Hi Jinny,
To some degree, you should already be removing those clusters,
assuming they do not come from mistakes in processing.
It is common to have small effects outside the brain from things
such as gaussian blurs, ringing effects from blurring with Fourier
interpolation, or ghosting of the data straight out of the scanner.
If the cause seems to come from those things, you are golden. It
is good to see what happens outside the brain in any case, mostly
as a sanity check.
By my statement of "already be removing those clusters", I mean
that when you measure the FWHM blur in the data in order to get
cluster values from AlphaSim, you probably provided a brain-only
mask to AlphaSim. If that is the case, then the assumption is
that you only have data within that mask, i.e. you are already
applying that mask to your data.
In this case, it is appropriate to apply your brain-only mask to
your results even before clustering.
- rick
