Benchmark Algorithm for Surface Electromyography-Based Periodic Limb Movement Detection.

Area of Science:

• Sleep Medicine
• Biomedical Engineering
• Signal Processing

Background:

• Periodic limb movement disorder (PLMD) is characterized by involuntary leg movements during sleep.
• Current detection methods rely on manual labeling of surface electromyography (sEMG), which is subjective and may underestimate movements.
• There is a need for standardized, objective algorithms for accurate LM detection.

Purpose of the Study:

• To develop and evaluate an automated algorithm for detecting leg movements (LMs) in PLMD.
• To establish a benchmark algorithm compatible with clinical scoring manuals.
• To reduce subjectivity and improve reproducibility in LM detection.

Main Methods:

• Proposed an algorithm evaluating sEMG in 0.5-second segments using a percentage parameter.
• Tested the algorithm on sEMG data from 20 individuals in clinical settings.
• Compared algorithm performance against manual labeling and scoring manual interpretations.

Main Results:

• The algorithm achieved an F1 score of 0.751 at a 40% parameter setting.
• Achieved a recall of 0.895 and a precision of 0.604.
• Demonstrated superior performance compared to strict scoring manual interpretation, which yielded zero recall and precision.

Conclusions:

• The developed algorithm shows promise for objective and reproducible LM detection in PLMD.
• It offers a standardized approach to complement existing clinical scoring manuals.
• This work represents a step towards an automated benchmark for PLM labeling.