Addressing Pitfalls in Phase-Amplitude Coupling Analysis with an Extended Modulation Index Toolbox

Area of Science:

• Neuroscience
• Signal Processing
• Computational Biology

Background:

• Phase-amplitude coupling (PAC) is crucial for coordinating neural information processing.
• Existing PAC detection methods can yield spurious or waveform-dependent results.
• Reliable PAC detection is essential for understanding brain function.

Purpose of the Study:

• To develop a novel, accessible tool for reliable phase-amplitude coupling detection.
• To differentiate authentic PAC from spurious or waveform-dependent coupling.
• To provide statistical significance estimation for PAC in continuous and epoched data.

Main Methods:

• Introduction of the Extended Modulation Index (eMI), based on the Modulation Index.
• Application of extreme value statistics for estimating statistical significance.
• Comparison of eMI with direct PAC estimator and standard Modulation Index using simulated and real LFP data.
• Development of a heuristic algorithm for classifying PAC as Reliable or Ambiguous.

Main Results:

• eMI demonstrated comparable sensitivity and specificity to existing PAC measures.
• eMI showed enhanced selectivity in the frequency dimension for phase.
• The heuristic algorithm effectively classified PAC reliability based on spectral properties.
• eMI provides visualizations and introduces the polar phase-histogram for detailed analysis.

Conclusions:

• eMI offers a more reliable and selective method for detecting phase-amplitude coupling.
• The tool addresses known challenges in interpreting PAC.
• eMI provides a robust framework for analyzing neural oscillations and their coordination.
• A freely available MATLAB toolbox (EEGLAB plugin) implements eMI and reference methods.