Table 3.

Kernel contribution (d_m) to the MKL models performing significantly above chance level, along with the expected ranking, across folds

Model MKL [TYP] vs. [COM]			Model MKL [TYP] vs. [DCD-COM]			Model MKL [TYP] vs. [DD-COM]
ROI#	dm (%)	ER	ROI#	dm (%)	ER	ROI#	dm (%)	ER
Cereb7 (‘default-mode’)	44.4370	14	Cereb7 (‘default-mode’)	39.8886	14	Cereb7 (‘default-mode’)	46.3638	14
Cereb3 (‘dorsal attention’)	13.4819	12.4103	Cereb4 (‘ventral attention’)	13.1845	12.2619	Cereb6 (‘frontoparietal’)	14.7626	12.5476
Stria6 (‘frontoparietal’)	12.4501	11.7586	Stria6 (‘frontoparietal’)	12.2180	11.9048	Stria6 (‘frontoparietal’)	13.8192	12.3571
Cereb4 (‘ventral attention’)	11.0945	11.1379	Cereb3 (‘dorsal attention’)	10.4024	11.0952	Cereb3 (‘dorsal attention’)	7.0694	10.2857
Stria7 (‘default-mode’)	8.1143	9.8966	Cereb2 (‘somatomotor’)	7.7157	9.6667	Cereb4 (‘ventral attention’)	7.0532	9.9762
Stria4 (‘somatomotor’)	6.1296	8.7931	Cereb6 (‘frontoparietal’)	6.0691	8.9524	Cereb2 (‘somatomotor’)	4.0890	8.0714
Cereb1 (‘visual’)	3.9389	7.6552	Stria4 (‘somatomotor’)	5.9496	8.4048	Cereb5 (‘limbic’)	3.2966	7.6667
Cereb6 (‘frontoparietal’)	0.3346	1.4828	Stria7 (‘default-mode’)	3.4399	6.7619	Stria7 (‘default-mode’)	3.1173	7.2381
Cereb5 (‘limbic’)	0.0191	0.2414	Cereb5 (‘limbic’)	0.4929	2.3810	Stria5 (‘limbic’)	0.3052	2.0476
Cereb2(‘frontoparietal’)	0	0	Cereb1 (‘visual’)	0.3341	2.3571	Stria4 (‘somatomotor’)	0.1237	1.0714
Stria1 (‘somatomotor’)	0	0	Stria5 (‘limbic’)	0.3052	2.0476	Cereb1 (‘visual’)	0	0
Stria2 (‘somatomotor’)	0	0	Stria1 (‘somatomotor’)	0	0	Stria1 (‘somatomotor’)	0	0
Stria3 (‘somatomotor’)	0	0	Stria2 (‘somatomotor’)	0	0	Stria2 (‘somatomotor’)	0	0
Stria5 (‘limbic’)	0	0	Stria3 (‘somatomotor’)	0	0	Stria3 (‘somatomotor’)	0	0

Expected ranking was computed as the average of the ranking across folds, thus evaluating the stability of the kernel contribution. Only the kernels that carry the most predictive information about group classification (>10% of contribution to MKL) are discussed in the text.