Empirical analysis of phase-amplitude coupling approaches

Area of Science:

• Neuroscience
• Signal Processing
• Computational Biology

Background:

• Phase-amplitude coupling (PAC) analysis of brain oscillations offers insights into memory and movement disorders.
• Technical choices in PAC analysis significantly influence results.
• Existing methods like Kullback-Leibler distance have limitations.

Purpose of the Study:

• To empirically explore various analysis design choices for measuring PAC.
• To compare different filtering approaches, including wavelet-based, constant filter, and optimized filtering.
• To introduce and evaluate a time-dependent PAC analysis technique.

Main Methods:

• Compared three filtering approaches: wavelet, constant filter settings, and optimized individual frequency bands.
• Introduced a time-dependent PAC analysis using wavelets.
• Assessed the impact of data duration, oscillation stability, and artifacts on the modulation index.
• Evaluated computational costs.

Main Results:

• Wavelet-based PAC analysis demonstrated superior performance with comparable or lower computational cost.
• Wavelet methods enable examination of temporal changes in PAC.
• PAC measurement reliability is contingent on data duration, oscillation stability, and artifact presence.

Conclusions:

• Wavelet-based PAC analysis is a robust and efficient method for studying brain dynamics.
• Careful consideration of data characteristics and analysis objectives is crucial for selecting appropriate PAC parameters.
• Time-resolved PAC analysis is recommended for initial data exploration.