Table 6.
Classification performance of different methods to distinguish different stages of MCI.
| Article | Method | Cohort | ACC (%) | SEN (%) | SPE (%) | AUC |
|---|---|---|---|---|---|---|
| This paper | Proposed | EMCI/LMCI (33/29) | 83.87 | 86.21 | 81.21 | 0.905 |
| Biao Jie ( 26 ) | Spatio-temporal interaction patterns of dynamic connectivity networks | EMCI/LMCI (56/43) | 78.8 | 74.4 | 82.1 | 0.783 |
| Seyed Hani Hojjatia ( 52 ) | Graph theory and machine learning approach (mRMR, FS) | MCI-C/MCI-NC(18/62) | 91.4 | 83.24 | 90.1 | N/A |
| Mohammed Goryawala ( 58 ) | fMRI volumes and neuropsychological scores | EMCI/LMCI (114/91) | 73.6 | 74.3 | 72.7 | N/A |
| Heung-Il Suk ( 59 ) | 93 features from a MR image and the same dimensional features from a FDG-PET image. | MCI-C/MCI-NC (43/56) | 74.04 | 58 | 82.67 | 0.696 |
| Zhang and Shen ( 60 ) | MRI, PET and cognitive scores, Leave-one-out cross-validation |
MCI-C/MCI-NC (38/50) | 78.4 | 79.0 | 78.0 | 0.768 |
| Moradi et al. ( 61 ) | MRI, age and cognitive measures 10-fold cross-validation |
sMCI/pMCI (100/164) | 81.72 | 86.65 | 73.64 | 0.902 |
| Ardekani et al. ( 62 ) | Hippocampal volumetric integrity (HVI) from structural MRI scans RF with 5,000 trees |
sMCI/pMCI (78/86) | 82.3 | 86.0 | 78.2 | N/A |
The best multivariate predictors of MCI conversion are shown for each study. ACC, accuracy; SEN, sensitivity; SPE, specificity; AUC, area under the curve; FDG-PET, fluorodeoxyglucose positron emission tomography; RF, Random forest.