Table 3.
Thresholds of statistical significance used by the 338 surveyed studies reporting whole brain results.
| Inferential method | n | % |
|---|---|---|
| Cluster-level inference (n = 244) | ||
| Corrected p = 0.05 | 228 | 93.4 |
| Corrected p = 0.025 | 1 | 0.4 |
| Corrected p = 0.01 | 8 | 3.3 |
| Corrected p = 0.001 | 7 | 2.9 |
| Voxel-level inference (n = 71) | ||
| Corrected p = 0.05 | 67 | 94.4 |
| Corrected p = 0.025 | 1 | 1.4 |
| Corrected p = 0.01 | 1 | 1.4 |
| Corrected p = 0.005 | 1 | 1.4 |
| Corrected p = 0.001 | 1 | 1.4 |
| Threshold free cluster enhancement (n = 7) | ||
| Corrected p = 0.05 | 7 | 100.0 |
| Uncorrected inference (n = 16) | ||
| p = 0.05, k = 40 | 1 | 6.3 |
| p = 0.005, k = 50 | 1 | 6.3 |
| p = 0.005, k = 20 | 1 | 6.3 |
| p = 0.005, k = 10 | 1 | 6.3 |
| p = 0.005 | 1 | 6.3 |
| p = 0.001, k = 20 | 4 | 25.0 |
| p = 0.001, k = 15 | 1 | 6.3 |
| p = 0.001, k = 10 | 3 | 18.8 |
| p = 0.001, k = 5 | 1 | 6.3 |
| p = 0.001 | 2 | 12.5 |
k means the minimal cluster size expressed in number of voxels.