The MSPTDfast photoplethysmography beat detection algorithm: design, benchmarking, and open-source distribution.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Physiological Monitoring

Background:

• Photoplethysmography (PPG) is crucial for physiological monitoring in clinical and consumer devices.
• Accurate heartbeat detection in PPG signals is essential for reliable physiological data.
• The existing Multi-scale Peak & Trough Detection (MSPTD) algorithm offers high accuracy but lacks computational efficiency.

Purpose of the Study:

• To develop a computationally efficient, open-source implementation of the MSPTD algorithm for PPG beat detection.
• To introduce MSPTDfast (v.2) as an optimized solution for faster and accurate heartbeat detection.

Main Methods:

• Evaluated five potential improvements to the original MSPTD algorithm on four datasets.
• Designed MSPTDfast (v.2) by integrating efficiency-enhancing modifications while preserving a high F1-score.
• Benchmarked MSPTDfast (v.2) against state-of-the-art algorithms on four additional datasets.

Main Results:

• MSPTDfast (v.2) significantly reduced execution time (1/3 to 1/20th of MSPTD) by downsampling PPG signals to 20 Hz and optimizing scalogram calculations.
• Achieved a comparable F1-score to the original MSPTD algorithm during internal validation.
• Demonstrated the highest F1-score alongside MSPTD during benchmarking, with competitive execution times.

Conclusions:

• MSPTDfast (v.2) provides an accurate and highly efficient solution for PPG beat detection.
• The open-source availability of MSPTDfast (v.2) in a Matlab toolbox facilitates wider adoption and research.
• This optimized algorithm enhances the feasibility of real-time physiological monitoring using PPG signals.