Hi,
> If one wanted to extract fMR data from a particular small region of the brain
> - let's say the anterior inferior cingulate cortex - is this the best way to
> handle in spm99?
>
> 1. draw VOIs for each subject (how can you make sure it's exactly the same are
> in each subject?)
> 2. estimate a specified model just for the VOIs?
A team of us have written an SPM toolbox that is designed to do just this sort
of thing. It is currently more or less in alpha state, but we have started to
use it in our lab. You would be welcome to a copy of the software if you want
to try it. One approach to ROI analysis is indeed to define the VOI for each
subject (without ever normalizing the data to a template). This has
advantages if you have some reliable method of knowing where the region is -
for example from anatomy, or another task from the same subject. Another
approach is to spatially normalize the data, and define the region on the
template. In neither case can you be sure your VOI is in exactly the same
place - in the first case due to imprecision in the metrics you have used to
define the region, in the second, because we cannot easily know how well the
normalization has succeeded in overlapping homologous areas,
Best,
Matthew
|