Design and validation of a periodic leg movement detector

Area of Science:

• Sleep Medicine
• Biomedical Engineering
• Computational Neuroscience

Background:

• Periodic Limb Movements (PLMs) are involuntary muscle contractions during sleep that can fragment sleep and impact health.
• Accurate scoring of leg movements (LM) during polysomnography is crucial for diagnosing sleep disorders.
• Existing automated methods for PLM detection face challenges with noise and artifact interference.

Purpose of the Study:

• To develop and validate the Stanford PLM automatic detector (S-PLMAD), a robust automated system for scoring PLMs.
• To improve the accuracy and efficiency of PLM detection in nocturnal polysomnography (NPSG).
• To assess the performance of S-PLMAD across diverse sleep disorder populations.

Main Methods:

• The S-PLMAD algorithm was developed using NPSG data from the Wisconsin Sleep Cohort (WSC) and Stanford Sleep Cohort (SSC).
• The algorithm was refined based on American Association of Sleep Medicine scoring rules, incorporating adaptive noise cancellation and adjustable thresholds.
• Validation involved comparing S-PLMAD outputs against expert visual scoring in 78 studies across various sleep conditions.

Main Results:

• The S-PLMAD demonstrated high correlation with expert visual scoring for PLM indices (r² = 0.94 in both WSC and SSC).
• The detector effectively handles noise, cardiac interference, and artifacts, and excludes leg movements associated with respiratory events.
• S-PLMAD provides comprehensive metrics including PLM indices for sleep (PLMS) and wake, and periodicity index.

Conclusions:

• The S-PLMAD is a robust, high-throughput automated detector for Periodic Limb Movements.
• The system performs accurately in both healthy individuals and patients with various sleep disorders.
• S-PLMAD offers a reliable tool for objective PLM scoring in clinical and research settings.