Table 6.

Comparison with the classification scores of previous studies using multimodal data.

Study reference	Modality	Sample size	Features	Feature count	Classifier	Accuracy (%)
Proposed Method	sMRI + fMRI	144	Structural ROI measures, Global functional connectivity, Group ICA	748	ELM	99.29
Du et al., 2012	rs-fMRI + task-related fMRI (auditory oddball task)	56	Kernel PCA with spatial ICA maps	53	Fisher's linear discriminant	98
Silva et al., 2014	sMRI + fMRI	144	Gray matter density based ICA features (sMRI), group ICA features (fMRI)	410	Gaussian process classifier	94
Cetin et al., 2016	rs-fMRI + task-related fMRI (auditory oddball task) + MEG	55	Group ICA based functional connectivity scores, MEG data for each frequency	103	LDA, Naïve Bayes classifier, and non-linear SVM	90
Ota et al., 2013	sMRI + diffusion MRI	50	Volume and fractional anisotropy (FA) of certain regions (insula, thalamus, ACC, ventricles, and corpus callosum)	31	Linear discriminant analysis	88
Yang et al., 2010	task-related fMRI (auditory oddball task) + SNP (genetic data)	40	Voxels in the fMRI map, ICA components, SNPs	411	Support vector machine	87
Sui et al., 2013	sMRI + rs-fMRI + diffusion MRI	63	GM density, ALFF (amplitude of low-frequency fluctuation), FA	1,863	Support vector machine	79