Table 3.

Visual interpretation of VBM maps: cross tables of the reader consensus versus the ground truth diagnoses for the differentiation between AD, FTLD, and normal, separately for each of the three different VBM methods

		Scanner-specific VBM			Multiple-scanner VBM			CNN-VBM
		AD	FTLD	Normal	AD	FTLD	Normal	AD	FTLD	Normal
Ground truth	AD	38	2	11	13	2	36	38	7	6
	FTLD	3	25	2	1	23	6	3	25	2
	Normal	0	0	37	0	0	37	3	1	33

AD Alzheimer’s disease, CNN convolutional neural network, CNN-VBM CNN-based VBM without reference to a normal database, FTLD frontotemporal lobar degeneration, multiple-scanner VBM conventional VBM with a mixed normal database comprising T1w-MRI images from multiple scanners as reference, scanner-specific VBM conventional VBM with a scanner- and sequence-specific normal database as reference, VBM voxel-based morphometry