (2015) who reported a moderate testretest reliability between two sessions of tDCS. Discrepancies between these studies might be partially explained by the number of sessions. The authors interpreted the variability of the MEPs as a potential cause for their findings. We compared three tDCS sessions, whereas L opez-Alonso et al. (2015)) compared only two sessions. The TMS coil position and orientation by not using stereotactic system have been shown to impact MEP amplitudes (Guggisberg et al., 2001;Julkunen et al., 2009;Kidgell et al., 2013). There might be several methodological differences compared to previously stated studies. For instance, it is known that MEP amplitude changes induced by tDCS is affected by the size of electrodes (Nitsche et al., 2007;Datta et al., 2009;Chew et al., 2015;L opez-Alonso et al., 2015) and by the duration of stimulation (Nitsche and Paulus, 2001;Nuzum et al., 2016;Woods et al., 2016). On the other side, the different time-periods between sessions (6-12 months (Moliadze et al., 2014); 1-7 weeks (Chew et al., 2015); 2-9 days (Parkin et al., 2015); 3-4 days (Inukai et al., 2016); may also have an impact on ultimate results. 2019 | Vol. 7 | Iss. 13 | e14087 Page 5 There was no baseline MEP difference between the three stimulation duration conditions. MEP size significantly differed between the longest tDCS condition (20 min) as compared to 5 and 10 min stimulation. Stimulation parameters, such as duration and intensity, as well as the electrode montage, likely interact with one another, possibly resulting in nonlinear effects on CSE (Nitsche et al., 2007). With regard to the electrode montage, the present study employed the conventionally used "M1-contralateral superior frontal orbit" arrangement, with an enlarged reference electrode (35 cm 2 ). We chose this montage because it was previously shown to reduce unwanted physiological effects under the reference electrode, at least up to a 2.0 mA setting (Datta et al., 2009). Further studies are required to compare the conventional montage with other montages that use multiple small electrodes in concentric ring arrangements, as these have been shown to induce more focused electric fields and also result in slightly enhanced motor cortical excitability (Ridding and Ziemann, 2010;Kuo et al., 2013). The present study collected MEP size over three sessions and a longer period of monitoring (60 min), as compared to previous studies that mainly assessed retest reliability over two sessions (Kidgell et al., 2013;Chew et al., 2015;L opez-Alonso et al., 2015;Hsu et al., 2016). We classified group response (increase) or nonresponder (decrease or no change) based on prior research work on cluster (Roche et al., 2011;Horvath et al., 2015;Parkin et al., 2015), or choosing an arbitrary value (Roche et al., 2011;Tremblay et al., 2013;Inukai et al., 2016).
Consequently, only the 1.5-mA condition tDCS-T20 min showed a clear and consistent response pattern when performing individual tracking of tDCS responses across sessions (Fig. 5).

Methodological considerations, limitations, and future directions
The TMS pulse intensity at 120% RMT was used to obtain MEP which is in accordance to previous studies

tDCS-20 tDCS-10 tDCS_5
Responder 67% Responder 54% Responder 60% Non-responder 46% Non-responder 40% Non-responder 33% Figure 6. The grand average value expressed as the percentage of "responders" (favorable MEP increase after anodal tic's stimulation) and "non-responders" using the mean grand average poststimulation criterion. Subjects with grand averages > 1 were classified as "nonresponders" and subjects with grand averages < 1 were classified as "responders" for tDCS-T5 , tDCS-T10, and tDCS-T20 min.  Brasil-Neto et al., 1992;Lang et al., 2004;Madhavan and Stinear, 2010;Scelzo et al., 2011;Di Lazzaro et al., 2012;Suzuki et al., 2012;Pellicciari et al., 2013;Teo and Chew, 2014). Both human and animal studies propose that the use of 120% RMT boosts the reliability of MEP measurements (Miyaguchi et al., 2013).Use of navigation system for obtaining the reliable MEP across the subject may have affected our outcome and masked or reduced any possible tDCS effect. An important limitation of this study is that use of 15 MEPs per batch to establish average MEP values, which is not uncommon in tDCS-induced MEP studies (Pitcher et al., 2003;Lang et al., 2004;Quartarone et al., 2004;Power et al., 2006;Datta et al., 2009;Bastani and Jaberzadeh, 2012;Suzuki et al., 2012;Pellicciari et al., 2013;Chang et al., 2016) and this number has been proved to elicit highly reliable response patterns of MEPs per batch. In the current study, there was no significant effect of gender and is consistent with Pitcher et al. (Cuypers et al., 2014), who did not find a main effect of gender on MEP variation when exploring TMS recruitment curve characteristics. The major limitations of the study are MEP batch, sample size of subject and tDCS intensity, and type of montage (Pitcher et al., 2003;Quartarone et al., 2004;Power et al., 2006;Bastani and Jaberzadeh, 2012;Suzuki et al., 2012;Pellicciari et al., 2013;Chang et al., 2016). We used navigation system to control technical factors including coil position or orientation to reduce the variability of MEPs. We monitored subject for physiological factors by background muscle activity. However, researchers using tDCS should consider high inter-and intraindividual variability of MEPs induced by coil, number of trials, the intertrial interval (ITI), and stimulus intensity as well as physiological and psychological factors such as attention and muscle fatigue.
We need to be careful in how we interpret with extrapolating our findings to other populations. A different number of consecutive stimuli might be required to estimate CSE in the elderly or in a population with neurodegeneration, as this data was obtained in healthy young subjects (Magistris et al., 1999).

Availability of Data and Material
All data are included in the manuscript. However, the datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Conflict of Interest
The author(s) declare(s) that there is no conflict of interest regarding the publication of this paper.

Consent for Publication
Consent was obtained from all participants before start of the procedure.