Hi Mike,
| Does anyone have a view or experience of applying inhomogeneity
| correction to T1-weighted structurals prior to spatial normalisation
| in SPM99? Intuitively, I would guess the less inhomogeneity the better
| the match to the T1 template, but I could be wrong.
Inhomogeneity correction may improve spatial normalisation, but I haven't checked
this out properly. Because the correction is based on grey and white matter,
it has to extrapolate outside the brain. I have found that this can change the
scalp intensities in T1 weighted images by quite a bit. However, the influence
of the scalp is minimal when brain-masking is enabled.
Incedentally, a small amount of intensity non-uniformity is modelled by the
nonlinear part of the spatial normalisation, but this is just a linear intensity
gradient.
All the best,
-John
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|