statistic problem

[Fox Zhong]

Hi Frank

I have performed the automatic fiber tracking and obtained detailed tract information with 33 metric values. I am trying to find the group difference between 4 groups. However, after I performed correlational tractography, I found the results are different between the two analyses. I use text tract information for data analysis and group comparison through multiple steps, including 1. Kruskal-Wallis tests on each metric to detect significant differences between groups 2. Calculate and store raw p-values 3. Apply Benjamini-Hochberg method for FDR correction to control false positive rate 4. Determine significance based on adjusted p-values (threshold of 0.05) 5. Conduct pairwise comparisons (Mann-Whitney U tests) for metrics with significant Kruskal-Wallis results 6. Apply Bonferroni correction to post-hoc test p-values (multiply p-values by the number of comparisons, 7 . Identify which group differences are substantial (adjusted p < 0.05). Although some results overlapped, there were some differences, as I know that correlational tractography uses regression model rather than post hoc and group comparison.

Best

Yiwen Zhong