Table S2. The weights of the ROIs (for GMM) and inter-ROI FCs (for FC) selected in the multimodal feature two-group classification are shown. The corresponding t-values from the two-sample t-test group comparison of the features are shown.

HC vs. MI / weight / t-value / HC vs. MP / weight / t-value / MP vs. MI / weight / t-value
PCAL.R / 1.54 / 5.17*** / PCL.L / 1.12 / 3.13** / HG.L / 0.38 / 2.44*
CUN.L-CUN.R / 1.03 / 5.08*** / vDC.R / 0.93 / 3.50** / CUN.R / 0.33 / 3.38**
rACC.L-PALL.R / 0.96 / 4.32*** / THA.R / 0.80 / 3.05** / CUN.L / 0.29 / 2.65*
LOFC.R-PALL.L / 0.59 / 4.07*** / PUT.L / 0.79 / 3.76** / PCAL.R / 0.11 / 2.95*
cACC.R-PALL.R / 0.57 / 4.15*** / PUT.R / 0.36 / 2.86* / LO.L / 0.10 / 2.54*
ENT.L-CAU.L / 0.52 / 4.20*** / LING.L / 0.24 / 3.06** / INS.R / 0.09 / 2.38*
PCAL.L-LING.L / 0.44 / 4.03*** / LING.R / 0.22 / 3.45** / STG.L / 0.05 / 2.03§
LING.R-LING.L / 0.43 / 5.51*** / TP.L-CAU.L / 1.94 / 3.51** / NA.L / -0.84 / -2.87*
CUN.R-LING.L / 0.39 / 5.21*** / PCAL.L-PCAL.R / 1.31 / 3.83** / rACC.L-PALL.R / 0.40 / 3.60**
cACC.R-PALL.L / 0.15 / 3.65** / cACC.L-TP.L / 1.24 / 3.90** / rACC.L-PALL.L / 0.39 / 3.78**
PCAL.R-LING.L / 0.12 / 4.22*** / rACC.L-TP.L / 0.27 / 3.26** / ENT.R-LOFC.R / 0.38 / 3.84**
THA.L-CAU.L / 0.12 / 3.71** / CUN.L-CUN.R / 0.19 / 3.27** / cMFG.L-FUS.R / 0.33 / 3.35**
rACC.L-PALL.L / -0.39 / 4.07*** / ITG.R-LOFC.R / 0.26 / 3.58**
vDC.L / 0.06 / 2.69* / LOFC.L-ENT.R / 0.26 / 3.77**
rMFG.R-ENT.L / 0.04 / 3.57** / cMFG.R / 0.03 / 2.82* / IFGtriang.R-CRBL.R / 0.26 / 3.37**
LOFC.R-HIP.L / 0.03 / 3.48** / PALL.L / 0.03 / 2.41* / SPL.L-LOC.R / 0.22 / 3.53**
HG.R-HG.L / 0.03 / 3.50** / NA.R / 0.01 / 2.50* / IFGorb.R-ENT.R / 0.22 / 3.53**
IFGoperc.R-HIP.L / 0.02 / 3.59** / STG.R / 0.01 / 2.44* / SPL.L-FUS.R / 0.20 / 3.78**
LOC.L-cACC.L / 0.02 / -3.29** / HIP.L / 0.01 / 2.60* / ICC.L-PALL.R / 0.15 / 3.21**
vDC.R-cACC.R / 0.02 / 3.50** / PCL.R / 4.94x10-3 / 2.41* / rMFG.L-PALL.L / 0.12 / 3.38**
IFGoperc.L-ICC.L / 0.02 / 3.06** / THA.L / -6.40x10-5 / 2.67* / IFGtriang.R-CRBL.L / 0.06 / 3.25**
CUN.L-LING.R / 0.01 / 3.45** / banksSTS.R-MTG.R / 0.07 / 2.96** / SMG.R-cMFG.R / 0.06 / 3.10**
LOFC.R-ENT.L / 0.01 / 3.50** / SMG.L-AMY.L / 0.04 / 3.04** / cACC.L-TP.L / -0.44 / -3.65**
PUT.R-cACC.R / 0.01 / 3.33** / TP.L-IFGtriang.L / 0.01 / 2.55* / PCAL.L-NA.L / -0.32 / -3.27**
