help
488 Subscribers 
help > gPPI values (from a contrast/cond) to behavior correlation
Showing 1-2 of 2 posts
Apr 27, 2022 11:04 AM | Panagiotis Iliopoulos
gPPI values (from a contrast/cond) to behavior correlation
Dear Alfonso / CONN users,
I have a pair of significant connections (in a contrast cond. B > A, task gPPI connectivity). I am interested to see whether the gPPI ROI-to-ROI connectivity values correlate to behavior (e.g. False alarms).
[see attached pic from the 2nd level viewer for overview these results]
I am not sure how I can proceed here.
My understanding says to get the 1st level gPPI values (must be the BETA values/map image files located at the 1st level folder created by CONN) and correlate these to the behavioral variable.
Q:
1) how I can link these maps to a behavioral variable?
(these BETA 1st level values/maps will need to be inserted in SPM? is it possible to do this analysis inside CONN?); not sure overall how to do this.
1b) how I can get the contrast of conditions (B minus A) values to do this? (instead of just condB or condA that exist in that 1st level folder)
2) These 1st level BETA maps are based on seed-to-voxel connectivity (not ROI-to-ROI) from what I understand. Where does CONN store the ROI-to-ROI connectivity values (see my results) to use those to do brain-behavior correlation ?
Many thanks in advance :)
I have a pair of significant connections (in a contrast cond. B > A, task gPPI connectivity). I am interested to see whether the gPPI ROI-to-ROI connectivity values correlate to behavior (e.g. False alarms).
[see attached pic from the 2nd level viewer for overview these results]
I am not sure how I can proceed here.
My understanding says to get the 1st level gPPI values (must be the BETA values/map image files located at the 1st level folder created by CONN) and correlate these to the behavioral variable.
Q:
1) how I can link these maps to a behavioral variable?
(these BETA 1st level values/maps will need to be inserted in SPM? is it possible to do this analysis inside CONN?); not sure overall how to do this.
1b) how I can get the contrast of conditions (B minus A) values to do this? (instead of just condB or condA that exist in that 1st level folder)
2) These 1st level BETA maps are based on seed-to-voxel connectivity (not ROI-to-ROI) from what I understand. Where does CONN store the ROI-to-ROI connectivity values (see my results) to use those to do brain-behavior correlation ?
Many thanks in advance :)
Apr 28, 2022 10:04 PM | Alfonso Nieto-Castanon - Boston University
RE: gPPI values (from a contrast/cond) to behavior correlation
Dear Panagiotis,
In the results explorer window (your attached image) you may click on 'import values' to import into CONN the gPPI values for each significant subject/condition/connection separately as a series of new second-level covariates. Then you can go to Tools.Calculator to compute the association between those values and your behavioral variable. For example if in your Setup.Covariates 2nd-level you have something like the following variables (after the 'import values' step):
- cluster 1/1 connection X during taskA
- cluster 1/1 connection X during taskB
- cluster 1/1 connection Y during taskA
- cluster 1/1 connection Y during taskB
and you also have imported there a variable 'behaviorScores' with the behavioral scores for each subject, then in the Tools.Calculator tab you may:
a) select in the 'subject effects' list the variables 'AllSubjects' and 'behaviorScores', and enter a between-subjects contrast [0,1]
and b) select in the 'measures' list the variables 'cluster 1/1 connection X during taskA' and 'cluster 1/1 connection X during taskB', and enter a between-measures contrast [-1 1]
and that will compute the correlation between the taskB>taskA differences and the behavioral scores across subjects.
Hope this helps
Alfonso
Originally posted by Panagiotis Iliopoulos:
In the results explorer window (your attached image) you may click on 'import values' to import into CONN the gPPI values for each significant subject/condition/connection separately as a series of new second-level covariates. Then you can go to Tools.Calculator to compute the association between those values and your behavioral variable. For example if in your Setup.Covariates 2nd-level you have something like the following variables (after the 'import values' step):
- cluster 1/1 connection X during taskA
- cluster 1/1 connection X during taskB
- cluster 1/1 connection Y during taskA
- cluster 1/1 connection Y during taskB
and you also have imported there a variable 'behaviorScores' with the behavioral scores for each subject, then in the Tools.Calculator tab you may:
a) select in the 'subject effects' list the variables 'AllSubjects' and 'behaviorScores', and enter a between-subjects contrast [0,1]
and b) select in the 'measures' list the variables 'cluster 1/1 connection X during taskA' and 'cluster 1/1 connection X during taskB', and enter a between-measures contrast [-1 1]
and that will compute the correlation between the taskB>taskA differences and the behavioral scores across subjects.
Hope this helps
Alfonso
Originally posted by Panagiotis Iliopoulos:
Dear Alfonso / CONN users,
I have a pair of significant connections (in a contrast cond. B > A, task gPPI connectivity). I am interested to see whether the gPPI ROI-to-ROI connectivity values correlate to behavior (e.g. False alarms).
[see attached pic from the 2nd level viewer for overview these results]
I am not sure how I can proceed here.
My understanding says to get the 1st level gPPI values (must be the BETA values/map image files located at the 1st level folder created by CONN) and correlate these to the behavioral variable.
Q:
1) how I can link these maps to a behavioral variable?
(these BETA 1st level values/maps will need to be inserted in SPM? is it possible to do this analysis inside CONN?); not sure overall how to do this.
1b) how I can get the contrast of conditions (B minus A) values to do this? (instead of just condB or condA that exist in that 1st level folder)
2) These 1st level BETA maps are based on seed-to-voxel connectivity (not ROI-to-ROI) from what I understand. Where does CONN store the ROI-to-ROI connectivity values (see my results) to use those to do brain-behavior correlation ?
Many thanks in advance :)
I have a pair of significant connections (in a contrast cond. B > A, task gPPI connectivity). I am interested to see whether the gPPI ROI-to-ROI connectivity values correlate to behavior (e.g. False alarms).
[see attached pic from the 2nd level viewer for overview these results]
I am not sure how I can proceed here.
My understanding says to get the 1st level gPPI values (must be the BETA values/map image files located at the 1st level folder created by CONN) and correlate these to the behavioral variable.
Q:
1) how I can link these maps to a behavioral variable?
(these BETA 1st level values/maps will need to be inserted in SPM? is it possible to do this analysis inside CONN?); not sure overall how to do this.
1b) how I can get the contrast of conditions (B minus A) values to do this? (instead of just condB or condA that exist in that 1st level folder)
2) These 1st level BETA maps are based on seed-to-voxel connectivity (not ROI-to-ROI) from what I understand. Where does CONN store the ROI-to-ROI connectivity values (see my results) to use those to do brain-behavior correlation ?
Many thanks in advance :)