Table 2. Classification performance using different types of feature.
| Feature types | ACC (%) | SEN (%) | SPE (%) | AUC |
|---|---|---|---|---|
| ALFF | 65.31 | 85.71 | 38.10 | 0.69 |
| ReHo | 67.35 | 71.43 | 61.90 | 0.67 |
| RFCS | 63.27 | 82.14 | 38.10 | 0.68 |
| GM | 71.43 | 85.71 | 52.38 | 0.83 |
| ALFF+ReHo | 69.39 | 82.14 | 52.38 | 0.70 |
| ALFF+RFCS | 64.58 | 85.71 | 33.33 | 0.54 |
| ALFF+GM | 70.83 | 89.29 | 42.86 | 0.74 |
| ReHo+GM | 72.92 | 85.71 | 52.38 | 0.75 |
| ReHo+RFCS | 71.43 | 82.14 | 57.14 | 0.75 |
| RFCS+GM | 75.00 | 92.86 | 47.62 | 0.78 |
| ALFF+ReHo+RFCS | 72.92 | 85.71 | 52.38 | 0.75 |
| ALFF+ReHo+GM | 75.51 | 89.29 | 57.14 | 0.78 |
| ALFF+RFCS+GM | 79.59 | 89.29 | 66.67 | 0.84 |
| ReHo+RFCS+GM | 73.47 | 85.71 | 57.14 | 0.71 |
| Concatenation | 67.35 | 78.57 | 52.38 | 0.74 |
| M3 method | 73.47 | 66.67 | 78.57 | 0.82 |
| Proposed | 83.67 | 92.86 | 71.43 | 0.83 |
SEN = sensitivity, SPE = specificity, ACC = accuracy, AUC = area under receive operating characteristic curve. “+” indicates combination of the given types of features; “Concatenation” means all four types of feature were concatenated into a long feature vector.