Table 7.
Significance levels when testing the positivity or negativity of the region-wise mean using permutation tests for different sequence types
| P1 | P2 | Maximum T | Minimum T | T threshold | |
|---|---|---|---|---|---|
| SPGR BC+N3 | 0.394 | 0.008 | 4.42 | −6.45 | ±7.35 |
| SPGR BC−N3 | 0.471 | 0.015 | 5.55 | −8.33 | ±7.3 |
| MP-RAGE BC+N3 | 0.362 | 0.046 | 9.90 | −7.30 | ±7.03 |
| MP-RAGE BC−N3 | 0.976 | 0.301 | 4.93 | −6.49 | ±6.90 |
| IR-SPGR BC+N3 | 0.173 | 0.038 | 7.49 | −9.99 | ±10.58 |
| IR-SPGR BC−N3 | 0.992 | 0.303 | 5.54 | −8.96 | ±10.8 |
| SYN BC+N3 | 0.827 | 0.039 | 5.40 | −6.02 | ±7.29 |
| SYN BC−N3 | 0.772 | 0.021 | 5.28 | −6.22 | ±7.28 |
P1 (P2) is the significance level of the alternative hypothesis that the region-wise mean is positive (negative), while Max (Min) T is the maximal (minimal) voxelwise T statistic inside the ROI defined by a mask of the average brain template (ICBM-53) in ICBM space. The T threshold is the corrected cut-off T value obtained using 10,000 permutations on the maximum/minimum statistic (at the P = 0.05 level). Voxels with T values more extreme than the corrected cut-off T threshold are considered active (since we should detect no change within a 2-week period, these voxels are considered false positives under the null hypothesis).