Hello,
I ran DCM analysis on n=18 stroke data set and was able to get a clear winning model. Thereafter, I correlated the DCM.Ep.B values of the winning model with some clinical scores. Because of the variability in the subject data(differences in lesion location) I decided to also extract the BMA scores for each connection/subject and run correlations with the clinical data as well. When presenting the results should I only include the significant correlations that I got with the B parameters?or is it acceptable to present correlations with both B parameter scores and BMA scores(even it its for the same modulatory connection)?
Also, if I were to divide my 18 subjects in to two groups(ex:cortical vs. non-cortical lesioned subjects) then could I use the models created using an 18 subject group level coordinate to compare between these two groups (since that way we would be comparing the same network)?or do you recommend creating models for each group based on the maximum activation coordinates of each group separately?
I would greatly appreciate any feedback.
thank you,
Thushini
|