TABLE 3.

Results of mediation analyses testing whether a proposed causal effect of X (predictor) on Y (outcome) may be transmitted through a mediating (M) variable.

							95% CI		Std. Coeff.
		Path	B	SE	t	p	UL	LL	β
	V1-PL	a	0.420	0.120	3.650	0.001	0.180	0.660	0.580
X = V1	PL-STSdp	b	1.060	0.340	3.130	0.005	0.360	1.760	0.590
Y = STSdp	V1-STSdp	c	0.540	0.230	2.360	0.030	0.070	1.020	0.420
M = PL	V1-STSdp | PL	c’	0.100	0.250	0.400	0.694	−0.410	0.600	0.080
	Indirect	a*b	0.450	0.190	–	–	0.076	0.840	–
	TPOJ1-PM	a	0.467	0.124	3.752	0.001	0.211	0.723	0.590
X = TPOJ1	PM-STSdp	b	1.258	0.468	2.688	0.013	0.294	2.221	0.540
Y = STSdp	TPOJ1-STSdp	c	0.707	0.330	2.139	0.042	0.027	1.386	0.390
M = PM	TPOJ1-STSdp | PM	c’	0.120	0.369	0.325	0.748	−0.639	0.879	0.070
	Indirect	a*b	0.587	0.286	–	–	0.002	1.125	–
	PI-MST	a	0.424	0.147	2.884	0.008	0.122	0.726	0.490
X = PI	MST-TPOJ1	b	0.615	0.162	3.807	0.001	0.282	0.948	0.620
Y = TPOJ1	PI-TPOJ1	c	0.388	0.149	2.600	0.015	0.081	0.695	0.450
M = MST	PI-TPOJ1 | MST	c’	0.127	0.139	0.915	0.369	−0.159	0.414	0.150
	Indirect	a*b	0.261	0.118	–	–	0.129	0.596	–

For each mediation analysis, the coefficients (B), standard error (SE), t-statistic, p-value and lower (LL) and upper (UL) levels for the 95% confidence interval (CI) and standardized beta coefficient are given for the direct paths: (a) X and M; (b) M and Y; (c) X and Y; (c’) X and Y, conditional on M and for the indirect (a*b) effect. In the first analysis (top), activity from V1 and STSdp (pSTS) parcels are the predictor and outcome variables, respectively, mediated by lateral pulvinar (PL) activation. Middle section shows the path from TPOJ1 (TPJ) to pSTS, mediated by medial pulvinar (PM). Bottom section shows the relationship between inferior pulvinar (PI) and MST mediated by TPJ. *p < 0.05.