Dear Peter and Florian,
> in a recent reply by Karl Friston (Thank You !),
> he stated:
> > This approach
> > eschews any problem with asynchrony between acquisition and trials and
> > automatically ensures approoriate convolution with the HRF in
> > 'microtime' (i.e. time-bins of TR/16).
>
> However, whenever we difine events, SPM asks for 17 covariates, when we
> define a period of exactly 1 TR. If we define a period of 2 TR SPM asks
> for 33 covariates, the system seems to be 16*TR + 1. Where does this
> one extra covariate come in ?
I should think one at the beginning (or end). I am a little confused
by your question because I am not quite sure what you mean by
'covariates'. The number of elements in the vector specifying the
parametric modulation (i.e. your behavioural covariate) should be equal
to the number of events you specified. This should be equal to the
number of onset times specifed. These onset times should be chosen to
reflect the number and sampling interval appopriate for your
behavioural index. This could be every second in continuous
performance tasks or every SOA for trial-specific indices.
I hope this helps - Karl
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|