A Hybrid Multi‑Detector Framework with Bayesian Uncertainty Quantification and Machine Learning Verification for Robust QRS Detection and HRV Analysis
DOI:
https://doi.org/10.29284/bs8x4n45Keywords:
Heart rate variability (HRV), Bayesian bootstrap, MIT-BIH database, multi-detector fusion, QRS detection.Abstract
Heat rate variability (HRV) analysis is crucial for accurate QRS detection. As several detectors struggle with noises and diverse data sources. A hybrid QRS detection algorithm was developed which combines three complimentary detectors. These detectors are Pan-Tompkins, stationary wavelet transform, and Hilbert transform along with machine learning validation and Bayesian uncertainty quantification for HRV metrics.
A SNR adaptive preprocessing pipeline was developed that merges beat candidates from the three detectors. A logistic regression classifier was used which was trained on four interpretable features, that verifies each candidate. HRV parameters were computed with 95 % confidence intervals through Bayesian bootstrap of the Lomb scargle periodogram. Sliding window analysis was used to assessed the reproducibility.
On all 48 records of the MIT‑BIH Arrhythmia Database, the detector achieved a median F1‑score of 0.936 (interquartile range 0.884–0.962) and a median false‑positive rate of 9.07 per minute (IQR 5.54–15.51). On record 100 (clean sinus rhythm), Bayesian 95 % confidence intervals for RMSSD and SDNN were [49.1, 60.3] ms and [43.2, 48.5] ms, respectively. Sliding‑window reproducibility on this record gave a coefficient of variation of 4.7 % for SDNN. On a challenging noisy record (record 105), the detector maintained a sensitivity of 0.967 and F1 of 0.934. A demonstration on locally recorded ECG through self developed ECG acquisition circuit based on IC LM325, the large duration ECG (>6min) yielded plausible HRV metrics (mean RR 735.8 ms, SDNN 39.4 ms, RMSSD 26.5 ms). While other recordings aquired through the circuit were excluded due to insufficient duration, underscoring the need for adequate recording length in HRV analysis.
The proposed work gave accurate QRS detection and HRV analysis for various clinical recordings. This work quantifies built-in uncertainty which enhances the interpretability of the model. When its performance was tested on a consumer grade ECG, it suggests its potentiality for remote monitoring and also the recordings met with the recommended duration guidelines.
References
[1] Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology. Heart rate variability: standards of measurement, physiological interpretation, and clinical use. Circulation. 1996;93(5):1043–1065.
[2] Shaffer F, Ginsberg JP. An overview of heart rate variability metrics and norms. Front Public Health. 2017;5:258.
[3] Peltola MA. Role of editing of R‑R intervals in the analysis of heart rate variability. Front Physiol. 2012;3:148.
[4] Pan J, Tompkins WJ. A real‑time QRS detection algorithm. IEEE Trans Biomed Eng. 1985;32(3):230–236.
[5] Elgendi M. Fast QRS detection with an optimized knowledge‑based method: evaluation on 11 standard ECG databases. PLoS One. 2013;8(9):e73557.
[6] Li C, Zheng C, Tai C. Detection of ECG characteristic points using wavelet transforms. IEEE Trans Biomed Eng. 1995;42(1):21–28.
[7] Kadambe S, Murray R, Boudreaux‑Bartels GF. Wavelet transform‑based QRS complex detector. IEEE Trans Biomed Eng. 1999;46(7):838–848.
[8] Kalyakulina A, Yusipov I, Moskalenko V, et al. LU‑DB: a new open‑access database for electrocardiogram analysis. Sci Data. 2020;7:376.
[9] Sahoo S, Kanungo B, Behera S, Sabut S. Multiresolution wavelet transform based feature extraction and ECG classification to detect cardiac abnormalities. Measurement. 2022;188:110572.
[10] Zidelmal Z, Amirou A, Adnane M, Belouchrani A. QRS detection using S‑transform and Shannon energy. Comput Methods Programs Biomed. 2012;107(2):218–226.
[11] Kwon S, Kim H, Park KS. Validation of heart rate extraction using an iPhone camera in various postures. IEEE J Biomed Health Inform. 2017;21(4):1034–1041.
[12] Singh B, Bharti N, Singh D. Uncertainty quantification in heart rate variability analysis: a review. Biocybern Biomed Eng. 2021;41(3):1029–1046.
[13] Sandercock GR, Bromley PD, Brodie DA. The reliability of short‑term measurements of heart rate variability. Int J Cardiol. 2005;103(3):238–247.
[14] Moody GB, Mark RG. The impact of the MIT‑BIH Arrhythmia Database. IEEE Eng Med Biol Mag. 2001;20(3):45–50.
[15] Nunan D, Sandercock GR, Brodie DA. A quantitative systematic review of normal values for short‑term heart rate variability in healthy adults. Pacing Clin Electrophysiol. 2010;33(11):1407–1417.
[16] Rubin DB. The Bayesian bootstrap. Ann Stat. 1981;9(1):130–134.
[17] Lomb NR. Least‑squares frequency analysis of unequally spaced data. Astrophys Space Sci. 1976;39(2):447–462.
Downloads
Published
Issue
Section
License
This work is licensed under a Creative Commons Attribution 4.0 International License.
This work is licensed under a Creative Commons Attribution 4.0 International License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
