Hello everyone,
I conducted seed-based analyses using functionally defined seed regions from previous studies. My hypothesis is that despite the proximity of the seeds (smaller divisions of a PFC region), distinct connectivity patterns will emerge for each seed. The analyses, performed in conn, revealed distinct clusters at the group level. However, there are similarities in clusters across different seed analyses.
My follow-up question is: What is the optimal method for quantitatively comparing results from the four seed analyses? Can this be done in conn?
Current setup: I ran the analysis individually for each seed. This produced the cluster results. Now I need to compare the 4 results to identify whether some seeds yield more similar clusters and and whether there are unique connections that identify with a specific one compared to the others.
Any guidance on this matter would be greatly appreciated.
Best Kelly
Hi Kelly,
Assuming that you have already defined a SBC first-level analysis that includes the four seeds, in the second-level tab you can simply select all four seeds under the 'Sources' list, and then in the 'between-sources contrast' field simply select the option that reads 'any differences (F-test)' (or enter manually the contrast [-1 1 0 0; 0 -1 1 0; 0 0 -1 1]) in order to evaluate any differences between the four seeds (for these analyses typically you would also select the 'AllSubjects' term in the 'subjects' list, and the 'rest' condition in the 'conditions' list, unless you want to evaluate some higher order interaction).
Hope this helps
Alfonso
Originally posted by akc23:
Hello everyone,
I conducted seed-based analyses using functionally defined seed regions from previous studies. My hypothesis is that despite the proximity of the seeds (smaller divisions of a PFC region), distinct connectivity patterns will emerge for each seed. The analyses, performed in conn, revealed distinct clusters at the group level. However, there are similarities in clusters across different seed analyses.
My follow-up question is: What is the optimal method for quantitatively comparing results from the four seed analyses? Can this be done in conn?
Current setup: I ran the analysis individually for each seed. This produced the cluster results. Now I need to compare the 4 results to identify whether some seeds yield more similar clusters and and whether there are unique connections that identify with a specific one compared to the others.
Any guidance on this matter would be greatly appreciated.
Best Kelly
Alfoson,
Thank you extremely helpful
As a followup- does conn perform something like a post hoc analysis on the result from the any differences (F-test)?