Table 2. Discrimination accuracy of four kinds of measures.
Accuracy (mean/std) | p-value | Specificity | Sensitivity | AUC | |
---|---|---|---|---|---|
fMRI+sMRI+DTI | a: 86.52% / 6% | <0.01 | 87.34% | 85.98% | 0.92 |
b: 84.96% / 6.8% | <0.01 | 84.88% | 85.26% | 0.91 | |
fMRI+sMRI | a: 85.86% / 6.4% | <0.01 | 85.81% | 85.98% | 0.92 |
b: 84.61% / 6.7% | <0.01 | 84.26% | 84.57% | 0.91 | |
fMRI+DTI | a: 83.94% / 6.8% | <0.01 | 84% | 82.44% | 0.9 |
b: 81.57% / 6.7% | <0.01 | 81.43% | 81.65% | 0.89 | |
fMRI | a: 84.47% / 6.9% | <0.01 | 84.6% | 84.98% | 0.91 |
b: 81.43% / 6.7% | <0.01 | 81.83% | 80.85% | 0.89 | |
sMRI+DTI | 83.16% / 6.12% | <0.01 | 83.48% | 82.27% | 0.9 |
sMRI | 83.94% / 6.21% | <0.01 | 83.7% | 84.3% | 0.9 |
DTI | 65.76% / 8.1% | 0.01 | 75% | 56.52% | 0.7 |
Note: a: all components of fMRI; b: first 5 principle components of fMRI; AUC: area under ROC curve.