0
$\begingroup$

I am working on a project that uses EEG signals of the brain to identify emotional states. While surveying the literature, I came across several references where "derived features of bispectrum" are used as features for the purpose of classifying motor and emotional signals.

Although I have a list of some these "derived features" at my disposal, I would like to have some intuition as to which of these features I should use and why consider a bispectrum in the first place. I would highly appreciate it if the physical significance of the bispectrum could be explained.

$\endgroup$
0
$\begingroup$

The bispectrum is technique to detect phase relationships (or phase coupling) between different components of a signal. With EEGs, I can imagine it would be useful to understand how & which signal components (or latent sources captured by a single EEG) are covariates.

The relevance of phase coupling depends on what is being studied by the EEG analysis, for example in this paper, the context is medical. Bispectrum features are used to detect decreases in the stability of the phase relationships of fast rhythms in EEG signals, this is considered an abnormality that corresponds to the existence of Alzheimer's disease.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.