Optimal fixed pipelines for Weak-Contrast data.

The optimal fixed pipeline combinations for weak contrast (TaskB vs. TaskA), identified via Nemenyi test (p = 0.05). Preprocessing includes ICA (M MELODIC; P = PESTICA), motion correction (MC), motion parameter regression (MPR), physiological noise correction via RETROICOR (RET) and polynomial detrending for orders 0–5 (DET). Pipelines with the same Cluster Group number have no significant difference in SPMs, based on overlapped 95% confidence ellipses in DISTATIS space, shown in Fig. 4.