Table 3.

The classification results for the DSI calculated by using bootstrapping for each case separately, excluding the case being tested from the randomized training sets

	C vs. FTD	C vs. AD	AD vs. FTD all	AD vs. FTD autopsy	AD vs. FTD clinical
AUC	0.99±0.01	1.00±0.00	0.89±0.01	0.97±0.02	0.94±0.02
Accuracy	0.95±0.02	0.98±0.01	0.83±0.02	0.91±0.04	0.89±0.03
Sensitivity	0.92±0.03	0.98±0.01	0.81±0.03	0.92±0.05	0.81±0.05
Specificity	0.99±0.02	1.00±0.01	0.83±0.03	0.90±0.06	0.93±0.03

The means and standard deviations are calculated from each round of bootstrapping (n = 100) for all tested cases. DSI classification results calculated with bootstrapping. Values are expressed as mean ± standard deviation.