Hi,
The answer to this depends in what space you've run the vertex analysis. If it is run in MNI space, then there is already a scaling that should account for the head-size changes (though some people prefer to measure TIV or total GM instead, and it is also debatable whether a non-linear correction is useful, but in general this normalisation to MNI space gets rid of the bulk of the effect). If you ran the vertex analysis in native space then you should account for the head-size (or TIV, or total GM) separately.
All the best,
Mark
On 30 Sep 2013, at 15:49, Sarah Winsland <[log in to unmask]> wrote:
> When you extract volumetric info for a structure (e.g., fslstats output_name_all_fast_firstseg -l 16.5 -u 17.5 –V) and then run between-group comparisons using a different software package (e.g., matlab) do you need to include total GM volume as a covariate or is this somehow accounted for in the segmentation and registration process that is run as part of run_first_all? Thanks!
|