Propósitos específicos

16 healthy volunteers (13 male, 16 right handed) aged 19-33 (mean age 23.13, SD 3.84) participated in these experiments. We did not conduct a prior power calculation to determine the sample size for this study. Instead, we decided to recruit a similar number of participants as in Chapter 3. Since collection of the data, a paper by Brown et al. has shown the predicted sample sizes required in TMS experiments for particular ICC(2,1) values, for different effect sizes (190). The numbers in the paper are calculated taking into account the day-to-day reliability of MEP measures within an individual. If we assume that the ICC(2,1) for the MEP amplitude at rest is 0.8, then we are adequately powered (approximately 80%) to see a paired t-test with an effect size of 0.8 or more with around 15 subjects. This makes the analyses of the paired t-test analyses of MEPs prior to movement and during movement preparation, adequately powered. However, to see a significant effect in a two-way ANOVA (i.e. different time courses of MEPs in two different conditions), we would need slightly larger numbers (approximately 20 subjects).

The study was approved by UCL Ethics Committee and none had contraindications to TMS, which was assessed by a TMS screening questionnaire.

4.2.1.2 Transcranial Magnetic Stimulation and Electromyography Recordings

TMS was delivered in the same fashion described in Chapter 3, experiment 1, with PA120

pulses only. EMG was collected as described in Chapter 3, experiment 1.

90

4.2.1.3 Stop-signal task and Go-only task

Participants were asked to perform both two blocks of the SST and two blocks of a simple reaction time (Go-only) task, which were driven by custom-made MATLAB (MathWorks) scripts using Psychtoolbox. For the SST, subjects were first presented with a white fixation cross on a black background. After 500 ms, an imperative stimulus (right arrow) was presented, which instructed the subject to press the ‘M’ key on the keyboard as fast as possible with their right index finger (go trials, n=105). On 25% of trials, a stop signal (red cross) appeared above the imperative stimulus at a variable delay after the imperative stimulus (stop trial, n=35). This delay, known as the stop signal delay (SSD) was controlled by a dynamic tracking algorithm, whereby the SSD would change depending on the outcome of the previous stop trial. The starting SSD was always set at 150 ms. If the subject successfully prevented their button press on a stop trial, the next stop trial would have its SSD set 50 ms later, whereas if the subject failed to stop, the next stop trial would have its SSD set 50 ms earlier. This dynamic tracking algorithm has been shown to reliably induce a convergence onto 50% successful inhibition across subjects.

The SSDs ranged from 100-250 ms (100, 150, 200 and 250 ms). There were also 15 baseline trials, where no signals were given, but TMS was given to give a representation of baseline corticospinal excitability. These trials also served as catch trials. The order of trials was pseudorandomised, such that one in every four trials contained a stop signal.

The Go-only task was similar to the SST, except no stop signals appeared in the block.

Hence, this was a block where no proactive control would be required. 105 go trials were given, with 15 trials without imperative stimuli to act as baseline.

To measure CSE during response preparation and execution, TMS was given in all trials, in all blocks to the M1 representation for the right FDI muscle, at an intensity required to produce a test MEP of 0.5 mV peak-to-peak amplitude. During go trials, one TMS pulse was given randomly at one of seven time points (at the imperative signal and 50, 100, 150, 200, 250 and 300 ms after the go signal). As such, 15 MEPs were taken at each time point. During stop trials, TMS was given 50 ms after the stop signal. In the 15 baseline

91

trials, TMS was given 1000 ms into the beginning of the trial to assess corticospinal excitability at rest. This has been summarised in figure 4.1.

Figure 4.1: TMS delivery in the Stop-signal and Go-only tasks.

SST: Go trials consist of a presentation of a fixation cross, followed by an imperative stimulus (right arrow) 500 ms later. In 25% of trials, the right arrow is followed by a stop signal (red cross) at one of four SSDs (100, 150, 200 or 250 ms after the arrow). Subjects must attempt to abort their button press on presentation of a stop signal. Failure to do so will result in the next stop signal having a shorter SSD (-50 ms) whereas success will lead to the next SSD becoming longer (+50 ms). TMS is delivered on go trials at one of seven time points (counterbalanced and randomised) or 1000 ms into a trial where no signals are shown (baseline trial). The Go-only task comprised of go and catch trials only; TMS was delivered at the same timepoints described above.

Behavioural measures taken included Go reaction time (reaction time on go trials), Stop Respond reaction time (reaction time on failed stop trials), average SSD and p(inhibit) (proportion of correct stop trials in the SST). We also calculated the SSRT using the mean method (mean go reaction time – mean SSD).

92

4.2.1.4 Data analyses

To investigate motor preparation used between the tasks, we plotted log normalised Go reaction time distribution histograms for each condition (Go-only and SST), for each participant. In rise-to-threshold models, if drift rate, boundary separation or non-decision time differs between conditions, this gives rise to different reaction time distributions (62). To assess whether these distributions were statistically different, we computed Levene’s test of equality of variances. As we were testing a null hypothesis in experiment 1, the variance of reaction distributions during go and stop blocks were equal, we computed a Bayesian paired t-test on standard deviations from each participant, using JASP (JASP Team (2018). JASP (Version 0.9.2)).

MEPs at each time point were collapsed into a grand average and then expressed as a fraction of the MEP at the go cue. A two-way repeated measures ANOVA with conditions BLOCK TYPE and TIME was performed. Based on the outcome of this analysis, post-hoc paired t-tests were performed between MEPs at each time point until they differed significantly from the go cue MEP, to assess if a delay had occurred in the rise of CSE.

To assess CSE during movement execution between blocks of potentially stopping in the SST compared to never stopping in the Go-only task, we controlled for reaction time differences (response-locked analysis). To this end, we calculated the time difference between TMS delivery and reaction time for each trial. MEPs were then categorised into 50ms time bins and a two-way repeated measures ANOVA was performed with main factors TIME BIN and BLOCK TYPE. Post-hoc paired t-tests were then performed to compare which were significant interactions.

4.2.2 Experiment 2: Conditional stop-signal task