Statistical methods


Are the statistical methods used valid?

Reviewer Comment:

No. While the statistical methods used are reasonable thresholds are quite liberal for a resting state imaging paper

Reply to this comment:

As reviewer mentioned, we also have paid much attention to this issue. In present study, we used the new version of DPABI software, which have fixed the bug of AlphaSim method reported in Eklund et al. (PNAS 2016; 113: 7900-7905) according to the study by Chen at al. [1]. In the revised manuscript, all results were recalculated according to both reviewers’ suggestions. The significance level of all result was set at p < 0.01in revision. There is no statistically significant difference between the two groups when the significance level was set at p < 0.001 (AlphaSim corrected). Before we used this correctional method, we had tried more rigorous method of FDR or FWE, but there was no significant result, which may result from our relatively small sample or significance of the difference.

[1]Chen X, Lu B, Yan C G. Reproducibility of R-fMRI metrics on the impact of different strategies for multiple comparison correction and sample sizes[J]. Human Brain Mapping, 2017.

Reviewer Comment: The PNAS paper recommends as far as I am aware that a minimum safe p value after correction is p<0.001. If your results can only reach significance at p<0.01 this means there is a high risk of false positives

老师您好, 最近投文章审稿人提了个统计问题,不知道怎么回比较好,麻烦您帮我看看,非常感谢!