Dear Griselda,
> I have converted a set of Interfile SPECT images to Analyze format using
> the MRIcro program. The original images had negative values outside the
> brain, which have been preserved after the conversion.
> I would now like to convert those values to zero before performing
> statistical analyses on the data using SPM. Does anyone know of a good
> way to do that? I fear that if I leave the negative values unchanged,
> these could influence the global normalization using proportional
> scaling.
It depends on how large the sum of the negative values are in comparison
to the sum of intracranial voxel intensities, whether there is a
measurable effect on the results. The values used for the global scaling
are computed by spm_global, where spm_global attempts to estimate the
mean of intracranial voxels. If the voxels outside of the brain are very
negative, this would lower the estimate of spm_global. You could check
this by comparing the global mean on the original image and a masked
version (using spm_global and image calc within spm99).
One way to change negative values to e.g. 0, is to use image calc for
each image. To do this use a formula like
(i1 < 0)*0
If you are sure that negative values influence the results your SPECT
analyses, the best way would not to change your images, but rather to
use/implement your own estimator of the global mean of intracranial
voxels.
Stefan
--
Stefan Kiebel
Functional Imaging Laboratory
Wellcome Dept. of Cognitive Neurology
12 Queen Square
WC1N 3BG London, UK
Tel.: +44-(0)20-7833-7478
FAX : -7813-1420
email: [log in to unmask]
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|