What works best will depend on the amount of bias artifact in your
data. If there is no bias artifact, you would use very heavy
regularisation (or disable bias correction) because there should be
nothing to correct. If your data are heavily corrupted by bias
artifact, then you'd use much less regularisation in order that the
model has more flexibility.
The FWHM option also depends on the nature of the artifact. If it is
very low frequency, you'd be better off with a broad FWHM. If it
contains high frequency, you may need a smaller FWHM.
Best regards,
-John
On 8 January 2012 17:16, Jordan Larson <[log in to unmask]> wrote:
> Hey SPMers,
>
> When using standard T1 weighted images in a VBM study with Estimate and
> Write, is it better to use very light or light bias regularization or
> something else entirely? What is the standard or best option for pediatrics?
> How does the option you use under bias regularization affect what you choose
> in FWHM (full width half maximum)? I appreciate any advice anyone may
> provide.
>
>
> Thank you
>
>
> Jordan Larson
>
> Undergraduate Research Assistant
> Department of Psychology
> Southern Illinois University Carbondale
|