Dear Stephan,
> When I specify "stochastic" under the fMRI models: specify design option,
> it asks me whether there is a null event type, which is good, but then it
> asks for a single SOA, which doesn't make sense because I need to enter a
> vector of onsets for each condition.
I wouldn't use the stochastic option. I believe that it was created in order
to *generate* stochastic design matrices, like your randomisation program,
which could then serve both for analysis and to trigger the stimulus
presentation program. Hence it only asks for a single SOA as SPM is going to
arrange the order of conditions for you!
> If I take the non-stochastic/deterministic option, it asks me for vectors
> of onsets (which is good) but it never asks whether there are null events
> (bad) and I am concerned that by telling SPM99 this is a non-stochastic
> model that the analysis will not proceed correctly.
No, the only difference between so-called 'stochastic' and so-called
'non-stochastic' is how the events are distributed, not how they are
analysed. Simply specify the 'null events' as a separate condition with
their own vector of onset times. You will then be able to look explicitly at
contrasts between active events and 'null events'.
> Perhaps the problem is that SPM99 wants to design the experiment for me
> (by default) when I select "stochastic" and I need to somehow circumvent
> its help and hand-enter the vectors of onsets?
Correct! Just use the non-stochastic option.
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
------------------------------------------------------------------
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|