Dear Giovanni,
>I've got a question about the High-pass-filter in SPM2 (in temporal
>autocorrelation options).
>SPM2 displays by default the value 128. With this new version of SPM, I'm
>not sure if I can use this value of 128 every time whatever the paradigm, or
>do we need to change it and to adapt it to the paradigm frequencies as we
>used to do with SPM99 (i.e. 2*SOA MAX) ?
Unless you have very low frequency treatment effects, I would stick to
128 seconds. This is based on the noise spectrum in fMRI and the necessary
constraints on signal frequencies imposed by the biophysics of the HRF.
It is therefore not so much a function of experimental design.
I hope this helps - Karl
|