To answer the “When” question, paired and independent t-tests were computed for all time-samples (3.91 ms each per sample) of the epoch to determine differences between conditions and groups, respectively. The average reference was used and 5000 random permutations (i.e., bootstrapping) were performed. The levels of significance were corrected for multiple comparisons and false positives (Holmes et al., 1996; Nichols and Holmes, 2002). For each comparison, statistically significant (< .05, 2-tailed) blocks of five time-samples within the time-range of the N2 component were selected for further analysis. Out of the five, the third time-sample in the block had the highest significance. By selecting only five instead of all time-samples those were statistically significant in the N2 time-range, this analysis was made more conservative.
