Hi,
On 26 Nov 2008, at 18:27, Gareth Gaskell wrote:
> Dear FSL users,
>
> I have a question about using Featquery on a high-level cope
> directory in
> order to conduct an ROI analysis on fMRI data.
>
> I have an event-related experiment that has been analysed at three
> levels,
> with separate runs from different participants combined at the mid-
> level.
>
> For the ROI analysis in featquery, I selected particular cope
> directories
> from the top level analysis and then from the output html page
> selected the
> "Mean time series (masked/weighted)" link. This contains as many
> data points
> as I have participants, and I assume that each data point is the mean
> parameter estimate for that contrast within my mask for one
> participant.
that's correct
> So I was wondering whether this assumption is correct, and if so
> whether it
> makes sense to then conduct simple t-tests on these by-participant
> values. I
> was a little concerned that the mean of these values doesn't match
> the mean
> parameter estimate quoted lower down in the feat report, although
> they are
> fairly similar.
Right - yes if you use the same ROIs then the only difference between
mean [highest-level input data] and mean [lowest level COPE values]
will be related to interpolation effects - probably mainly the
difference will be that the ROI mask will have very different
resolution when applied in those two cases (in the latter case it will
be transformed into native (lowres) resolution). Hence it won't be
_exactly_ the same mask, but results should be close.
Cheers.
>
>
> The alternative, I guess, would be to go back to the first level
> analyses
> and run featquery on each run of each participant.
>
> Advice on whether my interpretation is correct would be appreciated!
>
> thanks,
>
> Gareth
>
---------------------------------------------------------------------------
Stephen M. Smith, Professor of Biomedical Engineering
Associate Director, 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
---------------------------------------------------------------------------
|