Classification accuracy between patients with MHE and NHE using a single Bayesian network
| Voxelwise Threshold and Classification Network | Accuracy | Sensitivity | Specificity |
|---|---|---|---|
| MD alone | |||
| 0 SD above mean value | 78.5% | 75.9% | 80.6% |
| 0.25 SD above mean value | 80.0% | 75.9% | 83.3% |
| 0.50 SD above mean value | 75.4% | 58.6% | 88.9% |
| 0.75 SD above mean value | 81.5% | 65.5% | 94.4% |
| 1.0 SD above mean value | 80.0% | 62.1% | 94.4% |
| FA alone | |||
| 0 SD below mean value | 84.6% | 75.9% | 91.7% |
| 0.25 SD below mean value | 83.1% | 79.3% | 86.1% |
| 0.50 SD below mean value | 89.2% | 86.2% | 91.7% |
| 0.75 SD below mean value | 92.3% | 100% | 86.1% |
| 1.0 SD below mean value | 86.2% | 72.4% | 97.2% |