Systematic evaluation of the impact of stimulation intensity on neuroplastic after-effects induced by transcranial direct current stimulation

Abstract

Key points: Applications of transcranial direct current stimulation to modulate human neuroplasticity have increased in research and clinical settings. However, the need for longer-lasting effects, combined with marked inter-individual variability, necessitates a deeper understanding of the relationship between stimulation parameters and physiological effects. We systematically investigated the full DC intensity range (0.5-2.0 mA) for both anodal and cathodal tDCS in a sham-controlled repeated measures design, monitoring changes in motor-cortical excitability via transcranial magnetic stimulation up to 2 h after stimulation. For both tDCS polarities, the excitability after-effects did not linearly correlate with increasing DC intensity; effects of lower intensities (0.5, 1.0 mA) showed equal, if not greater effects in motor-cortical excitability. Further, while intra-individual responses showed good reliability, inter-individual sensitivity to TMS accounted for a modest percentage of the variance in the early after-effects of 1.0 mA anodal tDCS, which may be of practical relevance for future optimizations.

Abstract: Contemporary non-invasive neuromodulatory techniques, such as transcranial direct current stimulation (tDCS), have shown promising potential in both restituting impairments in cortical physiology in clinical settings, as well as modulating cognitive abilities in the healthy population. However, neuroplastic after-effects of tDCS are highly dependent on stimulation parameters, relatively short lasting, and not expectedly uniform between individuals. The present study systematically investigates the full range of current intensity between 0.5 and 2.0 mA on left primary motor cortex (M1) plasticity, as well as the impact of individual-level covariates on explaining inter-individual variability. Thirty-eight healthy subjects were divided into groups of anodal and cathodal tDCS. Five DC intensities (sham, 0.5, 1.0, 1.5 and 2.0 mA) were investigated in separate sessions. Using transcranial magnetic stimulation (TMS), 25 motor-evoked potentials (MEPs) were recorded before, and 10 time points up to 2 h following 15 min of tDCS. Repeated-measures ANOVAs indicated a main effect of intensity for both anodal and cathodal tDCS. With anodal tDCS, all active intensities resulted in equivalent facilitatory effects relative to sham while for cathodal tDCS, only 1.0 mA resulted in sustained excitability diminution. An additional experiment conducted to assess intra-individual variability revealed generally good reliability of 1.0 mA anodal tDCS (ICC(2,1) = 0.74 over the first 30 min). A post hoc analysis to discern sources of inter-individual variability confirmed a previous finding in which individual TMS SI1mV (stimulus intensity for 1 mV MEP amplitude) sensitivity correlated negatively with 1.0 mA anodal tDCS effects on excitability. Our study thus provides further insights on the extent of non-linear intensity-dependent neuroplastic after-effects of anodal and cathodal tDCS.

Keywords: neuromodulation; neurophysiology; transcranial direct current stimulation; transcranial magnetic stimulation; variability.

© 2016 The Authors. The Journal of Physiology © 2016 The Physiological Society.

Figures

Figure 1. Course of study

Participants were randomly divided into two groups for tDCS polarity (Anodal: n = 20; Cathodal: n = 18). Each participant took part in five randomized sessions during which either sham, 0.5, 1.0, 1.5 or 2.0 mA stimulation with the respective polarity was applied. Prior to receiving stimulation, baseline MEP amplitude and SI1mV was measured over the determined motor cortical ‘hotspot’ which produced the largest MEP from the right ADM muscle. Next, DC stimulation for 15 min was delivered, and MEP measurements were taken again from the hotspot immediately after stimulation, as well as every 5 min up to 30 min, and then every 30 min up to 2 h after stimulation.

Figure 2. Intensity‐dependent effects in motor‐cortical excitability following anodal tDCS

A, after‐effects of cortical excitability following 15 min of anodal stimulation at intensities ranging from sham to 2.0 mA on the mean MEP amplitude. Error bars represent standard error. Filled symbols indicate a significant difference in cortical excitability against the respective baseline (Student's paired t test, two‐tailed, P < 0.05). Floating symbols (●, 0.5; ▲, 1.0; ■, 1.5; ◆, 2.0 mA) indicate a significant difference between the active intensity and sham stimulation (paired t test, two‐tailed, P < 0.05). Anodal stimulation over all active intensities resulted in significant increases of excitability lasting up to 30 min. Sham stimulation did not induce any significant change in cortical excitability. B, effect sizes and 95% confidence intervals of active tDCS intensities versus sham. MEP amplitudes were averaged into two time bins of early (0–30 min) and late (60–120 min) excitability changes, followed by calculation of Cohen's effect size d. Error bars represent 95% confidence intervals based on the pooled variance. Differences between active intensities were generally not discernable in the first 30 min; however, 0.5 and 2.0 mA resulted in slightly larger effects, especially in the time window 60–120 min.

