Stringology-based motif discovery for electrophysiological time series: A framework for temporal pattern analysis...

Area of Science:

• Computational Neuroscience
• Data Science
• Biomedical Engineering

Background:

• Traditional time series analysis often relies on global statistics, overlooking specific recurrent temporal patterns.
• Identifying meaningful sequential structures in complex data like electroencephalography (EEG) remains challenging.

Purpose of the Study:

• To introduce a computational framework using stringology for discovering and characterizing temporal motifs in numerical time series.
• To apply this framework to EEG data from children with and without attention-deficit/hyperactivity disorder (ADHD) to identify group differences.

Main Methods:

• Adapted order preserving matching (OPM) and Cartesian tree matching (CTM) algorithms from stringology.
• Applied the framework to multichannel EEG recordings from 121 children (61 ADHD, 60 controls).
• Quantified motif frequencies, lengths, and hierarchical structures, validating differences using Mann-Whitney U test with FDR correction.

Main Results:

• Discovered motifs with high support (≥0.9) across EEG channels.
• ADHD participants showed significantly higher motif frequencies, shorter OPM motif lengths, and greater gradient instability.
• CTM analysis revealed reduced hierarchical complexity in ADHD, including shallower Cartesian trees and fewer hierarchical levels.

Conclusions:

• Stringology-based motif analysis effectively captures structural, stability, and hierarchical differences in temporal patterns, surpassing conventional signal processing.
• The framework offers a generalizable computational tool for analyzing sequential biomedical data, particularly EEG.
• This approach has potential applications in various domains requiring precise characterization of recurrent structures in numerical time series.