. 2021 Apr 5;16(4):e0249460. doi: 10.1371/journal.pone.0249460

Table 2. Mean ± standard deviation (min to max) cross-validated error and 95% confidence interval (min to max, last row) of myelin model accuracies.

Model	Loss	R2	Pearson Correlation	Residual Correlation
Linear GAM	RMSE	0.36 ± 0.17	0.68 ± 0.11	0.83 ± 0.05
	RMSE	(−0.05 to 0.60)	(0.47 to 0.85)	0.83 ± 0.05
	MAE	0.36 ± 0.21	0.68 ± 0.11	0.76 ± 0.07
	MAE	(−0.17 to 0.64)	(0.47 to 0.85)	0.76 ± 0.07
Segmentation Regression	RMSE	0.46 ± 0.16	0.70 ± 0.12	0.71 ± 0.10
	RMSE	(0.17 to 0.70)	(0.43 to 0.84)	0.71 ± 0.10
	MAE	0.43 ± 0.18	0.70 ± 0.11	0.55 ± 0.15
	MAE	(0.06 to 0.72)	(0.43 to 0.85)	0.55 ± 0.15
Segmentation Regression+	RMSE	0.59 ± 0.05	0.78 ± 0.04	0.60 ± 0.09
	RMSE	(0.51 to 0.64)	(0.73 to 0.81)	0.60 ± 0.09
	MAE	0.57 ± 0.05	0.78 ± 0.04	0.41 ± 0.13
	MAE	(0.50 to 0.61)	(0.73 to 0.80)	0.41 ± 0.13
Markov- GAM	RMSE	0.39 ± 0.18	0.69 ± 0.11	0.81 ± 0.07
	RMSE	(0.00 to 0.64)	(0.45 to 0.86)	0.81 ± 0.07
	MAE	0.37 ± 0.22	0.69 ± 0.11	0.70 ± 0.11
	MAE	(−0.11 to 0.66)	(0.44 to 0.85)	0.70 ± 0.11
Markov GAM	RMSE	0.41 ± 0.20	0.69 ± 0.12	0.75 ± 0.10
	RMSE	(−0.09 to 0.63)	(0.37 to 0.84)	0.75 ± 0.10
	MAE	0.40 ± 0.22	0.70 ± 0.12	0.70 ± 0.11
	MAE	(−0.13 to 0.67)	(0.39 to 0.85)	0.70 ± 0.11
Markov GAM+	RMSE	0.58 ± 0.11	0.78 ± 0.08	0.59 ± 0.14
	RMSE	(0.43 to 0.67)	(0.67 to 0.84)	0.59 ± 0.14
	MAE	0.58 ± 0.11	0.79 ± 0.07	0.53 ± 0.17
	MAE	(0.44 to 0.67)	(0.68 to 0.85)	0.53 ± 0.17
Myelin Feature	-	-	0.67	-
Myelin Feature	-	-	(CI: 0.66 to 0.69)	-

Note: Bold highlights the best results. Segmentation Regression+: segmentation regression plus lesion masks; Markov- GAM: GAM based on T2 signal intensity alone; Markov+ GAM: Markov GAM plus lesion masks.