Conventional or High Definition tDCS to Enhance Implicit Motor Sequence Learning in Healthy Young Adults?

STUDY DESIGN

The investigators will conduct a single-blind, sham-controlled, counterbalanced study. For the sequence-specific aspect of IMSL (primary outcome), a mixed factorial repeated measures ANOVA will be carried out with "device" (2 levels: conventional tDCS, HD tDCS) as between-subjects factor and "stimulation" (2 levels: anodal, sham), "blocks" (2 levels: random block, mean of adjacent blocks) and "time" (3 levels: during, post5min, post1week) as within-subjects factors. Similarly, for general learning (secondary outcome), a mixed factorial repeated measures ANOVA will be executed with "device" (2 levels: conventional tDCS, HD tDCS) as between-subjects factor and "stimulation" (2 levels: anodal, sham), "blocks" (7 levels: Blocks 1-6, Block 8) and "time" (3 levels: during, post5min, post1week) as within-subjects factors. Participants are randomly assigned to either the conventional tDCS group or the HD tDCS group by block randomization. All participants will receive both anodal (real) and sham (placebo) tDCS in a random order. Counterbalancing will be done by an independent investigator using Microsoft Excel®.

RECRUITMENT STRATEGY

Healthy young adults will be recruited from the Vrije Universiteit Brussel. There are no restrictions or prohibitions for the subjects.

MATERIALS

For the Conventional tDCS, a 1x1 Low Intensity Direct Current Stimulator (Soterix Medical Inc, New York, USA) will be used to generate and deliver tDCS through a pair of identical square rubber electrodes (size 35 cm2), placed in rectangular saline-soaked sponges. For the stimulation of M1, electrodes will be placed over C3 or C4 according to the international 10-20 electroencephalogram system, matching with the M1 contralateral to the performing hand. The reference electrode will be positioned on F1 or F2, ipsilateral to the performing hand.

The current stimulation will be slowly ramped up from 0 milliampere (mA) to 2 mA in one minute. For the anodal tDCS condition, this intensity will be maintained for the duration of the SRT-task (approximately 20 minutes), which is well within evidence-based safety standards for tDCS. This will result in a current density of 0,057 mA/cm2. For the sham tDCS condition - unbeknown to the subject - stimulation will be gradually decreased towards 0 mA immediately after the one-minute ramp-up. During the last block of the SRT-task, this gradual ramping-up and -down of the current stimulation will be repeated to optimize the process of blinding of participants. To control for blinding of the subjects, after the last session subjects will be asked whether they were aware of the stimulation condition or not. Transient side-effects will be inventoried by the experimenter during and two weeks after the tDCS protocol, and may include a slight itching sensation under the electrode, redness of the skin under the electrode, headache, nausea, fatigue or insomnia.

For the High Definition tDCS, a 4x1 Multichannel Stimulation Adapter (Soterix Medical Inc, New York, USA) will be used to deliver HD tDCS over M1. By connecting the conventional tDCS device (described above) to this Multichannel Stimulation Adapter, the direct current is delivered along the 4x1 HD tDCS configuration, allowing for neuromodulation restricted to the desired area. In contrast to the relatively large sponge-electrodes used in conventional tDCS montages (Figure 1A), stimulation via HD tDCS is delivered by means of one central gel-electrode and four return-electrodes placed in plastic encasings embedded in an EEG cap (Figure 1B). Set-up of the device will be done according to the extensive experimental protocol provided by Villamar and colleagues (2013). For the stimulation of the left M1, contralateral to the performing right hand, the anodal stimulation will be delivered via the central electrode corresponding with C3 and held in place using the specially designed synthetic cap to hold the HD-tDCS electrodes on the head. The return-electrodes are positioned at Cz, F3, T7 and P3 (international 10-20 electroencephalogram system), see Figure 2. For the right M1, contralateral to the performing left hand, the central electrode will be positioned at C4 with the return-electrodes at Cz, F4, T8 and P4. The stimulating (anodal) direct current will be ramped up for 60 seconds to an intensity of 2 mA, and maintained for the duration of the SRT-task (about 20 minutes).

The Serial Reaction Time task (SRT-task) will be used to determine IMSL. The SRT- task will be performed on a laptop using E-Prime® software (Psychology Software Tools, Inc., Pittsburgh, Pennsylvania, USA). Participants will be asked to press the horizontally aligned response keys C, V, B, N of a standard azerty keyboard for a leftmost, left, right, rightmost target, respectively. Responses will be given with the index finger of the least affected hand in the case of the persons with PD and with the dominant hand in the case of the healthy controls. If both hands are equally affected in the persons with PD, the dominant hand will be used as well. The response keys C, V, B and N will be the only visible keys, all other keys will be covered.

PROCEDURE

The experiment will take place at a laboratory of the Vrije Universiteit Brussel (VUB) or in a silent room at the participant's home, under supervision of the investigator.

Following a screening session (T0), in which baseline demographic characteristics (gender, age, dominant hand) will be collected. Participants will be assigned to either the conventional tDCS group or the HD tDCS group and will be seen four times (T1-T4) over the course of minimally five to maximally eleven weeks. All four sessions will start with a practice SRT-task of one random block of 25 trials, followed by the actual experimental SRT-task of eight blocks of 72 trials with a thirty second break between consecutive blocks. In Blocks 1 through 6 and Block 8, the order of the target (i.e. black dot) locations will follow a repeating sequence. This is unbeknown to the participant. The rationale is that reaction times will decrease with repetition of the sequence throughout Blocks 1-6 and 8, denoting a general training effect (secondary outcome measure). When the repeating sequence suddenly changes to a random sequence in Block 7, reaction time will increase in Block 7 and decrease once more in regularly sequenced Block 8, denoting a sequence-specific learning effect (primary outcome measure). To control for possible carry-over effects, the SRT-task will follow a different sequence in each stimulation condition (e.g. 132342134142 in T1-T2 and 243413241213 in T3-T4). To make sure IMSL is independent of a specific sequence, six different, structurally identical, sequences of 12 elements long will be counterbalanced between the participants.

