Table 3.
Accuracy (%), sensitivity (%) and specificity (%) of each modality and the multimodal combination using different training sets for the classification of AD pathology (discrimination of pNC, pMCI and sAD from sNC).
| FDG-PET (Metabolism) | T1-MRI (Volume) | Multimodal (Metabolism + Volume) | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Acc. | Sens. | Spec. | Acc. | Sens. | Spec. | Acc. | Sens. | Spec. | |
| Training Set | |||||||||
| sNC vs. sAD | 84.5 (1.4) | 79.9 (1.6) | 91.9 (6.9) | 81.9 (1.2) | 75.5 (1.3) | 92.3 (4.9) | 84.6 (1.5) | 80.2 (2.0) | 91.8 (6.8) |
| sNC vs. (pMCI and sAD) |
85.5 (2.8) | 85.0 (2.9) | 86.2 (10.1) | 82.8 (3.4) | 79.8 (4.1) | 87.7 (6.3) | 86.0 (2.5) | 85.7 (3.2) | 86.5 (8.6) |
| sNC vs. (pNC, pMCI and sAD) |
85.9 (4.8) | 85.6 (3.8) | 86.3 (7.8) | 82.5 (5.2) | 80.2 (7.6) | 86.1 (7.6) | 86.4 (4.7) | 86.5 (5.2) | 86.3 (8.6) |
The numbers in each cell are the average value and standard deviation from the 10-fold cross validation experiments. Sensitivity is the fraction of correctly classified pNC, pMCI and sAD images, specificity is the fraction of correctly classified sNC images and accuracy is correctly classified images taken together.