Origin

Submitted by Veronica on Fri, 12/31/2010 - 02:05

Hello,
We´ve been using REST to analyse resting state scans.
First we did it following the indications of the presentation, using a mask for the ROI and a mask for the brain.
Well, we finally obtained the default mode networks, as expected.
However, we realized that we had used images with different origins (brain mask, ROI mask and scan images), so we changed this parameter so that the three had the same origin. After doing the analysis with REST again, we were consternated to see a different zmap from the first one, and no default mode network could be seen.

We´d like to know how does REST deal with origin point. Does it take into account the origin of the images or the borders of the brains to overlay the ROI to the scan images? What possible explanation can there be?

Thank you very much,

Submitted by YAN Chao-Gan on Sat, 01/08/2011 - 00:36

Hi!
Are your images NIfTI format or ANALYZE format? Why do you need to change the origin for NIfTI images?
You need to make sure the functional images and your mask files are in the same space.

Submitted by Veronica on Mon, 01/10/2011 - 21:14

Hi!
I am using ANALYSE. Yes, I´m asking because images are supposed to have the same origin, but when I change the origins so that they do, the result is different from the result when I don't change the origins, and I trust more in the last ones (they are the correlations that I expect to obtain; the others seem to be noisy). I wanted to know if REST takes into account the origins or the borders to overlay the images, masks and ROIs...
Thank you,
Verónica

Submitted by YAN Chao-Gan on Mon, 01/10/2011 - 22:37

Hi!
If you define your ROI use spherical ROI by setting coordinates and radius, then the information of origin is taken into account.
If you just use a predefined ROI mask file, then the origin information will not affect the correspondence. Could you upload some of your functional images, ROI files before and after origin changing here?