Symmetric Projection Attractor Reconstruction: Sex Differences in the ECG

Area of Science:

• Cardiology
• Biomedical Engineering
• Signal Processing

Background:

• Electrocardiogram (ECG) analysis is crucial for patient management.
• Traditional methods often discard valuable time-series data by focusing on fiducial points.
• A novel Symmetric Projection Attractor Reconstruction (SPAR) method utilizes complete waveform data for analysis.

Purpose of the Study:

• To apply the SPAR method to ECG signals to assess its clinical value.
• To investigate if detailed ECG morphology analysis adds value beyond traditional methods.
• To demonstrate SPAR's accuracy in discriminating between biologically distinct groups, specifically sex differences.

Main Methods:

• Applied the SPAR method to 9,007 10-second, 12-lead ECG recordings from Physionet.
• Investigated sex differences in normal sinus rhythm ECGs using SPAR visualization and quantification.
• Utilized a stacked machine learning model for sex classification based on SPAR-generated features.

Main Results:

• SPAR revealed clear visual differences between female and male ECGs.
• Machine learning model achieved 86.3% cross-validation and 91.3% test accuracy for sex classification.
• Classification accuracy was highest in young adults and declined with age; mid-precordial leads were individually best, but all 12 leads yielded highest overall accuracy.

Conclusions:

• SPAR quantifies ECG morphology without fiducial points, reducing bias by using all data.
• SPAR's visualization accurately discriminated ECG sex differences and amplified subtle waveform variations.
• This method can enhance automated ECG analysis, improving patient stratification and risk management.