Firstly, your math doesn't quite add up, 7 functions at 2.5 is 17.5 seconds.
Secondly, is there a compelling reason to not just set FIR width to the TR?
Relatedly, is your design jittered? if the event onset is jitters with regards to TR onset, I am not so sure there is a compelling reason to match the FIR width to the TR.
And did you slice time correct the data? I have the impression this is more important with FIR than it might be with, for example an informed basis set (e.g HRF plus derivative and dispersion).
Good luck,
Colin Hawco, PhD
Neuranalysis Consulting
Neuroimaging analysis and consultation
www.neuranalysis.com
[log in to unmask]
-----Original Message-----
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf Of Andy
Sent: September-12-17 10:15 AM
To: [log in to unmask]
Subject: [FSL] FIR Model Not Aligned with TR
Hello,
I am using an FIR model to estimate the activity at set time points after the onset of an event. However, my TR is 2s, and I would like to estimate activity every 2.5s (e.g., set the window at 15s and use 7 basis functions).
Is there any disadvantage to doing this? Is the estimation less efficient than if I used a basis functions that were synchronized to the TR time grid?
Best,
-Andy
|