Hi everybody,
sorry for posting two thread in such little time, but there is something really bothering me.
I am after the association of certain behavioral variables with connectivity of the reading network.
I have 5 seed regions and 5 behavioral variables. The idea is that certain variables will be associated with the connectivity of certain regions but not other and vice versa.
I was looking at my second level effect seed - to - voxel and i noticed something strange: for some reason, i have some of the variables entered as Z score and some other variable entered as row score. None of the variables entered as Z score show any significant association. On the other hand, all the variables entered as raw score show strong significant association. The funny thing is that they show association with the connectivity of all my seeds (and this is already pretty unexpected) but more strikingly, the association are almost exactly overlapping to the effect that I have when I look at the "All subjects" variable. Basically, it seems that all the variables entered as raw score are associated with exactly the connectivity network of each seeds. I have attached some picture for comparison.
At the beginning I thought it was something due to the deameaning (since Z scored variable did not show this behavior), but demeaning should not affect simple correlation (that is what is going on when you test the association of a behavioral variable with the seed to voxel connectivty, as far as I understand).
Have you got any idea about what is going on here ?
Best
A.
sorry for posting two thread in such little time, but there is something really bothering me.
I am after the association of certain behavioral variables with connectivity of the reading network.
I have 5 seed regions and 5 behavioral variables. The idea is that certain variables will be associated with the connectivity of certain regions but not other and vice versa.
I was looking at my second level effect seed - to - voxel and i noticed something strange: for some reason, i have some of the variables entered as Z score and some other variable entered as row score. None of the variables entered as Z score show any significant association. On the other hand, all the variables entered as raw score show strong significant association. The funny thing is that they show association with the connectivity of all my seeds (and this is already pretty unexpected) but more strikingly, the association are almost exactly overlapping to the effect that I have when I look at the "All subjects" variable. Basically, it seems that all the variables entered as raw score are associated with exactly the connectivity network of each seeds. I have attached some picture for comparison.
At the beginning I thought it was something due to the deameaning (since Z scored variable did not show this behavior), but demeaning should not affect simple correlation (that is what is going on when you test the association of a behavioral variable with the seed to voxel connectivty, as far as I understand).
Have you got any idea about what is going on here ?
Best
A.