Fitting Oscillations and 1/f (FOOOF)

Please refer to the FOOOF Github Repository for detailed methods of the FOOOF toolbox and to our own LNCD FOOOF Code

The Fitting Oscillations and 1/f (FOOOF) protocol derives the 1/f spectral slope (referred to as the aperiodic exponent) and an offset. The steepness of the slope (i.e., the exponent) has been previously linked to shifts in the Excitation Inhibition Balance, reductions of which reflect shifts toward excitatory activity relative to inhibitory, while increases in the exponent reflect shifts toward inhibitory activity. Independently, the PSD offset is thought to reflect broadband shifts in power, reflecting levels of neuronal population spiking.

FOOOF conceives of a model of the power spectrum as a combination of two distinct functional processes:

• An aperiodic component, reflecting 1/f like characteristics, with
• A variable number of periodic components (putative oscillations), as peaks rising above the aperiodic component

FOOOF Protocol

• Power spectral density (PSD) is calculated across the continuous resting state using Welch’s method implemented in MATLAB (2s Hamming window, 50% overlap)
• The aperiodic component L(f) was extracted from the 1-50Hz frequency range of each power spectrum
• FOOOF settings:
  • aperiodic_mode = ‘fixed’
  • peak_width_limits = [0.5, 12]
  • min_peak_height = 0
  • peak_threshold = 2
  • max_n_peaks = 4
  • default settings otherwise)
• When the PSD is plotted on a log-log axis, the constant offset b refers to the y-intercept and the aperiodic exponent x corresponds to the slope of a line
• The average R^2 of the spectral fits was 0.98 for the eyes closed condition and 0.97 for the eyes open condition, reflecting good fits