Dear Dave,
> Sorry to bore you with this problem again, but I think there is a bug in
> the implementation of variable epoch lengths in SPM99. SPM now asks you
> for a vector of variable scan lengths after you click 'variable epoch
> lengths' (which is good),but it still wants you to define epoch lengths
> after you click 'temporal derivatives yes/no' (which is bad). So you are
> effectively defining epoch lengths twice, and the second option seems to
> only want scalars as its input (i.e. it's identical to the previously
> implemented prompts for epoch lengths). Is it possible to take out this
> wee stumbling block?
The variable duration option allows you to emulate variable length
epochs in terms of a train of events. It is a new option in the fMRI
deisgn matrix setup that automatically produces an ontset for every
time-bin of specified duration. Clearly when it comes to convolving
this with a bsis set you should use those appropriate for event-related
responses. Therefor use the event-related options not epoch-related.
Strictly speaking the variable duration refers to a train of onsets.
I hope this helps - Karl
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|