spectral_connectivity.statistics.fisher_z_transform#

fisher_z_transform(coherency1, bias1, coherency2=0, bias2=0)[source]#

Transform the coherence magnitude to an approximately normal distribution.

If two coherencies are provided, then the function returns the fisher z transform of the difference of the coherencies with coherency1 - coherency2.

Parameters:

  • coherency1 (complex array) – The complex coherency between signals

  • bias1 (float) – The bias from independent estimates of the frequency domain.

  • coherency2 (complex array, optional) –

  • bias2 (float, optional) – The bias from independent estimates of the frequency domain.

Returns:

fisher_z_transform – Either the difference from 0 mean or, if another coherency is provided, the difference from that coherency.

Return type:

real array