Dear Will,
there are numerous valid approaches, each with advantages and disadvantages. You may want to incorporate prior knowledge (atlas based ROIs or activation locations reported in literature) or use a data-driven approach (e.g. functional localiser). These approaches would work both for small volume correction in SPM maps and for time series extraction for e.g. DCM analysis. Time series extraction would also be possible using peak clusters from an SPM map.
You will have to decide based on your experimental design and hypothesis details which method matches your needs best.
Volkmar
|