IRASA method

Quoting from the FieldTrip website

IRASA allows distinguishing rhythmic activity from concurrent power-spectral 1/f modulations. The technique virtually compresses and expands the time-domain data with a set of non-integer resampling factors prior to Fourier-based spectral decomposition. As a result, rhythmic components in the power-spectrum are redistributed while the arrhythmic 1/f distribution is left intact. Taking the median of the resulting auto-spectral distributions extracts the power-spectral 1/f component, and the subsequent removal of the 1/f component from the original power-spectrum offers a power-spectral estimate of rhythmic content in the recorded signal.

The math on this doesn't appear too bad, but figuring out an appropriate interface might be a little bit more complicated, including what to return and which classes to have this as a method.

Start by reading the linked FieldTrip IRASA example and the cited paper to understand the resampling and spectral-decomposition requirements. Then determine an appropriate interface, including which classes should expose the method and what it should return; done means the method's scope and outputs are clearly defined and implemented.