Table 1:

Three groups of image datasets are used throughout the experiments and split into train, validate, and test subsets of specified size.

Group	Train	Validate	Test	Datasets
Within-contrast	7,400	5,000	5,030	ABIDE (Di Martino et al., 2014)
				ADHD-200 (Milham et al., 2012)
				GSP (Dagley et al., 2017)
				MCIC (Gollub et al., 2013)
				OASIS-1 (Marcus et al., 2007)
				PPMI (Marek et al., 2011)
				UK Biobank (Sudlow et al., 2015)
				Buckner40 (Fischl et al., 2002)
Multi-contrast	496	496	496	HCP (Bookheimer et al., 2019)
				FSM (in-house data)
Longitudinal	48	48	48	OASIS-2 (Marcus et al., 2010)