Hi - yes, for a first-level (time series) analysis the data gets demeaned
so you shouldn't use a constant EV.
(Note though - we don't demean the data at second-level, because the mean
levels are usually of interest themselves, so the designs at second-level
often DO have a constant EV in them.)
Thanks, Steve.
> Hi Steve !
> Thanks for your reply.If the constant 'mean' term is implicitly assumed in the way FILM
> works, then I take it my practice of trying to include this column of ones in the design
> matrix is redundant, right?(I am actually reproducing a design I had inquired about in a
> previous posting).
> Charis.
>
> > Hi - yes, this is probably the same problem as Antti was seeing. The
> > fix I mentioned in the previous email will prevent feat_model from
> > crashing when there is a nearly-constant EV, however if you have a
> > _completely_ constant EV there will be problems later in the model
> > fitting, as FILM starts by demeaning the data (an alternative to
> > putting a constant column in the design matrix). If you want the
> > constant value then you can simply take the temporal mean of
> > filtered_func_data with
> >
> > avwmaths filtered_func_data -Tmean mean_image
> >
> > Thanks, Steve.
> >
> >
> > On Tue, 12 Nov 2002, Charis Tzagarakis wrote:
> >
> > > Hi there!
> > > I have a problem which may be similar to that of Antti.I have an EV
> > > made up of ones in its entirety(it represents the overall mean level
> > > of the signal).When I try to include it in my design (as a one-entry
> > > per time-point custom file), I get a 'Problem with processing the
> > > model:child killed:segmentation violation' error.Removing a single
> > > '1' and replacing it with a '0' seems to allow avoiding that
> > > error.The other EVs I use are also custom files made in a similar
> > > way but they do not seem to cause any problems whatsoever.What could
> > > be the problem? Is it equivalent to try to create this EV by
> > > selecting 'square' in the EV basic shape menu and then setting '0'
> > > off time and an 'on' time that is equal to the duration of the
> > > entire experiment?(with skip=0, phase=0 and stop after=-1).That is
> > > what I've been trying as a workaround, but I'm not sure whether the
> > > 'on' value for 'square' is '1' as for my custom made EVs. Thanks !
> > > Charis
> > >
> > > Charidimos Tzagarakis MD
> > > Brain Sciences Center and University of Minnesota.
> > > Address:
> > > Brain Sciences Center (11B)
> > > Minneapolis VA Medical Center
> > > One Veterans Drive
> > > 55417 Minneapolis MN
> > > USA
> > > Tel:612-725-2000 ext 1756
> > > email: [log in to unmask]
> > >
> >
> > Stephen M. Smith
> > Head of Image Analysis, FMRIB
> >
> > 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
> >
>
>
>
> Charidimos Tzagarakis MD
> Brain Sciences Center and University of Minnesota.
> Address:
> Brain Sciences Center (11B)
> Minneapolis VA Medical Center
> One Veterans Drive
> 55417 Minneapolis MN
> USA
> Tel:612-725-2000 ext 1756
> email: [log in to unmask]
>
Stephen M. Smith
Head of Image Analysis, FMRIB
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
|