Describe the bug
sometimes correlations calculated in nilearn have a rounding error making them slightly greater than 1 or less than -1. This results in a NaN when performing the fisher's r to z transform
To Reproduce
correlate two voxels with the exact same time series.
Expected behavior
fisher's r to z transform should only have numbers in the inclusive range [-1, 1]
Desktop (please complete the following information):
N/A
nibetaseries version
devel (post 0.3.2)
Additional context
This problem is being solved in another pull request
Describe the bug sometimes correlations calculated in nilearn have a rounding error making them slightly greater than 1 or less than -1. This results in a NaN when performing the fisher's r to z transform
To Reproduce correlate two voxels with the exact same time series.
Expected behavior fisher's r to z transform should only have numbers in the inclusive range
[-1, 1]Desktop (please complete the following information): N/A
nibetaseries version devel (post 0.3.2)
Additional context This problem is being solved in another pull request