I'm not sure you can. Unless you have multiple samples, you can't
compute a variance. If you can't compute a vvariance, then you can't
compute a significance. Maybe you can do some kind of bootstrapping
On 9/3/19 1:49 PM, Matthieu Vanhoutte wrote:
> External Email - Use Caution
>
> So
External Email - Use Caution
So if I obtained two correlation maps from the same population but based on
different input parameters, how could I demonstrate that vertex correlation
values are statistically « strongest »/« lowest » in one of the map compared
to the other ?
You want to see if the correlation values themselves? You can't do that
(unless you have a bunch of correlation maps). You have to incorporate
your question/hypothesis into the design and contrast.
On 9/2/2019 11:14 AM, Matthieu VANHOUTTE wrote:
> External Email - Use Caution
>
> Dear