Dear Steven,
> We are comparing two groups of 13 subjects with DARTEL VBM8. I know this
> is underpowered, but wondered if there is any way to use the multiple T1
> images that were acquired for some of the subjects (in the same session). My
> initial thought is to process all T1s and use the quality check to see if
> there is a "better" image when I have two for a given subject. Are there
> better ways to use multiple T1s?
In the "new segment" option in SPM8, you can specify multiple
modalities of images for segmentation. I assume you could also
provide SPM with 2 T1 images. SPM will then use the combined
information across images to assign tissue class probabilities, and in
general I would think that more data would provide a more accurate
result. However, I think you'd want to process all of your subjects
the same way. So if you only have multiple T1s for a subset, I would
probably stick with the approach you suggest—see if you can identify a
"better" image. (Or, if this is hard to determine, then just use the
first.) But if anyone else has alternate approaches, I'd be very
interested to hear!
Best regards,
Jonathan
|