Russ - It is an absolute statement if you really only have 6 data points.
Bedpost is estimating 6 parameters (including a variance) - imagine trying
to estimate distributions on your parameter estimates in FMRI if you have
6 datapoints and 5 regressors !!
If you have many averages, you can merge them to give Bedpost a little
more freedom to estimate the variances, but the uncertainty in orientation
will then be very dependent on the orientation of the fibre bundles with
respect to the orientation of your bvectors. Derek Jones did a study on
this at ISMRM in Toronto I believe. I think it may now be published in
MRM. The uncertainty stops being dependent on orientation wrt bvectors
when you get to about 25-30 isotropically oriented bvectors- hence we
recommend this number as the minimum for probabilistic tractography.
Hope this helps
T
-------------------------------------------------------------------------------
Tim Behrens
Centre for Functional MRI of the Brain
The John Radcliffe Hospital
Headley Way Oxford OX3 9DU
Oxford University
Work 01865 222782
Mobile 07980 884537
-------------------------------------------------------------------------------
On Tue, 7 Dec 2004, Russ Poldrack wrote:
> Thanks Andreas. Regarding the using a 6-direction data with bedpost,
> is this an absolute statement (i.e, "One will never be able to use
> 6-direction data to do probabilistic tracking") or a relative statement
> ("One needs to be careful with the output of bedpost using 6-direction
> data")? If the latter, can you suggest some metrics for robustness?
>
> cheers
> russ
>
> On Dec 7, 2004, at 4:00 PM, Automatic digest processor wrote:
>
> >
> > hi russ-
> >
> > 6 directions is generally too few for a robust probabilistic solution.
> > w/r to 1, i would tend to average the coregistered /
> > eddy-current-corrected runs and perform further computations on the
> > average. (simply because it would be less time consuming.)
> > w/r to 2, yes - flirt can register to the b0-image (corrratio puts up
> > well with the liquor being bright on b0 but dark on the other images),
> > i.e you can simply put it first to use it as the default reference
> > volume.
> >
> > best regards-
> > andreas
> >
> ---
> Russell A. Poldrack, Ph.D.
> Assistant Professor of Psychology, UCLA
> Franz Hall, Box 951563
> Los Angeles, CA 90095-1563
> email: [log in to unmask]
> phone: 310.794.1224
> fax: 310.206.5895
> web: http://www.poldracklab.org/
>
|