Extended Data Fig. 5: Control analyses confirmed inter-brain neural correlations.
We conducted a series of control analyses to validate the findings of inter-brain neural correlations (Supplementary Note 3). a-d, Circular shifted control. The t value of actual inter-brain correlations (in 30-150 Hz) for each state were compared against null distributions of t statistics created from the correlations of circularly shifted data between cooperators. We found increased inter-brain neural correlation of the actual data compared to that of circularly shifted data for both the amygdala (especially in the initiation state: p = 0.012, a; similar increasing trend in the maintenance state, p = 0.149, c) and TPJ (initiation: p = 0.027, b; maintenance: p < 2 × 10-4, d). e-h, Competition dyad as a control condition. We calculated the inter-brain correlation between competitors and compared it with that of cooperators. We found higher inter-brain correlations between cooperators than competition dyads during both the initiation (e, amygdala: t64 = 2.477, p = 0.016, Cohen’s d = 0.403, 95% CI: 0.003, 0.032; f, TPJ: t95 = 1.910, p = 0.059, Cohen’s d = 0.258, 95% CI: -2.803 × 10-4, 0.015) and maintenance states (g, amygdala: t65 = 2.040, p = 0.045, Cohen’s d = 0.347, 95% CI: 1.529 × 10-4, 0.014; h, TPJ: t92 = 3.322, p = 0.001, Cohen’s d = 0.435, 95% CI: 0.005, 0.190). Probability distributions of correlation coefficients for each channel (amygdala: n = 67, TPJ: n = 97) are shown as violin plots with black dots representing the mean and error bars representing 95% CI. Correlation coefficients were Fisher-transformed, and coefficient outliers (mean ± 3 s.d.) were excluded prior to statistical tests. Two-sided paired t-tests were used. i-n, Pseudo dyad control. The t value of real-interacting cooperators’ inter-brain correlations (in 30-150 Hz) for each state and state difference were compared against null distributions of t statistics calculated from neural correlations between pseudo dyads. The significant inter-brain neural correlations in the amygdala and TPJ were replicated during both initiation (i, amygdala: p < 2 × 10-4; j, TPJ: p = 0.005) and maintenance states (k, amygdala: p = 0.063; l, TPJ: p < 2 × 10-4). Moreover, the inter-brain neural correlations in the amygdala showed a significant increase during initiation than maintenance states (m, amygdala: p < 2 × 10-4), while TPJ showed an opposite pattern with larger correlation in the maintenance state than initiation state (n, TPJ: p < 2 × 10-4). In a-d and i-n, vertical solid lines and dashed lines indicate observed inter-brain correlations and the upper/lower 95% CI of the null distributions (one-sided non-parametric permutation tests). Asterisks denote a significant difference (†p < 0.10, *p < 0.05, **p < 0.01, uncorrected).
