Hello
I have been experimenting with connectivity analysis and using a script similar to the one Giuseppe Pagnoni was using in 3dDeconvolve. I am using -censor flags to indicate conditions. However, I find that artifacts, for example fluctuating susceptibility artifact, are more problematic for this sort of analysis than for those based on estimation of a beta coefficient across many trials. I tried using a global signal baseline but artifactual fluctuations are still a big contributor to spurious correlations. One way I thought to address this is to correlate peak timepoints in events relative to the onset time points:
Is there a way to select sub-briks in a 3d+time using a 1D file (specifically an onset vector) for writing out to a compressed dataset?
The goal as I see it would be to have one dataset containing 5-sec delay peaks for, say 30 events of a given condition, one dataset containing the onset timepoints for those events, and subtract the two to get a response estimate on a given trial.
The only way I have uncovered in the documentation to accomplish this is to specify the sub-briks exactly within the, e.g., 3dtcat command.
Thank you as always for your help.
Alison