Hi Hojat,
Yes, you can use the GLM for structural MRI -- this is extremely
common in VBM.
For voxel-wise analysis of either functional or structural MRI, the
GLM is applied with the response (regressand, if I understand your
terminology correctly) given by the set of image values at that voxel
(across time/subjects/groups/etc).
The regressors depend on the model you are fitting, and not in any
fundamental way on whether the data is fMRI, sMRI or something like
blood-pressure. For a two-sample t-test, the design matrix can contain
two columns each of which is 1 in the rows where the response belongs
to that group, e.g. for groups a and b, with three and four subjects
respectively:
response y = [a1 a2 a3 b1 b2 b3 b4]'
regressor a = [1 1 1 0 0 0 0 ]'
regressor b = [0 0 0 1 1 1 1 ]'
design X = [a b]
Best,
Ged.
Hojat Vaheb wrote:
>
> •
> •When we want to analyze fMRI Data, we use GLM. In this Model (
> Y(i)=Beta(i)*X(i) + ErrorTerm) we should find best parameters using
> least squares method. This is ok for fMRI data: we choose basis
> functions (such as HRF , ..) and convolve them with neural causes to
> obtain regressors (X(i)),we can also involve movement parameters in this
> model. But what about analysis of MRI data? Suppose we want to perform a
> two sample t-test between two sets of MRI images, what are regressors
> and regressand in this model. Do we use GLM for MRI data?
>
> ------------------------------------------------------------------------
> Cheap Talk? Check out
> <http://us.rd.yahoo.com/mail_us/taglines/postman8/*http://us.rd.yahoo.com/evt=39663/*http://voice.yahoo.com>
> Yahoo! Messenger's low PC-to-Phone call rates.
|