Table 3.
Results of p‐MCI versus s‐MCI classification using single atlas and multiple atlases
| Method | ACC (%) | SEN (%) | SPE (%) | AUC |
|---|---|---|---|---|
| PC_SA | 68.49 | 67.80 | 69.10 | 0.6285 |
| COMPARE_SA | 70.06 | 68.08 | 72.02 | 0.6356 |
| PC_MA | 72.78 | 74.62 | 70.91 | 0.7245 |
| COMPARE_MA | 73.35 | 75.76 | 70.83 | 0.7405 |
| RS | 69.05 | 68.10 | 72.94 | 0.6912 |
| Lasso | 75.32 | 81.36 | 69.17 | 0.7602 |
| Proposed method | 78.88 | 85.45 | 76.06 | 0.8069 |
Note: PC_SA and COMPARE_SA denote PC and COMPARE using single atlas, respectively; PC_MA and COMPARE_MA mean PC and COMPARE using multiple atlases, respectively; RS represents random subspace; ACC denotes accuracy; SEN means sensitivity; SPE represents specificity, and AUC denotes the area under ROC curve.