Table 4.

Comparison of the regression performance of all methods with bi-modality data on 4-JRMI, i.e., AD/MCI-C/MCI-NC/NC

Methods	ADAS-Cog		MMSE
	CC	RMSE	CC	RMSE
FS	0.491 ± 0.34	5.759 ± 1.05	0.440 ± 0.16	2.366 ± 0.38
PCA	0.510 ± 0.56	5.647 ± 0.97	0.453 ± 0.47	2.390 ± 0.27
CCA	0.522 ± 0.25	5.531 ± 0.91	0.462 ± 0.53	2.334 ± 0.31
SJCR	0.538 ± 0.85	5.639 ± 0.97	0.493 ± 0.33	2.389 ± 0.30
M3T	0.521 ± 0.71	5.452 ± 1.01	0.528 ± 0.32	2.321 ± 0.27
Proposed	0.553 ± 0.34	3.926 ± 0.19	0.567 ± 0.23	1.929 ± 0.13

The boldface denotes the best performance of each column. (CC: Correlation Coefficient; RMSE: Root Mean Squared Error)