Hi Lars,
> ${FSLDIR}/bin/fslmaths ${I2R}_warp -sqr -Tmean ${I2R}_tmp
> ${FSLDIR}/bin/fslstats ${I2R}_tmp -M -P 50 > ${I2R}_warp.msf
>
> 1. Are these numbers a sensible estimate of "amount of warping
> applied"?
...ish. It is rather hard to interpret displacements per-se (or spline
coefficients of displacements, which is what is contained in the _warp
file).
> 2. Given the age of the subjects included in the FMRIB template (20-50
> yrs), I would expect a U-relationship between amount of warping and
> age in
> a healthy lifespan sample (8-85 years). However, I find a negative
> linear
> (r=-.59) relationship between age and this value. The resulting warped
> volumes look excellent across all ages, but I find these values a bit
> puzzling. Any comments?
I wouldn't worry too much about this. A rather more interesting
measure would be e.g. the upper and lower 5% percentile of logarithms
of the Jacobians of the warps. That would indicate how much local
squeezing/expansion that was necessary to warp the individual subjects
to the template. You might then (fingers crossed) find your u-shape.
Good luck Jesper
|