Print

Print
Dear Dr. Specht,

Thanks a lot for you reply.

As you suggested, I designed the matrix with just one column using a periodic function (Sine function- sin(2*pi*freq*(1:n)) where n is number of scans and f is frequency (<0.10 Hz for resting state). I am confused with the next step- i.e. whether should I use T-contrast as [1] or F-contrast as [1].

If I use T-contrast, clusters I get are not that highly significant and the corresponding peak co-ordinates of these clusters don't match at all with the peak co-ordinates I got from 2nd level analysis using functional connectivity toolbox. On the other hand, if I use F-contrast, I get activations all over the brain even at very lower threshold.

Could you please let me know if I am on right track then should I go ahead with F-contrast as [1] and define VOIs for each subject at peak co-ordinates or I can use the same co-ordinates obtained from 2nd level analysis for each subject?

Thank you so much for your time and help !!
Regards,
Sahil


On Tue, Nov 6, 2012 at 8:16 AM, Karsten Specht <[log in to unmask]> wrote:
Dear Sahil,

Yes, you have to get a SPM.mat file for each subject for getting the time course extracted. Since you know the regions you want to look at, you can try in the simplest case a design matrix with just one column for the overall mean and perhaps the realignment parameter. You could in fact also include some other temporal regressors in the design matrix, if you want to, like the time course of one of the regions you are interested in, or other periodic functions, and you explore then the results with an F-contrast. The result may not look that meaningful in the first place but it will allow you to extract the time course from the regions you are interested in.

Good luck,

Karsten


Fra: Sahil Bajaj <[log in to unmask]>
Svar til: Sahil Bajaj <[log in to unmask]>
Dato: 44Freitag, 2. November 2012 20:25
Til: <[log in to unmask]>
Emne: [SPM] DCM for resting state fMRI

Dear SPM/DCM experts,

Currently, I am analyzing resting state fMRI data. I have done the preprocessing using SPM8 and 1st level and 2nd level analysis using functional connectivity toolbox (http://www.nitrc.org/forum/forum.php?forum_id=2295) to find the activations and peak co-ordinates of four clusters.

Could you please let me know how to use the these four nodes activations to define models in DCM because I don't have SPM.mat files of these four clusters from each subject to use in DCM but what I have is SPM.mat files of each of the regions, obtained from 2nd level analysis.

I would greatly appreciate any kind of help. It would also be great if you could please refer me to someone or some paper where DCM has been used for resting state.

Thanks a lot !!
Regards,
Sahil