Table 3.

Statistical evaluation of the effect of the preprocessing procedures for patients.

Analysis method	Parameter	Main effect (Friedman test)		Post-hoc (Wilcoxon signed ranks test)
				P1–P2	P2–P3	P3–P4	P4–P5
		χ2	p	p	p	P	p
Independent component analysis	Qualitative	43	<0.0001	0.01	1	1	1
	PCC: z-score	31	0.0002	0.008	1	1	1
	MPFC: z-score	19	0.04	0.4	0.04	1	1
	LPR: z-score	11	0.9	–	–	–	–
	LPL: z-score	14	0.3	–	–	–	–
	Extra-DMN: %	48	<0.0001	0.002	0.02	1	1
Seed-based analysis	MPFC: z-score	49	<0.0001	1	0.002	1	0.8
	LPR: z-score	58	< 0.0001	0.4	0.01	1	1
	LPL: z-score	49	< 0.0001	0.4	0.006	1	0.9
	Extra-DMN: %	102	<0.0001	0.0002	0.0002	0.005	0.01
Linear regression	LPL-LPR: r	66	<0.0001	0.001	0.01	0.9	1
	PCCL-LPL: r	50	<0.0001	0.014	0.001	1	1
	PCCR-LPR: r	56	<0.0001	0.004	0.004	1	1
	PCCL-MPFCL: r	37	<0.0001	0.006	0.0008	1	0.1
	PCCR-MPFCR: r	37	<0.0001	0.003	0.0006	1	1

As described in-text, Friedman tests followed, where appropriate, by Wilcoxon post-hocs were performed. All p-values are reported following Bonferroni-Holm correction. PCC, posterior cingulate and precuneus; LP, lateral parietal cortex; MPFC, anterior cingulate and medial prefrontal cortex. Subscripts “L” or “R” stand for left or right hemisphere. “Extra-DMN” refers to the proportion of activated voxels outside the expected DMN localization. The effect of preprocessing “grade” was significant for most parameters, with the greatest overall differences being observed between P1–P2 and P2–P3, corresponding to the addition of band-pass filtering and removal of co-variance with the movement vectors and global brain parenchyma signal. See text for details.