Machine Learning of Schizophrenia Detection with Structural and Functional Neuroimaging

Table 3

Top 15 most discriminative features for classification.

Regions	ALFF	DC	ReHo	VMHC	GMD	Weight
SPL_L_5_2	()	()	NS	()	NS	1.3275
IPL_L_6_2	NS	()	()	NS	NS	1.2037
MVOcC_L_5_1	()	()	()	()	NS	1.1467
ITG_L_7_3	()	()	()	NS	NS	1.0778
MFG_L_7_5	()	()	()	()	()	0.9969
MVOcC _R_5_1	NS	()	()	()	NS	0.9484
PhG_L_6_5	NS	()	NS	()	()	0.9476
LOcC_L_2_2	()	()	()	()	NS	0.9099
FuG_L_3_2	()	()	()	()	NS	0.8904
BG_R_6_5	NS	()	()	()	()	0.8877
MVOcC_R_5_3	()	()	()	()	NS	0.8413
CG_R_7_7	NS	NS	()	()	NS	0.7860
SFG_R_7_2	()	()	()	()	NS	0.7672
PhG_R_6_5	NS	()	NS	()	NS	0.7664
pSTS_R_2_2	NS	()	()	()	()	0.7283

NS: not significant, (optimal threshold); positive value means increased values in the SZ group. ALFF: amplitude of low-frequency fluctuations (ALFF); ReHo: regional homogeneity; DC: degree centrality; VMHC: voxel-mirrored homotopic connectivity; SPL: superior parietal lobule; IPL: inferior parietal lobule; MVOcC: medioventral occipital cortex; ITG: inferior temporal gyrus; MFG: middle frontal gyrus; PhG: parahippocampal gyrus; LOcC: lateral occipital cortex; FuG: fusiform gyrus; BG: basal ganglia; CG: cingulate gyrus; SFG: superior frontal gyrus; pSTS: posterior superior temporal sulcus; L: left; R: right.