Spectral parameterization (specparam / FOOOF)

Spectral parameterization (the specparam or FOOOF framework) fits a power spectrum as the sum of an aperiodic component, a 1/f-like function described by an offset and an exponent (optionally with a low-frequency knee), plus a small number of Gaussian peaks, each with a centre frequency, a power above the aperiodic level, and a bandwidth. It replaces ad hoc band-power measures with explicit parameters that separate what actually changed: an oscillatory peak, or the sloping background beneath it.

The method has made the aperiodic exponent a variable of interest in its own right (a putative index of excitation-inhibition balance), but it is a curve fit and inherits the usual cautions: the fit range, the presence of a spectral knee and the peak-detection thresholds all affect the results, a fitted peak is not proof of a true oscillation, and the exponent it returns depends on the frequency range chosen. It parameterizes a spectrum; it does not by itself validate that a rhythm exists.