Table 2.

One-sided permutation p-values based on a studentized Wilcoxon rank-sum statistic⁵⁶ of the global null hypothesis that the intervention (CE vs. control) had no effect on FA/λ_⊥ changes during the experiment (cf. Fig. 6). p-values have been adjusted for multiple comparisons using a closed testing procedure (FWE-correction) ⁵⁷. Fisher’s chi-square combining function⁵⁵ was used to summarize evidence over the partial tests (last column). Effect sizes for between-group comparisons at all time intervals are reported as Cliff’s delta (d)⁵⁸ with the related 95% CI. The magnitude of d can be interpreted using the following thresholds: |d|< 0.147 "negligible", |d|< 0.33 "small", |d|< 0.474 "medium", otherwise "large"⁵⁹.

	MRI_1–MRI_2	MRI_1–MRI_3	MRI_1–MRI_4	MRI_1–MRI_5	NPC
FA	p = 0.17, d = − 0.2, 95% CI [− 0.57, 0.23]	p = 0.16, d = − 0.19, 95% CI [− 0.55, 0.23]	p = 0.072, d = − 0.34, 95% CI [− 0.67, 0.1]	p = 0.044, d = − 0.49, 95% CI [− 0.78, − 0.03]	p = 0.027
λ⊥	p = 0.015, d = 0.48, 95% CI [0.05, 0.75]	p = 0.039, d = 0.38, 95% CI [− 0.05, 0.69]	p = 0.006, d = 0.64, 95% CI [0.21, 0.86]	p = 0.017, d = 0.43, 95% CI [0.01, 0.72]	p = 0.004