An Interpretable DSP Architecture for ECG Rhythm and Premature Ventricular Contraction Detection in MATLAB/Simulink

Electrocardiogram (ECG) monitoring requires reliable signal preprocessing and interpretable detection of abnormal cardiac activity while maintaining low algorithmic complexity. This study presents an integrated digital signal processing architecture developed in MATLAB/Simulink for ECG preprocessing, QRS and R-peak detection, heart-rate estimation, rhythm classification, and premature ventricular contraction (PVC) detection. Real ECG recordings from the MIT-BIH Arrhythmia Database were used for development and evaluation. The preprocessing stage employs minimum-order elliptic IIR high-pass and low-pass filtering, followed by differentiation, squaring, moving-window integration, thresholding, and R-peak detection. RR intervals derived from detected R peaks are used for instantaneous heart-rate calculation and classification of normal rhythm, bradycardia, and tachycardia. PVC detection uses two interpretable beat-level features, an RR-interval ratio and a QRS-energy ratio, combined through fixed rule-based logic. Three initial 10-s windows containing 18 annotated PVCs were used during algorithm development and threshold selection. After the detector parameters were frozen, evaluation was performed on nine additional non-overlapping 10-s windows containing 56 expert-annotated PVCs from MIT-BIH Records 106, 119, and 200. Using a +/-100 ms event-matching tolerance, the held-out evaluation produced 40 true positives, 9 false positives, and 16 false negatives, corresponding to 71.43% sensitivity, 81.63% precision, and 76.19% F1 score. Across the same nine windows, the R-peak detector achieved 94.39% sensitivity, 99.02% positive predictivity, 96.65% F1 score, and approximately 16.91 ms mean absolute timing error. The results demonstrate the feasibility and limitations of a lightweight and explainable streaming ECG-processing architecture and motivate future work incorporating adaptive decision thresholds and broader patient-independent validation.

Authors

Institutions

Publication Details

Journal
medRxiv
Published
2026-11-18
DOI
https://doi.org/10.64898/2026.11.16.26368189
Primary Topic
ECG Monitoring and Analysis
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

An Interpretable DSP Architecture for ECG Rhythm and Premature Ventricular Contraction Detection in MATLAB/Simulink

Mst. Jannatul Ferdous, Abdullah Atif
medRxiv
ECG Monitoring and Analysis
preprint

An Interpretable DSP Architecture for ECG Rhythm and Premature Ventricular Contraction Detection in MATLAB/Simulink

Mst. Jannatul Ferdous, Abdullah Atif
preprint en

Abstract

Electrocardiogram (ECG) monitoring requires reliable signal preprocessing and interpretable detection of abnormal cardiac activity while maintaining low algorithmic complexity. This study presents an integrated digital signal processing architecture developed in MATLAB/Simulink for ECG preprocessing, QRS and R-peak detection, heart-rate estimation, rhythm classification, and premature ventricular contraction (PVC) detection. Real ECG recordings from the MIT-BIH Arrhythmia Database were used for development and evaluation. The preprocessing stage employs minimum-order elliptic IIR high-pass and low-pass filtering, followed by differentiation, squaring, moving-window integration, thresholding, and R-peak detection. RR intervals derived from detected R peaks are used for instantaneous heart-rate calculation and classification of normal rhythm, bradycardia, and tachycardia. PVC detection uses two interpretable beat-level features, an RR-interval ratio and a QRS-energy ratio, combined through fixed rule-based logic. Three initial 10-s windows containing 18 annotated PVCs were used during algorithm development and threshold selection. After the detector parameters were frozen, evaluation was performed on nine additional non-overlapping 10-s windows containing 56 expert-annotated PVCs from MIT-BIH Records 106, 119, and 200. Using a +/-100 ms event-matching tolerance, the held-out evaluation produced 40 true positives, 9 false positives, and 16 false negatives, corresponding to 71.43% sensitivity, 81.63% precision, and 76.19% F1 score. Across the same nine windows, the R-peak detector achieved 94.39% sensitivity, 99.02% positive predictivity, 96.65% F1 score, and approximately 16.91 ms mean absolute timing error. The results demonstrate the feasibility and limitations of a lightweight and explainable streaming ECG-processing architecture and motivate future work incorporating adaptive decision thresholds and broader patient-independent validation.

medRxiv
Varendra University (BD)
Good health and well-being
ECG Monitoring and Analysis
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon