Dear SPMers,
we combine two imaging modalities (PET and DTI) to investigate an
association between microstructural and metabolic alterations in diseased
subjects.
What would be the most appropriate way to analyze such data?
May the following suggestion have sense:
a) one compares a group of diseased subjects with a group of control subjects
within one modality (let's say DTI);
b) one extracts (raw) values from significant clusters for the diseased
subjects;
c) one conducts a multiple regression analysis on the PET data of the same
diseased subjects introducing the values from the siginificant clusters of the
analysis "a)" as variables.
Since measures of the two modalities are of totally different magnitudes (but
both normally distributed), would be an SnPM (statistical nonParametric
Mapping) tool more suitable for such analysis?
Any other ideas?
Thanks,
Igor Yakushev
|