Figure 3. Intensity‐dependent effects in motor‐cortical excitability following cathodal tDC

A, after‐effects of cortical excitability following 15 min of cathodal stimulation at intensities ranging from sham to 2.0 mA on the mean MEP amplitude. Error bars represent standard error. Filled symbols indicate a significant difference in cortical excitability against the respective baseline (paired t test, two‐tailed, P < 0.05). Floating symbols (●, 0.5; ▲, 1.0; ■, 1.5; ◆, 2.0 mA) indicate a significant difference between the active intensity and sham stimulation (paired t test, two‐tailed, P < 0.05). Only 0.5 mA and 1.0 mA cathodal stimulation resulted in significant differences from baseline, and only 1.0 mA was significantly different from sham during the early time bins. Higher intensities such as 1.5 and 2.0 mA (the latter is highlighted with double lines) tended to return to baseline values after about 10 min. Sham stimulation did not induce any significant change in cortical excitability. B, effect sizes and 95% confidence intervals of active tDCS intensities versus sham. MEP amplitudes were averaged into two time bins of early (0–30 min) and late (60–120 min) excitability changes, followed by calculation of Cohen's effect size d. Error bars represent 95% confidence intervals based on the pooled variance. Note that larger effects correspond to greater reduction of excitability from baseline (see Methods). Greatest differences are again seen with lower intensities of 0.5 and 1.0 mA whereas higher intensities did not result in marked changes.

Figure 4. Intra‐individual response to 1.0 mA anodal tDCS

Time courses of cortical excitability, measured as MEP amplitudes normalized to baseline, in a subgroup of 7 participants across three separate sessions are shown. Error bars represent standard error. In the table below, the intra‐class correlation coefficient (ICC(2,1)) was used to assess the strength of the reliability and reproducibility of 1.0 mA anodal tDCS, at each time point (first row), as well as over first 30 min and final 60 min (second and third row, respectively).

Figure 5. Differences in anodal and cathodal tDCS‐induced excitability between subjects with low and high sensitivity to TMS

The sample was median‐split (anodal: 48%; cathodal: 46%) with ‘Low SI1mV’ consisting of participants who required less than the median value to achieve an MEP amplitude of 1 mV and ‘High SI1mV’ consisting of the rest. Pairwise comparisons (panels A and C) between subgroups of the pooled average from the first 30 min were conducted using Student's unpaired, two‐tailed t test. Error bars represent the standard error. *Significant differences between the groups (P < 0.05). Effect size comparisons (panels B and D) were conducted by calculating Cohen's d. Error bars represent 95% confidence intervals of the pooled variance. A, subjects with Low SI1mV showed significantly greater increases in 1.0 mA anodal tDCS compared to subjects with High SI1mV. Note that within the sub‐groups, 1.0 mA was not significantly better than 1.5 mA (P = 0.081) for subjects with Low SI1mV and 1.5 mA was not significantly better than 1.0 mA (P = 0.073) for the subjects with High SI1mV. B, intensity effects for the first 30 min appear to follow a trend‐wise pattern, whereby lower intensities favour subjects with Low SI1mV while higher intensities favour subjects with High SI1mV. C, a subgroup comparison of cathodal tDCS does not reveal any significant pairwise difference during the first 30 min. D, effect size comparisons (where larger effects equate to greater reduction of cortical excitability) show a marginal intensity‐dependent tendency for the lowest (0.5 mA) and highest (2.0 mA) intensity only, which also follows the same pattern as anodal tDCS.

Figure 6. Inter‐individual differences in cortical excitability modulation following anodal tDCS

Post anodal tDCS time course (0–30 min) differences and trends between median‐split groups of low and high thresholds to TMS, based on the stimulus intensity for 1 mV amplitude (SI1mV). Error bars represent standard error. At lower intensities of 0.5 mA and 1.0 mA, subjects with lower SI1mV showed greater effects in cortical excitability facilitation whereas with higher intensities of 1.5 mA and 2.0 mA, subjects with higher SI1mV responded with a greater change in excitability compared to the Low SI1mV subjects. Notice that an upward trend of excitability facilitation was observed for subjects with higher SI1mV which was more pronounced at higher intensities, although this three‐way interaction could not be inferred as significant: F(18,1045) = 1.281, P = 0.198.

Figure 7. Relationship between individual TMS SI1mV sensitivity and efficacy of anodal tDCS on cortical excitability

For each active anodal tDCS intensity, each individual's grand‐averaged response over 0–30 min following stimulation was plotted as a function of his/her baseline TMS SI1mV (stimulus intensity for 1 mV amplitude). A negative correlation was observed with 1.0 mA anodal tDCS (r = −0.474, P = 0.035). [Colour figure can be viewed at wileyonlinelibrary.com]