Table 2. Comparison of performance of different methods in predicting 24-month (M24) clinical scores of MCI patients, by using different numbers of longitudinal data.
Method | MMSE | ADAS-Cog | ||
CORR | RMSE | CORR | RMSE | |
CONCAT (bl) | 0.635±0.049 | 2.541±0.100 | 0.657±0.038 | 4.771±0.146 |
Ensemble (bl) | 0.666±0.039 | 2.637±0.081 | 0.677±0.037 | 4.943±0.155 |
Proposed (bl) | 0.659±0.043 | 2.457±0.123 | 0.682±0.044 | 4.763±0.084 |
Proposed (bl+M06) | 0.702±0.037 | 2.344±0.097 | 0.746±0.028 | 4.318±0.173 |
Proposed (bl+M06+M12) | 0.743±0.024 | 2.177±0.113 | 0.768±0.025 | 4.097±0.126 |
Proposed (bl+M06+M12+M18) | 0.786±0.013 | 2.035±0.076 | 0.777±0.027 | 4.004±0.086 |
The reported values are the correlation coefficient (CORR) and the root-mean-square error (RMSE), averaged on 10-fold tests (with standard deviation also reported).
MMSE = Mini-Mental State Examination, ADAS-Cog = Alzheimer's Disease Assessment Scale - Cognitive Subscale.