Hi Andrew,
On Wed, 10 Aug 2005, Andrew Kozel wrote:
> I recently attended the FSL course and have been looking through the
> documentation to try to answer the following question.
>
> My understanding of using the GLM approach in FEAT - When comparing
> event A with B at a voxel, PE are created for each EV (A & B) which is a
> measure of how well the voxel's time-course matches the expected
> hemodynamic response for that EV. You designate a contrast such as 1 -1
> to test if the PE are significantly greater for A than B (with
> appropriate error estimate term). If you get a positive result, then a
> "greater response" is attributed to A than B. My question is that you
> do not appear to be testing for magnitude of response but closeness of
> fit to your expected curve.
In fact, it is not exactly either of these. The most accurate answer is
that with the [1 -1] contrast you are testing the relative PEs from the
model fitting of EVs A and B. If you use the same HRF in each, then this
means that you are comparing the FITTED heights of response.
But yes you are right that if the model is the wrong shape (or timing,
etc.) wrt the data, then this will indeed impact on the fitted height -
for example if the HRF is much slower in reality than you expected, then
this will result in reduced fitted height - i.e. reduced PE.
Hope this helps? Cheers, Steve.
> Is it possible to have a greater magnitude
> in response but the time course does not match as well? In addition,
> when comparing different subjects or the same subject at different time
> points (e.g. before and after some treatment) will the interpretation of
> "greater response" be greater magitude in change and/or closer fit?
>
> Thanks for the great course and subsequent help,
> Andy
>
--
Stephen M. Smith DPhil
Associate Director, FMRIB and Analysis Research Coordinator
Oxford University Centre for Functional MRI of the Brain
John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
[log in to unmask] http://www.fmrib.ox.ac.uk/~steve
|