The comparison of a randomly selected subset of between-pair correlations that matched the number of within-pair correlations showed a similar, albeit slightly reduced, effect (marginally significant cluster: p = 0.0679; a cluster-randomization approach controlling for multiple comparisons over sensors; see Figure 3—figure supplement 1A for the group-averaged temporal similarity maps).
Instead, we found that the averaged spatial similarity across the −880 – −485 ms interval before the onset of the SFW (corresponding to 120–515 ms after the onset of SFW-1) remained larger for the within-pair sentences (R = 0.072 + /- 0.02) than the between-pair sentences (R = 0.063 + /- 0.03): t (25) = 1.81, p = 0.08 ( Figure 2—figure supplement 2 ), although in this subset analysis, the difference only approached significance due to the limited statistical power (on average there were only 40 within-pairs and 29 between-pairs after artifact rejection).