Characterizing pink and white noise in the human electroencephalogram

Area of Science:

• Neuroscience
• Signal Processing
• Biophysics

Background:

• Human electroencephalogram (EEG) power spectra contain brain oscillations and non-oscillatory noise.
• Previous methods for estimating pink and white noise in EEG have methodological flaws.
• Understanding these noise components is crucial for accurate analysis of neural activity.

Purpose of the Study:

• To propose and validate a novel approach for separating pink and white noise from human EEG power spectra.
• To compare the new method with established procedures using simulated and real EEG data.
• To investigate the topographic characteristics and neural origins of pink and white noise.

Main Methods:

• Development of a new algorithm to extract distinct pink (1/f) and white noise estimates from EEG power spectra.
• Validation using simulated datasets and comparison with existing techniques.
• Application to a new dataset of resting eyes-open (EO) and eyes-closed (EC) human EEG recordings across 60 participants and 30 electrodes.

Main Results:

• Successful extraction of valid pink and white noise estimates from 5400 individual EEG spectra.
• Pink noise exhibited a central scalp topography, white noise an occipital distribution, and alpha oscillations a parietal topography.
• Eyes-closed conditions showed globally greater pink and white noise power compared to eyes-open conditions.

Conclusions:

• The new method provides valid estimates of pink and white noise in human EEG, differentiating them from oscillatory activity.
• Distinct topographies suggest separate neural origins for pink noise, white noise, and alpha oscillations.
• This methodology enhances the assessment of neural activity in various populations and can be applied across scientific fields.