Print

Print
Hi


On 6 Oct 2015, at 10:30, MCLAREN, Donald <[log in to unmask]> wrote:

Christian and Stephen,

In MELODIC, the subject data is variance-normalized on a voxel-wise basis prior to any computation. In reading paper on dual-regression, authors don't typically say whether or not the subject data is variance-normalized or not.

typically we don't variance normalise the data when running dualreg

What is your recommendation for the subject's fMRI data input into dual-regression: variance-normalized or demeaned or another option?

I understand that in FSL, the variance-normalization option is applied to the subject-specific timecourses from Stage 1

by default (and recommendation), yes.

Cheers


, before the Stage 2 model, but the question about the subject's fMRI data is a separate question.

Best Regards, 
Donald McLaren, PhD



---------------------------------------------------------------------------
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)
[log in to unmask]    http://www.fmrib.ox.ac.uk/~steve
---------------------------------------------------------------------------

Stop the cultural destruction of Tibet