Dear Peter,
> sorry for not being clear about our question, here we try again:
> WE have for instance an fMRI session of 100 scans,
> 1 condition,
> SOA is variable,
> Variable duration: yes,
> Durations (in No of Scans): 1 2 1 1
> parametric modulation: other
> expansion: linear
> trial: 1
>
> SPM now asks for 84 parameters, while our calculation suggests 80
> parameters only: 16 + 2*16 +16 +16
> Now we do not know how to derive the 4 parameters, that we cannot account
> for (are they at the end or beginnig of each event?).
I think everthing would be simpler if you answered 'no' to 'Variable
duration' and simply model the epochs as trains of individually
specified events each of which has a 'covariate' value. This entials
specifying an onset time vector with 80 elements (if you have 80
covariates measured at those onset times).
Does this help? - Karl
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|