Print

---

It is suggested to always correct for multiple comparisons, and it is rare to be able to publish without it. If you want to be less conservative you could use FWE of .10, though It depends quite a lot on the type of data and analysis you are performing. 

Hope this helps

Eduardo

El sábado, 9 de abril de 2016, Carole Guedj <[log in to unmask]> escribió:

Dear FLS community,

I used the FSLnets script to assess changes in resting-state networks interactions between 2 conditions.
If I understand well (and please, fell free to correct me if I’m wrong), the glm_net script returns a plot with the result of permutation testing:
        - with significant p-values above the diagonal and,
        - significant + FWE corrected p-values below the diagonal.

I’m wondering if it is legitimate to use the uncorrected p-values as a result (i.e. the result shown above the diagonal). Indeed, I saw one paper using this analysis (i.e. FSLnets with randomise) without correction for multiple testing such as FWE correction.

Thank you in advance for your help !