help
help > RE: guidelines for including effects of condition
Apr 6, 2015 04:04 AM | Alfonso Nieto-Castanon - Boston University
RE: guidelines for including effects of condition
Hi Patrick,
I would include the "effects of condition" regressors as confounds during the Denoising step for any connectivity analysis type, including PPI. PPI focuses on the interaction between the psychological (task) term and the physiological (BOLD) term, after controlling their main effects (the "main psychological" term in PPI is the same as the "effect of condition" regressor in CONN). This means that in practice those effects will be effectively removed during PPI estimation irrespective of whether you have already removed them during the Denoising step or not.
Hope this helps
Alfonso
Originally posted by Patrick McConnell:
I would include the "effects of condition" regressors as confounds during the Denoising step for any connectivity analysis type, including PPI. PPI focuses on the interaction between the psychological (task) term and the physiological (BOLD) term, after controlling their main effects (the "main psychological" term in PPI is the same as the "effect of condition" regressor in CONN). This means that in practice those effects will be effectively removed during PPI estimation irrespective of whether you have already removed them during the Denoising step or not.
Hope this helps
Alfonso
Originally posted by Patrick McConnell:
Hello Alfonso,
Will you please clarify when it is advisable to include effects of condition regressors as confounds and when not to? I am looking at between-groups differences in functional connectivity between an ROI and the rest of the brain during successful response inhibition. My understanding is that for functional connectivity analyses, I want to include effects of condition as confounds. However, I would also like to run a PPI analysis using the same ROI as source, and test whether connectivity between that ROI and the rest of the brain is differentially modulated across groups by the condition effects (successful response inhibition, i.e., no-go correct). In this case, would I want to redo the denoising procedure excluding the condition effect (no-go correct) from the confounds list?
Thank you,
Patrick
Will you please clarify when it is advisable to include effects of condition regressors as confounds and when not to? I am looking at between-groups differences in functional connectivity between an ROI and the rest of the brain during successful response inhibition. My understanding is that for functional connectivity analyses, I want to include effects of condition as confounds. However, I would also like to run a PPI analysis using the same ROI as source, and test whether connectivity between that ROI and the rest of the brain is differentially modulated across groups by the condition effects (successful response inhibition, i.e., no-go correct). In this case, would I want to redo the denoising procedure excluding the condition effect (no-go correct) from the confounds list?
Thank you,
Patrick
Threaded View
| Title | Author | Date |
|---|---|---|
| Patrick McConnell | Mar 28, 2015 | |
| Alfonso Nieto-Castanon | Apr 6, 2015 | |