help
help > RE: Visual Task related functional connectivity
Jul 4, 2015 01:07 AM | Alfonso Nieto-Castanon - Boston University
RE: Visual Task related functional connectivity
Dear Elveda,
The results and interpretation of your test (2) depend exactly on the meaning of a zero value of your motion covariate. In particular, the described test (selecting 'controls' and 'ControlsMotion', and entering a contrast [1 0]) will evaluate the estimated connectivity within the control group at the zero-level of the motion covariate. If the values entered into the second-level covariate ControlsMotion for the control group are centered (i.e. their mean is zero, in other words a zero value represents the average of your control group), then the results and interpretation will be most similar to your test (3) (i.e. the effect sizes will be identical in both tests, but the statistics will typically be slightly different, since test (2) further removes any covariation with the motion covariate from the residual error). So, summarizing, if you center the motion values entered into the ControlsMotion covariate you will get most similar results between your test (2) and (3) (otherwise the difference reflects the estimated group connectivity at different levels of your motion covariate)
Hope this helps
Alfonso
Originally posted by Elveda Yildirim:
The results and interpretation of your test (2) depend exactly on the meaning of a zero value of your motion covariate. In particular, the described test (selecting 'controls' and 'ControlsMotion', and entering a contrast [1 0]) will evaluate the estimated connectivity within the control group at the zero-level of the motion covariate. If the values entered into the second-level covariate ControlsMotion for the control group are centered (i.e. their mean is zero, in other words a zero value represents the average of your control group), then the results and interpretation will be most similar to your test (3) (i.e. the effect sizes will be identical in both tests, but the statistics will typically be slightly different, since test (2) further removes any covariation with the motion covariate from the residual error). So, summarizing, if you center the motion values entered into the ControlsMotion covariate you will get most similar results between your test (2) and (3) (otherwise the difference reflects the estimated group connectivity at different levels of your motion covariate)
Hope this helps
Alfonso
Originally posted by Elveda Yildirim:
Dear Alfonso,
Thank you very much for your help.I listed below what I have done.
1- Subject-effects: 'Patients', 'Controls', 'PatientsMotion', and 'ControlsMotion'; Between-subjects:[0 0 1 0],[ 0 0 0 1]. I didn't get any connectivity for selected regions. It means motion doesn't effect the connectivity.
2-Subject-effects: 'Controls' and 'ControlsMotion'; Between-subjects:[1 0] and I did the same thing for patients. What the results are present?
3-Subject-effects: 'Controls' ; Between-subjects:[1]. This is simply connectivity of control group. The results are different from #2. If #1 is true, motion doesn't effect the connectivity, why I had different results?
Could you please help me to clear my mind?
Thanks,
Elveda
Many thanks,
Elveda
Thank you very much for your help.I listed below what I have done.
1- Subject-effects: 'Patients', 'Controls', 'PatientsMotion', and 'ControlsMotion'; Between-subjects:[0 0 1 0],[ 0 0 0 1]. I didn't get any connectivity for selected regions. It means motion doesn't effect the connectivity.
2-Subject-effects: 'Controls' and 'ControlsMotion'; Between-subjects:[1 0] and I did the same thing for patients. What the results are present?
3-Subject-effects: 'Controls' ; Between-subjects:[1]. This is simply connectivity of control group. The results are different from #2. If #1 is true, motion doesn't effect the connectivity, why I had different results?
Could you please help me to clear my mind?
Thanks,
Elveda
Many thanks,
Elveda
Threaded View
Title | Author | Date |
---|---|---|
Elveda Yildirim | Jun 23, 2015 | |
Alfonso Nieto-Castanon | Jun 30, 2015 | |
Elveda Yildirim | Jul 2, 2015 | |
Alfonso Nieto-Castanon | Jul 4, 2015 | |
Elveda Yildirim | Jul 6, 2015 | |