INS.R-HG.R / 0.01 / 3.44** / PCAL.L-LING.L / 1.52x10-5 / 2.76* / TP.L-PALL.L / -0.31 / -3.79**
PCC.L-PALL.R / 3.92x10-3 / 2.77** / FP.L-IFGtriang.R / -0.01 / -2.99** / LOFC.L-rACC.L / -0.09 / -3.17**
CUN.L-LING.L / -0.04 / 3.60** / TP.R-LOC.L / -0.02 / -2.53*
LOC.R-cACC.R / -0.03 / -3.26** / LOC.L-NA.R / -0.03 / -3.10** / PCAL.L / -0.02 / -2.59*
LOC.R-cACC.L / -0.03 / -3.24** / MTG.L-PUT.R / -0.04 / -3.03** / INS.L / -0.01 / -2.06*
cACC.R-LOC.L / -0.02 / -3.38** / IFGorb.R-rMFG.R / -0.04 / -2.97** / PCUN.L / -0.01 / -2.00
THA.R-PCL.R / -0.01 / -3.43** / PALL.L / 3.67x10-3 / 1.39
THA.L-PCL.R / -0.01 / -3.39** / IPL.L / -0.01 / 2.03§
PCL.R-cACC.L / -0.01 / -3.30** / HIP.L / -0.01 / -1.59
AMY.L / -0.02 / -1.77
INS.L-cMFG.R / 0.04 / 3.24**
FUS.L-SPL.L / 0.03 / 3.20**
SPL.L-banksSTS.R / 0.03 / 3.09**
PCC.R-cMFG.R / 0.03 / 3.02**
ENT.R-IFGorb.L / 0.02 / 2.96*
MTG.L-ITG.L / 0.02 / 2.81*
rACC.R-PALL.L / 0.02 / 3.24**
cACC.R-PCUN.L / 0.01 / 3.04**
LING.R-LING.L / 0.01 / 2.43*
ENT.R-MOFC.L / 0.01 / 2.82*
cACC.R-ICC.L / 0.01 / 2.85*
NA.L-PAL.R / 0.01 / 2.84*
vDC.R-CAU.R / 4.34x10-3 / 2.80*
CAU.R-PUT.L / 4.33x10-3 / 2.91*
SFG.R-cMFG.R / 4.11x10-3 / 2.71*
LOC.L-SPL.L / 4.03x10-3 / 3.11**
HIP.L-LOFC.R / 3.84x10-3 / 2.59*
CAU.R-PALL.L / 3.46x10-3 / 2.54*
HIP.L-IFGoper.R / 3.17x10-3 / 2.82*
IFGorb.R-rMFG.L / 2.66x10-3 / 2.79*
cACC.R-cACC.L / 2.50x10-3 / 2.88*
PALL.L-MOFC.R / 1.11x10-3 / 2.91*
rMFG.R-PCL.L / -0.01 / -2.55*
PCC.L-PCAL.L / -0.01 / -2.85*
MTG.R-bansSTS.R / -0.01 / -2.30*
vDC.R-LING.L / -4.44x10-3 / -2.82*
vDC.R-LING.R / -4.03x10-3 / -2.80*
PALL.L-PCAL.R / -4.01x10-3 / -2.54*
vDC.L-FUS.L / -1.59x10-3 / -2.91*
FUS.R-IFGorb.L / -9.63x10-4 / 2.93*

Note: Features in Bold are the features whose absolute weights are above the 5% of the maximum absolute weight over all the features in each classfication. Features underlined indicate their weights were of different signs from 2-sample t-test results. Asterisks represent the significance after FDR correction: §- p<0.05 (no correction), *- pFDR<0.05, **- pFDR<0.01, ***- pFDR<0.001. Positive signs of the weights and the t-values reflect that the features are higher in the former group compared to the latter group (i.e., HC vs. MI, HC vs. MP and MP vs. MI); negative signs of the weights and the t-values indicate the opposite. For full name of the ROIs, please refer to Table S1. .L- left hemisphere; .R- right hemisphere.