On Tue, 29 Jul 2008 09:16:48 +0100, Armin Blickenstorfer
<[log in to unmask]> wrote:
>I'm fairly new to SPM and I'm interessted in a greater detail on how you would
>imply a new regressor.
>I hope I do not ask too much
For each volume you want to "regress out," you create a separate regressor.
So if you wanted to regress out 4 volumes, you'd need 4 regressors.
The regressors are not "conditions" but rather "user-specified regressors".
The regressor is simple: it's all zeros, except for a single one in the row of the
volume you want to regress out.
If I recall correctly, in fMRI with multiple sessions (or "runs"), you specify the
regressors for each run separately. If that's indeed correct, you would have a
regressor with all zeros, except for a single 1 in the position of the volume you
want to regress out.
So if you want to regress out volume 151 out of a total of 200 in a session,
the regressor would have 199 zeros, and a 1 at position 151.
You can save yourself the trouble of typing that many zeros using a matlab
expression, since SPM accepts valid matlab expressions. In the example
above, it would be something like
[zeros(1, 150), 1, zeros(1,49)]
Warning: be careful if your volumes are numbered like 0000, 0001, ..., 0199.
In that case, volume 0151 is actually volume number 152.
>Institute of Neuroradiology
>University Hopital Zurich
>CH - 8091 Zurich
>Tel: +41 (0) 44 255 5603
>Fax: +41 (0) 44 255 4504