Table 4.

Comparison of results obtained using native volumes derived from label resampling and scaling factor correction.

	Native manual segmentation volumes based on resampling of labels (as presented in the article)		Native manual segmentation volumes based on calculation of scaling factors
	FreeSurfer	FSL-FIRST	FreeSurfer	FSL-FIRST
% volume diff. (SD)
L-hippocampus	60.38 (13.04)	27.61(14.49)	60.16 (12.90)	27.43 (14.28)
R-hippocampus	51.53 (13.17)	28.39 (13.07)	51.41 (13.09)	28.28 (12.89)
L-amygdala	100.29 (24.56)	50.32 (27.65)	99.53 (24.27)	49.76 (27.55)
R-amygdala	93.56 (25.78)	40.29 (26.09)	93.11 (24.56)	40.03 (25.83)
PCC with manual seg.
L-hippocampus	0.77	0.51	0.77	0.52
R-hippocampus	0.69	0.59	0.70	0.60
L-amygdala	0.67	0.31	0.66	0.30
R-amygdala	0.61	0.35	0.62	0.35
ICC with manual seg.
L-hippocampus	0.74	0.51	0.74	0.52
R-hippocampus	0.68	0.59	0.69	0.60
L-amygdala	0.65	0.30	0.64	0.28
R-amygdala	0.60	0.33	0.61	0.33

Percentage of volume difference (% volume diff.), Pearson correlation coefficients (PCC) and intraclass correlation coefficients (ICC) computed between manual segmentation volumes and automatic protocols. Results are presented with native manual segmentation volumes obtained by resampling labels in the native space using an inversion of the linear transformation (left column) and with manual volumes obtained by dividing volumes segmented in the standard space by scale factors associated with the linear transformation in the x,y,z directions (right column). This table shows that both methods of estimating manual segmentation volumes in the native space lead to highly similar results. L — left. R — right. SD— standard deviation.