Hi John
The PPI parameters are not convolved twice. The demo does mention an HRF
basis set when setting up the PPI, but that's only because it is a default
value. The PPI vectors are entered as user defined regressors, and as such
are not further convolved by SPM.
Darren
> -----Original Message-----
> From: SPM (Statistical Parametric Mapping)
> [mailto:[log in to unmask]] On Behalf Of John Herrington
> Sent: Tuesday, July 15, 2008 8:46 AM
> To: [log in to unmask]
> Subject: [SPM] PPI, spm_peb_ppi.m and HRF convolution
>
> I am trying to run psychophysiological interaction analyses
> created using spm_peb_ppi.m - both within the context of
> 1st-level SPM analyses and outside of it. In reading
> spm_peb_ppi.m, I realized that apparently the program
> convolves the critical interaction term with an HRF at the
> end. This happens at line 341 (after the interaction term
> has been constructed via the deconvolved task and brain
> signal parameters):
>
> % convolve and resample at each scan for bold signal
> %---------------------------------------------------------------------
> ppi = conv(PSYxn,hrf);
>
> I found this a bit surprising, as the one demo I know of
> regarding PPI:
>
> ftp://ftp.fil.ion.ucl.ac.uk/spm/data/attention/README_GLM_PPI.txt
>
> appears to implement a 1st-level PPI GLM that also does a HRF
> convolution of the input parameters (see line 140 of that
> file). I could be misunderstanding something, but I wanted
> to be sure... does this result in the PPI parameters being
> convolved twice with an HRF?
>
> Thanks - John
>
|