Dear SPM experts,
I conducted VBM analysis on two groups of subjects (group A and group B) and got gray matter volume maps for each subject. Then, I am interested in whether the two groups show common brain regions with a significant correlation between a continuous variable (e.g., age) and GM volume. To address this issue, I'm going to conduct a conjunction analysis in SPM12. I have several questions regarding the procedure.
I learned from the previous mails that if I want to find the shared brain regions for only one group of subjects with two variables (e.g., IQ and language ability), I can build a design matrix with IQ and language ability as independent variables. Then, in the 'Result' section, I choose the two 'con' files for these two variables >> done >>conjunction hypothesis to get the shared brain regions for two interested variables. However, I have two groups of subjects. My plan is to build a design matrix with an interaction between group (factor) and age so that I will get two columns: groupA*age and groupB*age. Then, in the 'Result' section, I do the same thing, choose the two 'con' files for groupA*age and groupB*age>>done>>conjunction hypothesis. I wonder whether this is the correct way to get the common region with the correlation between GM and age for the two groups?
Alternatively, I found a previous mail about the minimum-statistics approach by Nichols et al., (2005) [ https://www.jiscmail.ac.uk/cgi-bin/wa-jisc.exe?A2=ind2108&L=SPM&D=0&X=D6A0911B6773C01F93&Y=kuhnke%40cbs.mpg.de&P=41228 ]. I also have some questions. If I understand correctly, I can do GLM with age as an independent variable and GM as the dependent variable for two groups separately. Then, I get two T maps of age-GM correlation for group A and group B. By using the Imcalc function, I get a T map that shows the min t value of the two original T maps. I wonder what should I do next to test the significance of this T map?
Thanks in advance.
Best,
Di
|