Understanding ECG Signal Analysis: A Technical Breakdown

The Critical Role of ECG Signal Analysis

ECG signal analysis is the foundation of medical heart rhythm monitoring, decoding the heart's electrical activity into waveforms that reveal arrhythmias, ischemia, and conduction disorders. This technical breakdown from Vositone — a 16-year Shenzhen wearable OEM with 70+ R&D engineers and ISO 13485 capability — covers ECG waveform components, signal acquisition, processing pipelines, feature extraction, arrhythmia detection, and how PPG compares to ECG. Importantly, Vositone consumer devices use PPG optical sensing, not diagnostic ECG; this article is technical education for engineers, product teams, and OEM clients. For detection algorithm specs, see our Arrhythmia Detection Specifications.

Medical disclaimer: This article is for technical education only. ECG signal analysis requires medical-grade hardware and certified software for clinical use. Vositone consumer wearables are general wellness devices and do not provide diagnostic ECG. Furthermore, nothing in this article constitutes medical advice. For heart concerns, consult a qualified healthcare professional.

• ECG Waveform Fundamentals
• ECG Signal Acquisition
• Signal Processing Pipeline
• Feature Extraction and Arrhythmia Detection
• PPG vs ECG: Key Differences
• Vositone OEM ECG Capability
• Limitations and Medical Disclaimer
• Frequently Asked Questions
• Build Your Heart Health Wearable

ECG Waveform Fundamentals

The P-QRS-T Complex

In practice, a standard ECG waveform repeats with each heartbeat and consists of three key deflections. First, the P wave represents atrial depolarization (the atria contracting), lasting 80-100ms. Next, the QRS complex represents ventricular depolarization (the ventricles contracting), normally under 120ms wide. Then, the T wave represents ventricular repolarization (the ventricles recovering), lasting 160-200ms. Together, these deflections tell the full story of one cardiac cycle.

Key Intervals and Segments

Similarly, clinicians measure intervals between waveform components to detect conduction abnormalities. Specifically, the PR interval (start of P to start of QRS) normally ranges 120-200ms; prolongation suggests first-degree heart block. Meanwhile, QRS duration under 120ms indicates normal ventricular conduction; widening suggests bundle branch block or ventricular rhythm. Finally, the QT interval (start of QRS to end of T) normally ranges 350-440ms and must be corrected for heart rate (QTc); prolongation increases risk of torsades de pointes.

Normal Heart Rate and Rhythm

Furthermore, a normal resting heart rate ranges 60-100 bpm, with regular R-R intervals (time between consecutive QRS peaks). By contrast, rates below 60 bpm are bradycardia, while rates above 100 bpm are tachycardia. Notably, regularity of R-R intervals is the first check for arrhythmia — irregularly irregular intervals strongly suggest atrial fibrillation, whereas regularly irregular patterns may indicate premature contractions or heart block.

ECG Signal Acquisition

Electrodes and Lead Configuration

For example, medical ECG uses Ag/AgCl electrodes placed on the skin to detect the heart's electrical potentials. Typically, a standard 12-lead ECG uses 10 electrodes to produce 12 views of the heart. On the other hand, single-lead wearable ECG devices use two electrodes (often on the chest or wrist + finger) and capture one lead, typically Lead I or a modified lead. Consequently, single-lead ECG is sufficient for rhythm analysis but cannot detect all ischemic changes that require multiple leads.

Sampling Rate and Resolution

Specifically, medical ECG systems sample at 250-1000 Hz with 12-16 bit resolution, capturing the full frequency spectrum of the QRS complex (which contains energy up to 40-100 Hz). By comparison, consumer wearable ECG devices often sample at 100-250 Hz, which is adequate for rhythm and R-peak detection but may lose fine QRS morphology detail. Therefore, higher sampling rates improve R-peak timing accuracy and HRV precision.

Noise Sources in Wearable ECG

Moreover, wearable ECG faces significant noise challenges. Above all, motion artifact from body movement is the largest noise source, often overwhelming the signal during exercise. Additionally, baseline wander from respiration and electrode movement shifts the signal baseline. Likewise, power line interference (50/60 Hz) adds sinusoidal noise, and electromyography (EMG) from muscle contraction adds high-frequency noise. Thus, effective filtering and signal quality assessment are critical for reliable wearable ECG. For sensor optimization, see our Sensor Calibration Steps Guide.

Signal Processing Pipeline

Filtering Strategy

In practice, ECG signal processing begins with filtering. First, a bandpass filter (typically 0.5-40 Hz) removes low-frequency baseline wander and high-frequency muscle noise. Second, a notch filter (50 or 60 Hz) removes power line interference. Furthermore, some pipelines use a high-pass filter at 0.5-1 Hz to remove baseline wander, and a low-pass filter at 30-40 Hz to smooth the waveform. However, over-filtering can distort the QRS morphology, so filter design must balance noise removal with signal fidelity.

