Dear Simon,
> 1. Did I throw away the first three scans correctly, so that the
> estimation did not use these scans and the rest of the scans are put into
> the correct condition? When I looked at the "design image" of the stats
> analysis, it's hard to tell if the images were lined up properly or not.
Yes; the early 'dummy' scans should not be included in the analysis. If you
specified 72 scans in the design matrix, then SPM will only let you select
72 scans to analyse and assuming you correctly selected #4 onwards, all will
be OK!
If you are really paranoid (!) then you can look at the SPMcfg.mat data
structure directly by typing (in the MATLAB window, after starting SPM99 in
your analysis directory):
>> load SPMcfg.mat
>> VY(:).fname
...which will tell you the file names that you actually analysed, in order!
> 2. What does temporal derivative do? The model looked so different
> with this option. How do I set up contrasts for this?
If you select this option, SPM99 will construct an additional column for
each condition - the 'temporal derivative' of the original boxcar (or other
basis function). The idea is that for each condition, some linear
combination of the original boxcar and the temporal derivative can model a
wider range of brain responses than the boxcar alone. Specifically, the
combination of the two allows modelling brain responses that have the form
of a boxcar, but an onset and offset that is delayed (or shortened) relative
to the original boxcar (i.e. phase lag/advance). This can be helpful in two
situations. First, it may (or may not!) slightly improve the fit of the
original boxcar to brain responses. Second, you can explicitly compare the
parameter estimates for the temporal derivatives which may reflect different
onset latencies for haemodynamic responses between different conditions.
IMHO, simple boxcar is usually enough for most purposes!
best wishes,
Geraint
------------------------------------------------------------------
Dr. Geraint Rees
Wellcome Advanced Fellow, Lecturer,
Mailstop 139-74, Institute of Neurology,
California Institute of Technology, University College London,
Pasadena, 12 Queen Square,
CA 91125 London WC1N 3BG
voice (626) 395-2880
fax (626) 796-8876
web http://www.klab.caltech.edu/~geraint
------------------------------------------------------------------
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|