Hi Experts
I also have the same confusion.
According to the mathematical equation. Between a particular varcope and the sigmasquareds, the scalar resulting from c'*pinvX*pinvX'*c (which is from the equation varcope(i)=c'*pinvX*pinvX'*c*sigsq(i)), should be the same for each voxel.
Therefore, when dividing a particular varcope image by the sigmasquareds image, the resulting value should be the same for each voxel.
However, when I tried to do this, each voxel has different values which is not explained by rounding errors.
Especially, when you look at the resulting volume with fsleyes, it actually still resembles the structure of the brain. For instance, you can kind of tell apart the CSF, grey matter, and white matter.
Therefore, I am wondering whether some kind of tissue probability mask were used when calculating the varcope from the sigmasquareds?
Thank you very much!
Best wishes,
Chen-Chia
|