| we've done a motorial task with an epi acquisition.
| It works all right, but when we're trying to display the results over
| the T1 acquisition
| it seems that the region in which we are interested in doesn't appear.
| I think it's a problem about coregistration the T1 image on the mean
| image obtained from the realigning of
| the BOLD images. when we try to coregister, the T1 image appears out of
| the red frame, so normalization it's not possible, and of course we're
| not able to visualize the results over the T1 image.
Use the <Check Reg> button to see how closely aligned the images are
before you start. If the images are not within about 5cm of each other,
then the coregistration may not work. If the images are not well aligned,
then I would suggest roughly re-orienting them using the <Display> option.
You would do this by displaying one of your images, and trying different
rotations and translations until you get a nice transverse plane in the
lower left panel, with the anterior at the top. You can apply this
transformation to a number of images at once by selecting the images
after hitting the <Reorient images...> button.
Once you have reasonable starting estimates, then your coregistration is
more likely to work. The way the coregistration works is that it starts
with a particular estimate, and then it keeps making new parameter
estimates that it thinks should improve the fit. With poor starting
estimates, the coregistration can end up with a poor solution. Its a
bit like climbing Everest with a blindfold. If you keep walking
up hill, you will find the top only if your starting point is not too
far away - otherwise you will end up at the top of a smaller mountain/hill.
Best regards,
-John
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|