Hi,
I'm having some trouble with a masked analysis which I hope someone
can help with. I'm doing a basic T test at the second level and get
some reasonable results when I run a whole brain analysis. But when I
try to estimate exactly the same T test within a mask which is a
sphere with 10mm radius, the model won't estimate. A bit of detective
work has shown me that it is crashing around line 880 of spm_spm.m,
because spm_est_smoothness is returning NaNs for FWHM and then
spm_resels_vol fails
line 880 [FWHM,VRpv] = spm_est_smoothness(VResI,VM);
line 881 R = spm_resels_vol(VM,FWHM)';
So I'm guessing that I have a smoothness problem. But I don't know if
it arises from my data (which I originally smoothed at 9mm FWHM) or
fact that I'm using a small mask or some combination. I can estimate
within a much bigger mask (30mm radius), but that isn't much good for
an ROI analysis.
Is there any way to fix this problem? What are the limits on mask size and why?
Thanks for your help,
Antonia.
-----------------------------
Dr Antonia Hamilton
School of Psychology, University of Nottingham
University Park, Nottingham, NG7 2RD, UK
+44 115 846 7921
[log in to unmask]
www.antoniahamilton.com
-----------------------------
|