help
help > Correlation analysis in NBS
Aug 5, 2020 01:08 PM | giusy olivito
Correlation analysis in NBS
Hi everyone, I need your help/suggestion for the correlation
analysis in NBS. I created my connectivity matrices for a patient
sample (n 48) and I want to run a correlation analysis to check the
association between internodal functional connectivity within the
a-priori created network ( 5 rois) and some behavioral tests. I
selected one -sample t-test statistical model, and entered my
design matrix as follows:
1 23
1 43
1 25
1 34
1 23
1 56
and so on...
So my correlations with performance scores of 3 different tests were based on one-sample t-test, using 5000 permutations and setting the significant p-value at 0.05 (FDR method). I found some significant results involving all the 5 nodes of the networks for all the tests, but, surprisingly, the t value was the same for each edge across the three tests used.
My answer is...did something not work? it seems unlikely that the t-scores are exactly the same...
1 23
1 43
1 25
1 34
1 23
1 56
and so on...
So my correlations with performance scores of 3 different tests were based on one-sample t-test, using 5000 permutations and setting the significant p-value at 0.05 (FDR method). I found some significant results involving all the 5 nodes of the networks for all the tests, but, surprisingly, the t value was the same for each edge across the three tests used.
My answer is...did something not work? it seems unlikely that the t-scores are exactly the same...
Threaded View
Title | Author | Date |
---|---|---|
giusy olivito | Aug 5, 2020 | |
Andrew Zalesky | Aug 6, 2020 | |