These two cohorts were taken from two different samples, one collected in Boston, MA, USA and the other collected in Barcelona, Spain. We chose to analyse the data separately rather than combining the data because we felt that we had sufficient power to analyse the two samples separately, see more and this provided us with an opportunity to test the validity and generalizability of the finding. From the data from the first cohort, a receiver operating characteristic (ROC) curve was created and the area under the curve at the various timepoints was determined by calculating

the c-statistic. Based on this statistic a timepoint was chosen at which returning to baseline would optimally differentiate between the first cohort groups. This value was then applied to the new cohort’s data and diagnostic sensitivity and specificity values were obtained. All participants tolerated the TMS study without any side-effects or complications. Consistent with prior findings (Theoret et al., 2005), AS and control groups did not differ significantly in resting motor threshold (RMT) (mean ± SD: ASD, 42.6 ± 6.0; Control, 46.9 ± 6.6; P = 0.14)

Erismodegib clinical trial or in baseline MEP values prior to either cTBS (P = 0.48) or iTBS (P = 0.51). Consistent with our hypothesis, the AS group showed greater and longer-lasting modulation of their MEPs following both forms of TBS. The average time to return to baseline MEP values following cTBS was 35.5 ± 13.2 min for the controls, while the AS group did not return to baseline

levels until an average of 87 ± 26.3 min (Fig. 2). Similarly, for iTBS, the average time taken to return to baseline was 37.2 ± 35.3 min in the control group and 77.8 ± 31.3 min in the AS group. These differences were significant for both forms of TBS (cTBS: t19 = 8.20, P < 0.001; iTBS: t8 = 3.04, P < 0.05) and were not correlated with age, IQ, ADOS score or ADI score (all P > 0.05). In addition, following cTBS, the AS group was significantly different in baseline-corrected others MEPs as compared to the control group, beginning at 20 min post-TBS and lasting until 50 min post-TBS (all P-values < 0.004 Bonferroni-corrected). For the iTBS paradigm, the groups were not significantly different at any timepoint after Bonferroni correction was applied. We chose to use the cTBS paradigm to test the diagnostic potential of this TMS protocol in a different cohort. The cTBS paradigm was chosen for this second cohort based on several factors. Firstly, the cTBS paradigm was found to be more reliable than the iTBS paradigm in the first cohort. Secondly, to simplify the study we only wanted to include a single TBS session and we felt that the cTBS protocol, being a suppressive protocol, would be theoretically safer (i.e. less likely to induce a seizure). Using data from the first cohort, we calculated an ROC curve, which provided a c-statistic (area under the curve) of 0.966 ± 0.