Dear Justine,
> I have a conundrum with my DARTEL-VBM data.
> I have this region showing a difference - see image attached. But this
> single blob traverses two lobes: temporal and parietal. (It also follows
> the border of the insula but does not overlay it.)
> The question arises - what is the most likely explanation?
> - there is a single blob in one of the regions, but there is some inaccuracy
> involved in relating this to the canonical T1. If so how to deal with this?
> - there are 2 blobs in 2 regions, which (either coincidentally or not) are
> contiguous
The suggestion of looking at your results on a template other than a
single-subject template is a good one. I would suggest both the gray
matter template generated by DARTEL (normalized to MNI space, assuming
that your analysis was conducted in MNI space), and also perhaps an
average structural image of all of the subjects. That may give you a
slightly more accurate idea of the spatial resolution/location of your
results.
Also, as you no doubt know, smoothing can also contribute to putting
results outside of what is obviously gray matter.
(If this was an fMRI result I would also suggest looking at the
location of peaks in single-subject data, but that doesn't make sense
for a VBM study.)
Finally, you can use a probabilistic atlas to identify where the peak
lies, or where the cluster spans (XX% of the cluster lies in region 1,
XX% in region 2). Such probabilistic atlases include the SPM Anatomy
Toolbox (though I'm not sure if it covers the region you are
interested in), or atlases from Alexander Hammers (you can search
previous posts by him to find out more). The atlas-based approach may
not give you a single answer, but it's a way of quantifying the
options and being objective about your anatomical localization (though
personally I think this should come after having a closer look at the
templates—e.g. tedious anatomy, and then probabilistic atlases).
All of that being said, I don't know if there is a good way to tell
between the two possibilities that you list. If you have a result
that is statistically significant, it is certainly appropriate to
report it; as far as the interpretation, I think being straightforward
about the level of (un)certainty is the best way to go. Though,
hopefully some of these other suggestions will clear things up.
Hope this helps!
Best regards,
Jonathan
--
Dr. Jonathan Peelle
Department of Neurology
University of Pennsylvania
3 West Gates
3400 Spruce Street
Philadelphia, PA 19104
USA
http://jonathanpeelle.net/
|