Using Lempel-Ziv Complexity to Assess ECG Signal Quality

Area of Science:

• Biomedical Engineering
• Signal Processing
• Medical Diagnostics

Background:

• Poor quality wireless electrocardiography (ECG) recordings can lead to misdiagnosis and inefficient use of medical resources.
• Objective assessment of ECG signal quality is crucial for reliable clinical interpretation.

Purpose of the Study:

• To interpret Lempel-Ziv (LZ) complexity as a metric for ECG signal quality assessment.
• To evaluate the performance of LZ complexity in identifying various noise types in ECG signals.

Main Methods:

• LZ complexity was analyzed for clean ECG signals and signals contaminated with simulated high-frequency (HF) noise, low-frequency (LF) noise, power-line (PL) noise, and impulse (IM) noise.
• The impact of noise type, signal length, and signal-to-noise ratio (SNR) on LZ complexity was investigated.
• LZ complexity was further tested on real ECG signals with muscle artefacts (MAs), baseline wander (BW), and electrode motion (EM) artefacts.

Main Results:

• LZ complexity clearly differentiated HF noise from LF and PL noise, indicating its potential for HF noise detection.
• ECG signals with added HF noise exhibited the highest LZ values, while other noise types resulted in lower LZ values.
• LZ complexity generally increased with decreasing SNR for most noise types, and showed sensitivity to MA artefacts in real ECG signals.

Conclusions:

• LZ complexity is a sensitive indicator of noise level in ECG signals, especially for HF noise.
• LZ complexity serves as a valuable reference index for assessing the overall quality of wireless ECG recordings.
• This approach can contribute to more accurate ECG diagnoses and optimized resource utilization.