Dear FSL experts,
Iām hoping to get some guidance about the best way to run a PPI analysis on a dataset with a fast event-related design.
My understanding is that FSL constructs Interaction regressors by multiplying the convolved task regressor with the seed region time course, while SPM (and, I believe, AFNI) deconvolves the physiological time course to approximate the neural response, multiplies that by the raw task regressor, and then re-convolves. A few years ago, I was advised that, while the two methods were roughly equivalent for block designs, only the SPM method had any hope of getting an effect in an event-related design, following from Gitelman et al's 2003 paper.
My questions are the following:
ā Is it still true that FSL does not provide a straightforward way of deconvolving the seed region time series before constructing the PPI regressors?
ā If the answer to the first question is no, is there currently reason to believe that the FSL method is acceptable for an event-related design? Or is it still clearly preferable to use the SPM-style method of PPI with an event-related design?
Thanks,
Michael
--
Michael S. Cohen, Ph.D.
Postdoctoral Researcher
Reber Lab
Department of Psychology
Northwestern University
|