Hi - no you need to do something slightly more involved, so that vectors pointing in opposite directions do not get cancelled.
We typically average rank-1 tensors and extract the first eigenvector of the resulting tensor. The associated eigenvalue is related to the dispersion.
You can use the following program to do this: www.fmrib.ox.ac.uk/~saad/average_vectors
Cheers
Saad
On 22 Aug 2012, at 12:44, SUBSCRIBE FSL Jay wrote:
> Hello
>
> After thresholding the dyads using maskdyads routine, i would like to average the peaks of the dyads across a group of 20 subjects. But i see that the dyads have 3 volumes.
>
> Should i average the corresponding 3 volumes in each subject to get a final averaged dyad with 3 volumes? Is it alright to average dyads across subjects to see the aveage peaks within a ROI?
>
> BW
> Jay
>
--
Saad Jbabdi
University of Oxford, FMRIB Centre
JR Hospital, Headington, OX3 9DU, UK
(+44)1865-222466 (fax 717)
www.fmrib.ox.ac.uk/~saad
|