Table 2.
Comparison of performances of five different methods on Experiment 1. The reported values are the correlation coefficient (for MMSE and ADAS-Cog regression) and accuracy (for AD vs. HC and MCI vs. HC classification), averaged on 10-fold tests (with standard deviation also reported).
Methods | Correlation coefficient | Classification accuracy | ||
---|---|---|---|---|
MMSE | ADAS-Cog | AD vs. HC | MCI vs. HC | |
MRI-based | 0.504 ± 0.038 | 0.609 ± 0.014 | 0.848 ± 0.026 | 0.739 ± 0.028 |
PET-based | 0.658 ± 0.027 | 0.670 ± 0.018 | 0.845 ± 0.035 | 0.797 ± 0.023 |
CSF-based | 0.465 ± 0.019 | 0.474 ± 0.013 | 0.805 ± 0.022 | 0.536 ± 0.044 |
CONCAT | 0.658 ± 0.023 | 0.695 ± 0.011 | 0.920 ± 0.033 | 0.800 ± 0.024 |
Proposed M3T | 0.697 ± 0.022 | 0.739 ± 0.012 | 0.933 ± 0.022 | 0.832 ± 0.015 |
AD = Alzheimer’s Disease, HC = Healthy Control, MCI = Mild Cognitive Impairment, MMSE = Mini-Mental State Examination, ADAS-Cog = Alzheimer’s Disease Assessment Scale - Cognitive Subscale