Hello,
I'm trying to run some simulation data-sets through SPM for fMRI and I have
a few questions about the results that SPM generates.
Basically, I set up my simulated activation map (40x40x40 voxels) as a
checker board pattern of 10x10x10 ON/OFF voxels (value=100, 0 respectively)
and use it to modulate the intensity of the simulated BOLD signal (a
regular-spaced stimulus train (impulse) convolved with the canonical hrfs
viz. spm_hrf). I then added some AR(1) noise (sigma^2 = 0.01) and a linear
drift to the data.
When I run this data-set through SPM (canonical hrfs with no derivatives, p
< 0.01, no correction) I get the original region as a very strong
activation plus an additional cube-shaped regions of low activation. Given
the regular structure of this spurious regions - it does not seem likely to
be a Type I error.
Also if I examine the beta_0001 image, corresponding to the stimulus
regressor, the values vary between 1-10 (while the original map was 0/100),
and while the shape of the high-intensity region is approximately correct,
it is fairly blurred.
Given that I'm not smoothing my data-sets or doing any other kind of
pre-processing why do I observe these effects. Also, why doesn't the beta
map reflect the original intensity of the activation pattern ?
Thanks,
-firdaus
|