Hi Varina,
Without much deep thinking, I believe that ranking the targets within
subject, then averaging these ranks across subjects, could give the
information you want, although computing the significance may be a bit
challenging (not impossible, just requires some thinking and maybe some
coding).
Another possibility perhaps would be to mean center (separately) the
seed and the targets' time courses for each subject, then concatenate
them across subjects, and only then compute the correlation coefficients
between the seed and each of the targets.
Not sure if this helps...
All the best,
Anderson
On 13/01/2014 20:50, Varina Wolf wrote:
> Hello Experts,
>
> I have 2 groups of subjects with rs-fMRI, for which I have calculated partial connectivity coefficient maps between 13 target ROIs and a single seed area. I transformed the partial cc maps to zstat maps, then checked the absolute max and min of these zstat maps for each target, for each individual. I know that when -1.96 < z < 1.96 the p >0.05, and am thinking that for each map this would be a way of determining when the partial cc is significant, at least for that individual.
>
> I noticed a few things doing this, which have me thinking. For starters, I noticed that the Z values for each individual, across the target areas show similar values, regardless of the target. For example, for individual one, almost all of the Z mins were 0 and Z max were above or around 2; and for the next individual nearly all of the Z mins had negative values less than -1 or more and Z max were greater than 1 or 2; and so forth on the trend within each individual.
>
> Seeing this, it becomes obvious that averaging Z values across individuals does not seem a good way to find average significant connectivity for a target, since Z seems to depend more on the effect of the individual rather than the target brain area.
>
> So am left wondering the best way to determine a way to rank the significance of the target areas connectivity to the given seed across individuals then between the groups.
>
> Your advice is greatly appreciated!
> Thank you,
> Varina
|