Hi
Two related main reasons:
(i) FMRI data raw intensity values are largely meaningless, so expressing ICA maps relative to the residual noise makes sense
(ii) this gives a close correspondence to what is being done in classical (GLM type) of analysis
For details youmight want to check the original PICA paper (Beckmann & Smith 2004)
If you want the unscaled IC values you can get these with the --Oorig command line option
hth
Christian
On 16 May 2013, at 00:29, qiang wang <[log in to unmask]> wrote:
> Hi Experts
> Why the results (ICs) of MELODIC are all transferred into Z score?
> Thanks,
> Qiang
|