Dear Jeff,
> If one has an fMRI experiment using a delayed boxcar function (with a 6
> second delay), a TR of 6 seconds, and a cycle of 10 rest scans
> alternating with 10 task scans. If one was comparing active
> to rest scans, would this compare
>
> scans 3 to 12, 23 to 22, 43 to 52, etc. for the rest scans
> with
> scans 13 to 22, 33 to 42, 63 to 72, etc. for the active scans
> OR
>
> scans 3 to 10, 23 to 30, 43 to 50, etc. for the rest scans
> with
> scans 13 to 20, 33 to 40, 63 to 70, etc. for the active scans
>
> Thanks alot
>
> I could not seem to find this specification anywhere in the course notes
> regarding the general linear model
The regressor SPM would come up with will model all scans in your
time-series so that scans 1 and 2 would be attributed to the rest
condition as well. I guess a delay of 6 sec would mean that the first
rest condition starts at scan 1 and the first scan of the activation
condition starts at scan 12 and so on.
A delay of the boxcar basis function is modelled by a convolution with a
model of the HRF. Your convolved boxcar regressor will therefore not
only be delayed but also contain some values other than e.g. -1 and 1.
This means that one cannot really say for short TRs that SPM 'compares
scans', if you convolve your model with the HRF. To find out, how
exactly your regressor looks like, you could specify a model, load the
resulting SPM_fMRIDesMtx.mat and plot xX.X (which contains the temporal
basis functions).
Stefan
--
Stefan Kiebel
Functional Imaging Laboratory
Wellcome Dept. of Cognitive Neurology
12 Queen Square
WC1N 3BG London, UK
Tel.: +44-(0)20-7833-7478
FAX : -7813-1420
email: [log in to unmask]
|