An ANN-based HRV classifier for cardiac health prognosis

Area of Science:

• Cardiology
• Artificial Intelligence in Medicine
• Biomedical Signal Processing

Background:

• Cardiac health assessment traditionally relies on cardiologist interpretation of patient data.
• Heart Rate Variability (HRV) analysis offers a non-invasive method for evaluating cardiac status.
• Existing methods may lack objective, automated classification of cardiac health severity.

Purpose of the Study:

• To develop and validate an artificial neural network (ANN)-based classifier for cardiac health.
• To assess the performance of the ANN classifier using Heart Rate Variability (HRV) data.
• To compare the ANN classifier's output with expert cardiologist diagnoses.

Main Methods:

• Electrocardiogram (ECG) data were collected from 46 patients suspected of cardiac abnormalities.
• Heart Rate Variability (HRV) was calculated using a self-designed autoregressive-model-based technique.
• A multi-layer artificial neural network (ANN) was trained on 28 patients' HRV data and tested on 18 patients' data.

Main Results:

• The proposed ANN-based HRV classifier successfully categorized patients' cardiac health.
• The classification results from the ANN model demonstrated a high degree of agreement with cardiologist opinions.
• The system provides cardiac health status independently of direct cardiologist input.

Conclusions:

• An ANN-based HRV classifier can accurately and objectively assess cardiac health status.
• This automated approach shows potential for supporting clinical decision-making in cardiology.
• The findings suggest the utility of machine learning in analyzing ECG and HRV for cardiac diagnostics.