Hi,
I tried to use 3dttest++ to do it. There is no statistic significant region. I also tried to use 3dDeconvolve to remove the age and gender effects first, then to use 3dTcorr1D to calculate the correlation coefficient, the result is total different and looks reasonable. Is it correct using 3dTcorr1D? Otherwise, I tried to calculate both the spearman and pearson correlation. The regions of statistic significant using spearman are more than the result using pearson. If I assume the correlation is not linear I should use spearman coefficient, right?
Thanks,
Xiaozhen