help > RE: combining ROIs and one side doesn't pass FDR
Sep 19, 2014  02:09 AM | Mary Newsome
RE: combining ROIs and one side doesn't pass FDR
I apologize for a potentially confusing typo. In the first sentence below I meant to say, "to test connectivity with either L BA 8 or R BA 8"--BA 9 should not appear in it.
Originally posted by Mary Newsome:
Dear Alfonso,

I combined two ROIs (L BA 8 + R BA 8) and ran an overall F-test (thanks to extremely helpful previous advice from you) to test connectivity with either L BA 9 or R BA 8. The results yielded 4 clusters. When I ran post-hocs, there were 8 results (L and R for each cluster), and one of the clusters was NS for the right side only (the left side did not pass FDR correction as reported by Rex when using the Explore Clusters button).
 
My question is about reporting the significant areas. For the other clusters that passed FDR on both sides, I just refer back to the gyri/BA numbers reported in the gui generated by the Seed to Voxel Results Explorer (e.g., X voxels covering 6% of BA40, Supramarginal gyrus). I'm not sure what to do about the cluster in which only one side was significant, though (as that gui is based on both sides). Would you be able to help me tease out the gyri/BA numbers that are significant for just the right side?
 
Also, could you confirm that in the Rex/Explore Clusters output, that "m1" refers to the left ROI and "m2" refers to the right ROI in labels such as m1.results.ROIs.+4 -14 +10 and m2.results.ROIs.+4 -14 +10?
 
Thanks so much!
Mary

Threaded View

TitleAuthorDate
Mary Newsome Sep 17, 2014
Alfonso Nieto-Castanon Sep 22, 2014
Mary Newsome Sep 22, 2014
Alfonso Nieto-Castanon Sep 23, 2014
Mary Newsome Sep 23, 2014
RE: combining ROIs and one side doesn't pass FDR
Mary Newsome Sep 19, 2014