R-Peak Detection (Pan-Tompkins Algorithm)

Similarly, the Pan-Tompkins algorithm is the most widely used method for R-peak detection in real-time ECG. Essentially, it applies bandpass filtering, differentiation, squaring, and moving-window integration to enhance QRS complexes, then uses adaptive thresholding to detect peaks. Additionally, modern variants include search-back logic to recover missed beats and refractory period logic to reject false detections. Ultimately, accurate R-peak detection is the foundation for all downstream HRV and arrhythmia analysis.

Signal Quality Assessment

Furthermore, before analyzing any ECG segment, the system must assess signal quality. Specifically, quality indices include signal amplitude, noise level, motion detection from accelerometers, and percentage of valid R-peaks. Accordingly, low-quality segments should be flagged and excluded from analysis, rather than producing unreliable results. For comparison, Vositone's PPG pipeline uses similar quality gating — for measurement precision details, see our Health Data Precision Guide 2026.

Feature Extraction and Arrhythmia Detection

Heart Rate Variability Features

For example, after R-peak detection, the system computes R-R intervals and derives HRV features. On one hand, time-domain features include SDNN (standard deviation of all NN intervals, reflecting overall variability) and RMSSD (root mean square of successive differences, reflecting parasympathetic tone). On the other hand, frequency-domain features include LF (low frequency, 0.04-0.15 Hz, mixed sympathetic/parasympathetic) and HF (high frequency, 0.15-0.4 Hz, parasympathetic/respiratory). Collectively, these features support stress assessment, recovery scoring, and autonomic balance analysis.

Arrhythmia Classification Features

Specifically, arrhythmia detection extracts features beyond HRV. For instance, atrial fibrillation detection looks for irregularly irregular R-R intervals and absence of P waves. Similarly, premature contractions appear as early QRS complexes followed by a compensatory pause. In contrast, ventricular tachycardia shows wide QRS complexes (>120ms) at high rates (>100 bpm), while bradyarrhythmias show slow rates or pauses >2 seconds. Consequently, machine learning models (CNN, LSTM) can classify these patterns from raw ECG segments, but require large labeled datasets and clinical validation.

From Detection to Clinical Decision

Moreover, automated arrhythmia detection is only the first step. Admittedly, a consumer wearable can flag "possible AFib" or "irregular rhythm," but confirmation requires a doctor reviewing a medical ECG. Meanwhile, regulatory pathways (e.g., FDA De Novo, CE MDR) require clinical trials demonstrating sensitivity and specificity. Accordingly, Vositone supports OEM clients with algorithm licensing and clinical validation partnerships — for detection performance specs, see our Arrhythmia Detection Specifications.

PPG vs ECG: Key Differences

In practice, ECG and PPG measure different physiological signals. Specifically, ECG captures the heart's electrical command signal, while PPG captures the mechanical blood flow that results. Furthermore, ECG can identify specific arrhythmia types from waveform morphology, whereas PPG can only infer irregularity from pulse interval variability. For consumer wellness, PPG provides excellent heart rate and trend data at low power. For diagnostic arrhythmia detection, however, ECG remains the gold standard. For user-facing response guidance, see our Arrhythmia Response Guide.

Vositone OEM ECG Capability

ECG Module Customization

In practice, Vositone offers ECG hardware module integration for OEM clients who need diagnostic-grade or wellness-grade ECG in their wearable. Notably, options include single-lead ECG with Ag/AgCl or dry electrodes, sampling rates from 100-500 Hz, and integrated analog front-end (AFE) chips. Additionally, we can integrate ECG alongside PPG, SpO2, and other sensors in a single device, with custom firmware and algorithm licensing.

Clinical Validation and Regulatory Support

Similarly, Vositone supports OEM clients through clinical validation: study design, hospital partnership access (3 hospital partnerships), data collection, and algorithm performance analysis. Furthermore, we can help prepare documentation for FDA, CE MDR, or other regulatory pathways, working alongside your regulatory team. Ultimately, our ISO 13485 certified production and 16-year OEM track record ensure quality and traceability.

Production and Scale

Finally, Vositone produces ECG-enabled wearables at our 10,000 sqm factory with 50,000+ monthly capacity (peak 80,000). Specifically, MOQ starts at 300 units for white-label, with 10-15 day sampling and 30-55 day mass production. Moreover, our 0.3% defect rate, full certification support (CE/FCC/RoHS/ISO9001, ISO 13485 available), and 100% customer IP ownership ensure a reliable partnership. Learn more at our ODM and OEM Custom Wearable Solutions page.

