Dear experts,
Being discussed here several times, e.g.,
https://www.jiscmail.ac.uk/cgi-bin/wa-jisc.exe?A2=ind2007&L=SPM&O=D&P=52634
https://jiscmail.ac.uk/cgi-bin/wa.exe?A2=spm;af4afb49.1103
I would very much appreciate some clarifications and advice on the topic.
My fMRI experiment includes people with brain lesions and an event-related design.
I included time derivative (TD) in the basis function first-level analysis.
1. Since the derivative regressor is orthogonal to the HRF first regressor, does adding TD influence the group results in any way, if *NOT* taken to the second level? (comparing response delays is not my goal)
2. If the derivatives are included, F-test should be carried out to take the results to the second level.
In this case, should I compare the results of HRF to those of TD in order to interpret the results?
Following this, should the activation caught by the derivatives always be treated as an additive to that of the HRF regressor?
Reasons for NOT including TD:
* I have relatively long events (5s). See https://imaging.mrc-cbu.cam.ac.uk/imaging/DealingWithDifference
* It can increase between-subject variability. See https://www.frontiersin.org/articles/10.3389/fnins.2016.00322/full
Many thanks,
Tammar
|