Response variability of different anodal transcranial direct current stimulation intensities across multiple sessions

Abstract

Background: It is well known that transcranial direct current stimulation (tDCS) is capable of modulating corticomotor excitability. However, a source of growing concern has been the observed inter- and intra-individual variability of tDCS-responses. Recent studies have assessed whether individuals respond in a predictable manner across repeated sessions of anodal tDCS (atDCS). The findings of these investigations have been inconsistent, and their methods have some limitations (i.e. lack of sham condition or testing only one tDCS intensity).

Objective: To study inter- and intra-individual variability of atDCS effects at two different intensities on primary motor cortex (M1) excitability.

Methods: Twelve subjects participated in a crossover study testing 7-min atDCS over M1 in three separate conditions (2 mA, 1 mA, sham) each repeated three times separated by 48 h. Motor evoked potentials were recorded before and after stimulation (up to 30min). Time of testing was maintained consistent within participants. To estimate the reliability of tDCS effects across sessions, we calculated the Intra-class Correlation Coefficient (ICC).

Results: AtDCS at 2 mA, but not 1 mA, significantly increased cortical excitability at the group level in all sessions. The overall ICC revealed fair to high reliability of tDCS effects for multiple sessions. Given that the distribution of responses showed important variability in the sham condition, we established a Sham Variability-Based Threshold to classify responses and to track individual changes across sessions. Using this threshold an intra-individual consistent response pattern was then observed only for the 2 mA condition.

Conclusion: 2 mA anodal tDCS results in consistent intra- and inter-individual increases of M1 excitability.

Keywords: Cortical plasticity; Inter-individual variability; Intra-individual variability; Transcranial direct current stimulation.

Figure 1

Three identical sessions were performed for each tDCS condition at least 48 hours apart. Each session consisted of TMS pre-measures; defining stimulus intensity required to evoke mean MEP peak-to-peak amplitude of 1mV (SI_1mV, Pre). Then, atDCS 1 or 2 mA was applied for 7 min (in sham condition current passed only for 30s) with the active electrode on the FDI ‘hotspot’. Post-measurements were performed using the SI_1mV intensity immediately (Post₀), 15 (Post₁₅), and 30 (Post₃₀) min after tDCS.

Figure 2

Overall atDCS effect and reliability across consecutive sessions. A) Combined session average of MEP amplitude (mV) for each time point (Pre, Post₀, Post₁₅, Post₃₀) and condition (2mA, 1mA, sham). Two-way ANOVA_RM was performed with CONDTION and TIME, comparing pre with post-values. Post-hoc pairwise comparisons were performed using Bonferroni correction (*, p≤0.05; **, p≤0.01; ***, p≤0.001). B) Mean MEP amplitudes for each session and condition across time. ICC, calculated with all post-time points from the 3 sessions for each condition (n = 9/condition), showed fair to high reliability. Errors bars represent SEM.

Figure 3

Grand Mean values of MEP amplitude for each individual across sessions and conditions (2mA, 1mA, sham). Each colored line represents a single subject across all nine sessions. ICC calculated for each condition reveals poor to fair reliability of tDCS post-effects.

Figure 4

Normal frequency distribution plots of mean MEP ratios (post/pre) for all sessions within each condition (n= 36/condition). The grey area represents overlapping ratios found in the sham, 2mA and 1mA conditions. Dotted lines indicate SDs from mean sham values.

Figure 5

Pie charts of post-tDCS responses across sessions classified with the Sham Variability-Based Threshold. A) Responses were classified as ‘increase’ (> 1.3), ‘no change’ (0.7 < ratio < 1.3) and ‘decrease’ (< 0.7). B) Individual response tracking across sessions for each condition. Same classification scheme as in A but separated by session. Dot sizes are linearly proportional to the number of subjects (n = 12). The biggest dot size represents 9 subjects and the smallest dot size represents 1 subject.