Dear Tom:
The reference you're looking for should be:
Worsley et al. "Detecting changes in nonisotropic images"
Human Brain Mapping, Vol. 8, 2/3, 98-101
This is a concise treatment of the problem in question. I think K. Worsley was
working on a more complete and technically-oriented paper, but as far as I know
it hasn't been published yet.
I was a bit surprised to learn that, judging from the previous mails, the
implementation of this new smoothness estimator could lead to significant
changes in the corrected p-values _for the intensity test_, since, as far as I
could gather from the above mentioned paper, taking into account non stationary
correlation structures should only make a practical difference in the p-values
for the _extent test_ (because the distribution of cluster sizes will be biased
if smoothness anisotropy is strong).
I'd be interested to know about more people who have noticed p-value differences
between two SPM versions that could reasonably be attributed to the new
generalized estimator.
Cheers,
Alexandre
#--------------------------------------------------------#
| Alexandre Andrade |
| Service Hospitalier Frederic Joliot, CEA/DRM |
| 4, Place du General Leclerc |
| 91401 Orsay cedex, France |
| Phone: +33 1 69 86 78 21 |
| Fax: +33 1 69 86 78 68 |
| e-mail: [log in to unmask] |
#--------------------------------------------------------#
> SPM craftsmen,
>
> Excerpts from SPM-help#: 3-Feb-100 Re: Fw: comparision bet... by K Friston
> > The only difference between SPM96
> > and SPM99 (in terms of estimtation and inference) is that the spatial
> > smoothness estimator has been upgraded. This upgrade accommodates
> > non-stationary spatial correlations among the errors when correcting
> > for the search volume using tests based on peak height (but not on
> > those based on spatial extent, at this stage). This advance (due to
> > Keith Worsley) allows one to be more confident about corrected
> > inferences in situations where the smoothness may change from region to
> > region (e.g. VBM).
>
> Is there a reference for this improved smoothness estimator?
>
> -Tom
>
>
> -- Thomas Nichols -------------------- Department of Statistics
> http://www.stat.cmu.edu/~nicholst Carnegie Mellon University
> [log in to unmask] 5000 Forbes Avenue
> -------------------------------------- Pittsburgh, PA 15213
>
>
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|