Researchers' phase-based method tracks brain oscillation patterns in real time

The team — Robison J. Santos-Silva, Bruno R. R. Boaretto, Thiago L. Prado, and Roberto C. Budzinski — proposes a framework based on spatial ordinal patterns to characterize spatiotemporal dynamics in oscillatory systems. Unlike conventional methods that operate on amplitude, this approach acts directly on the phase, adding extra patterns for near-equal phases. The resulting symbolic representation encodes local spatial ordering, capturing both phase gradients and synchronized clusters in a single unified scheme. From this construction, they define a spatial permutation entropy that quantifies the diversity of spatiotemporal patterns at each point in time, enabling the detection of transient dynamics and regime transitions as they occur.

This method was tested on synthetic oscillator networks across multiple spatiotemporal regimes and on resting-state EEG recordings from human volunteers. It successfully distinguishes phase-locked states that share identical global synchronization levels but have distinct spatial organization, as well as partially synchronized states. Most notably, the approach can differentiate different conditions within individual volunteers, suggesting applications in personalized analysis. By making hidden neural dynamics visible in real time, this work offers a new tool for researchers studying everything from brain disorders to engineered network behavior.