Table 2.
Results of simulation study to determine the optimal bandwidth based on the signal size.
| Signal Size | MTPs | Bandwidth | pValue.thresh | True positive | TPRate | False positive | FPRate |
|---|---|---|---|---|---|---|---|
| Small | BH | 5 | 0.034 | 101 | 0.902 | 35 | 0.000 |
| Small | WY | 5 | 0.050 | 101 | 0.902 | 35 | 0.000 |
| Medium | BH | 10 | 3 × 10–4 | 1411 | 0.953 | 235 | 0.001 |
| Medium | WY | 10 | 1 × 10–3 | 1442 | 0.974 | 411 | 0.001 |
| Large | BH | 15 | 0.001 | 13,219 | 0.900 | 65,291 | 0.173 |
| Large | WY | 15 | 0.279 | 13,234 | 0.901 | 65,762 | 0.174 |
| Large | BH | 19 | 7.78 × 10–7 | 13,419 | 0.914 | 45,718 | 0.121 |
| Large | WY | 19 | 1 × 10–5 | 13,584 | 0.925 | 51,627 | 0.137 |
MTP, multiple comparison procedure; pValue.thresh, p value considered significant change for a voxel; True Positive, number of true positives detected; TPRate, true positive rate; False Positive, the number of false positives detected; and FPRate, the false positive rate.