Unveiling Hidden Patterns: Harnessing the Power of Short PPG-Traces for Atrial Fibrillation Detection

Published in IEEE 19th International Conference on Body Sensor Networks (BSN), 2023

Recommended citation: Peris-Lopez, P., Fuster-Barceló, C., Camara, C., & Martin, H. (2023). *Unveiling Hidden Patterns: Harnessing the Power of Short PPG-Traces for Atrial Fibrillation Detection*. IEEE BSN 2023. https://ieeexplore.ieee.org/abstract/document/10331002 https://ieeexplore.ieee.org/abstract/document/10331002

Unveiling Hidden Patterns: Harnessing the Power of Short PPG-Traces for Atrial Fibrillation Detection

Presented at: IEEE BSN 2023, Cambridge, UK


🔍 Summary

This study explores atrial fibrillation (AF) detection through short photoplethysmography (PPG) traces by drawing analogies with:

  • Cardiologists’ acoustic evaluation of cardiac rhythms
  • Musical signal analysis (pitch, timbre, and scalograms)

Using musical features extracted from PPG signals, the work compares:

  • A handcrafted feature approach reaching 89% accuracy
  • A CNN + CBAM attention module achieving up to 95% accuracy

💡 Significance

  • Shows that shortened PPG recordings (instead of ECG) are sufficient and reliable for AF detection.
  • Competitive with FDA-approved ECG-based solutions, but simpler, faster, and cheaper.
  • Demonstrates the universal potential of music-inspired signal interpretation in biomedical sensing.

Let me know when you’re ready for the next one.