Limitations and Medical Disclaimer

Technical Boundaries

For example, this article describes general ECG signal analysis principles. Importantly, actual implementation requires medical-grade hardware, certified algorithms, and clinical validation for any diagnostic use. Furthermore, consumer wearable ECG (even with electrodes) may not meet clinical accuracy standards, and single-lead ECG cannot detect all cardiac conditions. Therefore, always refer to regulatory guidelines and qualified medical professionals for clinical applications.

Vositone Consumer Device Boundary

Specifically, Vositone consumer wearables use PPG optical sensing for heart rate and wellness trends. Notably, they do not include diagnostic ECG unless specifically customized for an OEM client. Consequently, PPG-derived irregular pulse alerts are wellness notifications, not medical diagnoses. For any heart symptom, consult a doctor. The American Heart Association provides expert information on arrhythmia and ECG.

Frequently Asked Questions

ECG Signal Fundamentals

Q1: What are the P wave, QRS complex, and T wave in an ECG?

A1: To begin with, the P wave represents atrial depolarization (atria contracting), lasting 80-100ms. Next, the QRS complex represents ventricular depolarization (ventricles contracting), normally under 120ms wide. Then, the T wave represents ventricular repolarization (ventricles recovering), lasting 160-200ms. Together, they form one complete cardiac cycle, and their shape, timing, and intervals reveal heart rhythm and conduction health.

Q2: What is the Pan-Tompkins algorithm and why is it used?

A2: Essentially, the Pan-Tompkins algorithm is the most widely used real-time method for detecting R-peaks in ECG signals. Mechanically, it applies bandpass filtering, differentiation, squaring, and moving-window integration to enhance QRS complexes, then uses adaptive thresholding to identify peaks. Accordingly, it is valued for its balance of accuracy and computational efficiency, making it suitable for wearable devices with limited processing power.

Q3: What sampling rate does a wearable ECG need?

A3: As a best practice, medical ECG systems sample at 250-1000 Hz with 12-16 bit resolution to capture full QRS morphology. By contrast, consumer wearable ECG can work at 100-250 Hz for rhythm and R-peak detection, but higher rates (250-500 Hz) improve HRV precision and QRS detail. Because the QRS complex contains energy up to 40-100 Hz, sampling below 100 Hz risks aliasing and lost detail.

Analysis and Device Questions

Q4: Can PPG detect arrhythmia the same way ECG does?

A4: Put simply, no. Specifically, ECG captures the heart's electrical activity and can identify specific arrhythmia types from waveform morphology (P waves, QRS width, ST segments). Meanwhile, PPG captures blood volume changes and can only infer irregularity from pulse interval variability. Thus, PPG can flag irregular pulse trends that warrant further evaluation, but it cannot diagnose atrial fibrillation, ventricular tachycardia, or other specific arrhythmia types. Only ECG reviewed by a clinician can do that.

Q5: What is the difference between time-domain and frequency-domain HRV?

A5: Generally, time-domain HRV features (SDNN, RMSSD) are calculated directly from R-R interval statistics and reflect overall or short-term variability. In comparison, frequency-domain HRV features (LF, HF) apply spectral analysis to decompose variability into frequency bands, with HF reflecting parasympathetic activity and LF reflecting mixed sympathetic/parasympathetic balance. Overall, both are useful — time-domain is simpler and more common in wearables, while frequency-domain offers deeper autonomic insight.

Q6: Can Vositone build a wearable with ECG for my brand?

A6: Most importantly, yes. Specifically, Vositone offers ECG module integration for OEM clients, including single-lead ECG with dry or Ag/AgCl electrodes, 100-500 Hz sampling, and custom firmware. Additionally, we support clinical validation through 3 hospital partnerships and can assist with regulatory documentation (FDA, CE MDR). Finally, MOQ starts at 300 units, with 10-15 day sampling and 30-55 day production. Contact our OEM team for a tailored ECG wearable proposal. For broader heart health capabilities, see our Cardiac Health Management Handbook 2026.

Build Your Heart Health Wearable

Finally, ECG signal analysis is a deep technical field that bridges electrical engineering, signal processing, and clinical medicine. Specifically, Vositone provides the full stack: PPG heart rate sensing (standard), ECG module integration (OEM option), arrhythmia trend algorithms, HRV analysis, white-label app, clinical validation support, and ISO 13485 capable manufacturing.

To build your heart health wearable: share your target use case (wellness vs. medical), required sensors (PPG only, PPG+ECG, SpO2), sampling and accuracy requirements, regulatory targets, and expected volume. Then, our heart health OEM team will respond within two working days with a tailored proposal. For data privacy in health wearables, see our Health Data Privacy Protection Guide.

Decode the signal, build with Vositone.

Useful links:

1. Wearable Technologies Market Trends
2. FDA Digital Health Regulations
3. American Medical Association Wearable Guidelines
4. WHO Digital Health Innovations Report
5. Vositone ECG EKG PPG Smart Watch