help
help > Basic yet tricky question
Sep 18, 2014 01:09 PM | Macià Buades-Rotger
Basic yet tricky question
Dear all,
First of all, congrats on the toolbox, it works really well. I am currently performing seed-to-voxel connectivity analysis comparing between conditions. My connectivity measure is the bivariate correlation coefficient. So my question is: do 2nd level positive/negative connectivity values reflect positive/negative correlation values? That is, does the direction of connectivity values between conditions determine the sign of the correlation coefficient (+ vs -), or does it only determine the magnitude (high |r| in one condition vs high |r| in the other)?
Let's say my conditions are A and B. I run a [1 -1] contrast, I check the seed-to-voxel results explorer, select the "positive contrast" and find a "red" area. Then I select "negative contrast" and find a "blue" area.
Does the red area show a significantly higher correlation with my seed, or does it show a significantly higher and positive correlation with my seed? Conversely, is the blue area highly and negatively correlated with the seed in this contrast, or is it just higher in the opposite contrast? Or is it all just me overthinking?
I've checked other threads on interpretational issues and looked up papers in which a directional relationship is assumed, but still I can't fully get my head around this.
A short clarification would be greatly appreciated.
All the best,
Macià
First of all, congrats on the toolbox, it works really well. I am currently performing seed-to-voxel connectivity analysis comparing between conditions. My connectivity measure is the bivariate correlation coefficient. So my question is: do 2nd level positive/negative connectivity values reflect positive/negative correlation values? That is, does the direction of connectivity values between conditions determine the sign of the correlation coefficient (+ vs -), or does it only determine the magnitude (high |r| in one condition vs high |r| in the other)?
Let's say my conditions are A and B. I run a [1 -1] contrast, I check the seed-to-voxel results explorer, select the "positive contrast" and find a "red" area. Then I select "negative contrast" and find a "blue" area.
Does the red area show a significantly higher correlation with my seed, or does it show a significantly higher and positive correlation with my seed? Conversely, is the blue area highly and negatively correlated with the seed in this contrast, or is it just higher in the opposite contrast? Or is it all just me overthinking?
I've checked other threads on interpretational issues and looked up papers in which a directional relationship is assumed, but still I can't fully get my head around this.
A short clarification would be greatly appreciated.
All the best,
Macià
Threaded View
Title | Author | Date |
---|---|---|
Macià Buades-Rotger | Sep 18, 2014 | |
Alfonso Nieto-Castanon | Sep 22, 2014 | |
Macià Buades-Rotger | Sep 22, 2014 | |