HBClab / NiBetaSeries

Nipype implementation of BetaSeries Correlations (Beta)
https://nibetaseries.readthedocs.io
MIT License
32 stars 28 forks source link

Divide correlations by standard deviation #285

Open jdkent opened 4 years ago

jdkent commented 4 years ago

Describe the bug In the original Rissman (2004) paper, the fisher's z -transformed correlation coefficients were divided by their "known" standard deviation.

The transformed correlation coefficients were then divided by their known standard deviation (1 / sqrt(N-3) , where N is the number of data points used to compute the correlation coefficient) to yield z scores.

To Reproduce Use NiBetaSeries to create correlations

Expected behavior The correlations should be normalized

Desktop (please complete the following information): N/A

nibetaseries version v0.4.3

Additional context N/A