Hello,
I’m planning to deconvolve our data before running gPPI and second-level analyses in CONN, based on the method described in the paper “Comparison of whole-brain task-modulated functional connectivity methods for fMRI task connectomics” doi: Comparison of whole-brain task-modulated functional connectivity methods for fMRI task connectomics | Communications Biology (https://static-content.springer.com/esm/…), and I wanted to check if this approach makes sense and if it is feasible to do it in Conn.
Here’s what I was thinking:
-
My data is preprocessed using fMRIPrep.
-
Do spatial smoothing and denoising in CONN, after setting “acquisition type” to “sparse” in the Setup.Basic tab as suggested https://www.nitrc.org/forum/message.php?… Click or tap if you trust this link." href=“NITRC: CONN : functional connectivity toolbox: RE: Extra scans in condition specification?” target=“_blank” rel=“noopener noreferrer” data-auth=“NotApplicable” data-linkindex=“0”>here, so the design matrix isn’t convolved with the hrf.
-
Run a gPPI analysis in CONN to create the appropriate files & structure, but we will later overwrite its results
-
Use the denoised timeseries from conn*/results/preprocessing/ROI_Subject*_Condition*.mat.
-
Deconvolve those timeseries using spm_peb_ppi.m function.
-
Then reconvolve the interaction term or PPI regregressors using CONN.
Now the question is how/whether I can have CONN do the gPPI calculations before we reconvolve with the canonical hrf meaning which files we need to replace to allow CONN to use that data.