help
help > RE: Running contrasts in CONN
Mar 8, 2023 12:03 AM | Alfonso Nieto-Castanon - Boston University
RE: Running contrasts in CONN
Hi Jose,
Yes, typically to see what is driving those effects you may simply click on the 'plot effects' button (that will show, within each cluster, the average connectivity -corrected at the zero level of your control covariates- within each of your three groups).
If you prefer to run directly a post-hoc analysis looking for example at differences between the first two groups within each one of your significant clusters, that can also be done, for example, by using the following procedure:
1) in the results explorer window of your original analysis click on 'export mask' to create a mask defining these significant clusters
2) In CONN's GUI Setup.Results (2nd-level) tab, define and run a new second-level analyses looking at the desired group differences (e.g. with a contrast [1 -1 0 0 0 0 0 0 0 0 0 0])
3) in the results explorer window of this new analysis, select the 'plot effects' button and change the option that reads 'clusters of interest in current analysis' to 'other clusters of interest/ROIs (select exported mask/ROI file)'. Then select the *.ROIs.nii file that was created in step (1) and click 'Ok' to compute those post-hoc analyses on each cluster from your original analysis.
Hope this helps
Alfonso
Originally posted by Jose Maximo:
Yes, typically to see what is driving those effects you may simply click on the 'plot effects' button (that will show, within each cluster, the average connectivity -corrected at the zero level of your control covariates- within each of your three groups).
If you prefer to run directly a post-hoc analysis looking for example at differences between the first two groups within each one of your significant clusters, that can also be done, for example, by using the following procedure:
1) in the results explorer window of your original analysis click on 'export mask' to create a mask defining these significant clusters
2) In CONN's GUI Setup.Results (2nd-level) tab, define and run a new second-level analyses looking at the desired group differences (e.g. with a contrast [1 -1 0 0 0 0 0 0 0 0 0 0])
3) in the results explorer window of this new analysis, select the 'plot effects' button and change the option that reads 'clusters of interest in current analysis' to 'other clusters of interest/ROIs (select exported mask/ROI file)'. Then select the *.ROIs.nii file that was created in step (1) and click 'Ok' to compute those post-hoc analyses on each cluster from your original analysis.
Hope this helps
Alfonso
Originally posted by Jose Maximo:
Hi,
After running an ancova with 3 groups (group1 group 2 and group3 and 3 covariates (age1 age2 age3 sex1 sex2 sex3 motion1 motion2 motion3) [1 -1 0 0 0 0 0 0 0 0 0 0; 0 1 -1 0 0 0 0 0 0 0 0 0] I get some significant clusters. I extract data from these clusters and want to test for group differences (what is driving this omnibus effect) in SPSS, but I get no sig. results. Is there a way to do this within conn and not extracting the data and doing it in some statistics program??
After running an ancova with 3 groups (group1 group 2 and group3 and 3 covariates (age1 age2 age3 sex1 sex2 sex3 motion1 motion2 motion3) [1 -1 0 0 0 0 0 0 0 0 0 0; 0 1 -1 0 0 0 0 0 0 0 0 0] I get some significant clusters. I extract data from these clusters and want to test for group differences (what is driving this omnibus effect) in SPSS, but I get no sig. results. Is there a way to do this within conn and not extracting the data and doing it in some statistics program??
Threaded View
| Title | Author | Date |
|---|---|---|
| Jose Maximo | Mar 3, 2023 | |
| Alfonso Nieto-Castanon | Mar 8, 2023 | |