Table 2.

Mean MAE ± SD resulting from age prediction for each class and each harmonization technique for the three regression algorithms SVR, Random Forest, and Lasso. The best performances are highlighted in gray.

Harmonization Technique	NC			ASD
	SVR	RF	Lasso	SVR	RF	Lasso
No harmonization	$2.66\pm 0.38$	$2.59\pm 0.37$	$2.79\pm 0.37$	$2.56\pm 0.49$	$2.54\pm 0.45$	$2.80\pm 0.51$
Age covariate	$2.70\pm 0.36$	$2.62\pm 0.32$	$2.68\pm 0.37$	$2.46\pm 0.40$	$2.46\pm 0.39$	$2.68\pm 0.36$
No age covariate	$4.40\pm 0.55$	$3.67\pm 0.54$	$4.82\pm 0.53$	$4.28\pm 0.58$	$3.29\pm 0.55$	$4.56\pm 0.59$