HI - this is an implementation of the variance normalisation used by MELODIC -
see Beckmann TMI 2004.
The idea is to remove the 30 strongest PCA components (temporarily) and use the
remaining weakest components (ie residual "noise") to compute the voxelwise
scalings for the variance normalisation, to be applied to the original data.
30 is indeed somewhat arbitrary, but for this step, increasing it further
doesn't have a huge impact.
Cheers.
> On 3 Jun 2017, at 01:29, Chihuang Liu <liuchihu...@gmail.com> wrote:
>
> Hi HCP!
>
> As I was going through the scripts for generating HCP 820-subject PTN
> release, there is a step in script "subproc_DO_2_groupPCA.m" that seems quite
> confusing to me. The part of code is pasted as follows:
> BO=ciftiopen(f{r(1)},WBC); grot=demean(double(BO.cdata)'); clear BO.cdata;
>
> [uu,ss,vv]=ss_svds(grot,30); vv(abs(vv)<2.3*std(vv(:)))=0;
> stddevs=max(std(grot-uu*ss*vv'),0.001);
> grot=grot./repmat(stddevs,size(grot,1),1); % var-norm
>
> W=demean(grot); clear grot;
>
> This is done to every individual run and it's supposed to temporally demean
> and variance normalize the data. Why is this way (marked in bold) used to
> compute the stddevs? And how is this number 30 in the ss_svds chosen? It
> seems to be fixed and not relative with the dimension of ICA.
>
> Thanks for any clarification!
>
> _______________________________________________
> HCP-Users mailing list
> HCP-Users@humanconnectome.org
> http://lists.humanconnectome.org/mailman/listinfo/hcp-users
>
---------------------------------------------------------------------------
Stephen M. Smith, Professor of Biomedical Engineering
Head of Analysis, Oxford University FMRIB Centre
FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
st...@fmrib.ox.ac.uk http://www.fmrib.ox.ac.uk/~steve
<http://www.fmrib.ox.ac.uk/~steve>
---------------------------------------------------------------------------
Stop the cultural destruction of Tibet <http://smithinks.net/>
_______________________________________________
HCP-Users mailing list
HCP-Users@humanconnectome.org
http://lists.humanconnectome.org/mailman/listinfo/hcp-users
