[FieldTrip] WPLI question

[Martin Vinck]

Hi Mateo,

As it is computed in ft_connectivity_wpli it preserves the information about the sign, i.e. whether the average imaginary component of the cross-spectrum is positive or negative.

That is, it estimates the average imaginary component of the cross-spectrum divided by the average magnitude of this imaginary component.

By taking the absolute value over the output one gets a measure of the strength of the interaction,- WPLI as it is defined in the paper, by taking the sign one gets a measure of the direction of the interaction.

The debiased WPLI estimator gets negative however as a consequence of the debiasing method - different issue.

Best,

Martin.