Table 1.

Recently published recommended statistical practices for controlling false positives.

Publication name	Recommendations
Woo et al., 2014	1. Set the default cluster-defining primary threshold (CDT) at p < 0.001. 2. Use a stringent CDT or voxelwise inference for highly powered studies.
Eklund et al., 2016	1. The parametric method works well for voxelwise inferences but not for clusterwise inferences (unless a stringent CDT is set at p < 0.001). 2. The permutation method works well for both voxelwise and clusterwise inferences.
Roiser et al., 2016	1. For clusterwise inferences, choose a stringent CDT (e.g., p < 0.001) unless the permutation method was employed. 2. For voxelwise inferences, p-values should be corrected for multiple comparisons. 3. Complementary approaches, such as false-discovery rate or threshold-free cluster enhancement, can be considered. 4. Preregister the proposed studies in which the planned statistical analyses methods are documented clearly.
Carter et al., 2016	1. Studies investigating very small brain regions should use a high voxel threshold (e.g., p < 0.001). 2. Studies not targeting precise localization may consider a more liberal threshold and focus on controlling false negatives by data reduction (e.g., region-of-interest analyses), as studies with fewer than 50 subjects per group usually have limited power.