Dear andreas,
> (1) in an epoch related design with 2 baseline conditions (B1 & B2) and
> 2 active tasks (A1 & A2), both of the latter parametrically modulated,
> is there any reason why to avoid varying epoch lengths? Both A1 & A2
> may make it psychologically sensible to employ epochs of different
> duration. In addition, I thought it may be helpful to exclude errors
> introduced by drifts at low ferquencies. Is that correct?
If it is pyschologically important to use epochs of varying length then
you should certainly do so. Variable length epochs have no special
advantage in term of being able to remove drifts that I can think of.
The only drawback is that you will have to model each epoch as a train
of events. This precludes modeling within-epoch adaptation with
epoch-based basis functions (e.g. mean and exponential decay). Other
minor issues are that you will have to think in terms of event density
when interpreting your results and the epoch related plots will be to a
single event not the event train.
> (2) how to get SPM to help with setting up a paradigm design?
You can create the deisgn matrix before acquiring the data and then use
it to analyse the time-series afterwards. The design matrix and
associated parameters in the configuration file partially specifiy the
experimental paradigm.
I hope this helps - Karl
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|