The first interventional session (T1) will be planned at least 1 week after the screening session (T0) and will consist of active (anodal) tDCS or sham (placebo) tDCS administered during the SRT-task. Five minutes post-tDCS, subjects will be asked to carry out a short, three-block version of the SRT-task without application of tDCS to investigate potential short-term consolidation effects: Blocks 1 and 3 following the same regular sequence as earlier; Block 2 following a random sequence.

The second session (T2) will be planned one week later. During this one-week post-tDCS session the same, full version of the SRT-task from one week earlier will be performed, this time without the application of tDCS, to determine potential long-term consolidation effects. After T2, a washout period of at least three weeks will be planned to control for carry-over effects between the two stimulation conditions (active/sham tDCS). Cross-over will take place and the same procedure with the opposite stimulation condition will be repeated during the third (T3) interventional and fourth (T4) follow-up session. Half of participants in each group will have received active tDCS during T1 and sham tDCS during T3, while the other half of participants will have received these conditions in reversed and randomized order.

The current stimulation will be slowly ramped up from 0 mA to 2 mA in one minute. For the anodal tDCS condition, this intensity will be maintained for the duration of the SRT-task (approximately 20 minutes), For the sham tDCS condition - unbeknown to the subject - stimulation will be gradually decreased towards 0 mA immediately after the one-minute ramp-up.

A post-SRT-task questionnaire will be completed after the last session (T4) to determine whether participants became aware of the sequential nature of the task. If participants indicate that they believe a specific sequence appeared, they will have to reproduce the sequence of the last session as correct as possible.

STATISTICAL ANALYSES

All statistical analyses will be carried out using International Business Machines (IBM) Statistical Package for the Social Sciences (SPSS) Statistics version 26. The level of significance will be set at α = 0.05. A trend towards significance will be defined as 0.05 ≤ α < 0.10. Appropriate corrections for multiple comparisons will be made when necessary. Cohen's f effect sizes will be reported, with values of .10, .25, and .40 representing small, medium, and large effect sizes, respectively.

In the event of null-effects, the investigators will conclude that there is no evidence of a difference between conditions. However, the investigators will also calculate post-hoc Bayes factors for each group (conventional tDCS, HD tDCS) to assess whether a lack of difference in sequence learning between the anodal and sham stimulation conditions could be interpreted as evidence for the absence of an effect of tDCS on sequence learning.

Correlation analyses, Bonferroni-corrected for multiple comparisons, will be performed to investigate if the amount of IMSL correlates with demographical, variables (including clinical subtypes of PD). If assumptions for parametrical testing are violated, the non-parametric alternative Spearman's Rho will be calculated.

The outcomes of the SRT-tasks during, at 5 minutes post and one week post sham tDCS will serve as baseline measures to be compared with the outcomes following actual anodal tDCS.

The analyses of the SRT-task performance will be based on median reaction time (RT) per block instead of mean RT to minimize potential outlier effects. Practice trials, the first response after each break, erroneous responses and responses following an error will be excluded from the analyses. Median RTs per block will be analyzed to determine (1) a general learning effect (secondary outcome measure) and (2) a sequence-specific learning effect (primary outcome measure).

General learning effects during and at one-week post-tDCS will be derived from a decline in median RTs over the seven regularly sequenced blocks (i.e. Blocks 1-6 and Block 8). This will not be applicable to the 5-minutes post-tDCS SRT-task as it is a short version with only three blocks. A 2x2x2x7 repeated measures ANOVA will be carried out with device (conventional tDCS, HD tDCS) as between-subjects factor and stimulation (active, sham), time (during, post1week) and block (Blocks 1-6, Block 8) as within-subjects factors.

Sequence-specific learning effects during, 5 minutes and 1 week post-tDCS will be analyzed by subtracting the mean of the median RTs of adjacent sequenced blocks (Blocks 6 and 8 during stimulation and at one week post-tDCS; Blocks 1 and 3 at 5 minutes post-tDCS) from the median RT of the random block (Block 7 during stimulation and at 1-week post-tDCS; Block 2 at 5-minutes post-tDCS). A 2x2x2x3 repeated measures ANOVA (or Friedman and Wilcoxon signed rank tests as non-parametrical alternatives) will be carried out with device (conventional tDCS, HD tDCS) as between-subjects factor and stimulation (active, sham), sequence (random block, mean of adjacent sequenced blocks) and time (during, post5min, post1week) as within-subject factors.

In case assumptions of sphericity are violated, Greenhouse-Geisser or Huynh-Feldt corrections will be reported. Bonferroni-corrected t-tests will be implemented to further analyze significant main and interaction effects.

Error percentages in the SRT-task are generally small and thus, because of a limited number of observations, less sensitive to IMSL. The percentage erroneous reactions per block will be calculated for both stimulation conditions (anodal, sham) and for the three measurements over time (concurrent, post5minutes, post1week). The Shapiro-Wilk test of residuals will be carried out to evaluate normality of distribution.

The sequential score, as outcome measure for explicit knowledge, will be taken up as a covariate in the analyses. An independent samples t-test (or non-parametrical alternative Mann-Whitney U test) will be carried out to ascertain whether the sequential scores (x/12) of the conventional tDCS group are different from the HD tDCS group.