Dear Alfonso, dear Forum
1) we computed bivariate correlations and compared Group A vs. Group B. We also corrected for age, setting up 1 -1 0 (with age = 0).
However, how is this practically done in Conn that we can correct for the influence of age? Is Conn doing a partial correlation? In other words, how is the correction for age done using bivariate correlations?
2) We compared group-differences by bivariate correlations and by semi-partial correlations. Results look dramatically different. None of the group-differences using bivariate correlations was seen using semi-partial correlations (instead other connections show up). In the manual, it is written:
"Most people seem to be using the simpler bivariate correlations (as a measure of 'total' functional connectivity between two areas). Semi-partial correlations are used when you want to obtain instead the 'unique' contribution of a given source on a target area (controlling for the contributions of other additional source areas)."
In fact, it seems more logical to use semi-partial correlations. We are hence puzzled which results should be reported at the end. Although both results can be used to explain our hypothesis, we want to report the correct methodological approach of course.
Best wishes and thanks
Lars
1) we computed bivariate correlations and compared Group A vs. Group B. We also corrected for age, setting up 1 -1 0 (with age = 0).
However, how is this practically done in Conn that we can correct for the influence of age? Is Conn doing a partial correlation? In other words, how is the correction for age done using bivariate correlations?
2) We compared group-differences by bivariate correlations and by semi-partial correlations. Results look dramatically different. None of the group-differences using bivariate correlations was seen using semi-partial correlations (instead other connections show up). In the manual, it is written:
"Most people seem to be using the simpler bivariate correlations (as a measure of 'total' functional connectivity between two areas). Semi-partial correlations are used when you want to obtain instead the 'unique' contribution of a given source on a target area (controlling for the contributions of other additional source areas)."
In fact, it seems more logical to use semi-partial correlations. We are hence puzzled which results should be reported at the end. Although both results can be used to explain our hypothesis, we want to report the correct methodological approach of course.
Best wishes and thanks
Lars