Use of feature vectors and constellation maps for analysis of normal sinus rhythm, supraventricular arrhythmia and sudden cardiac death electrocardiogram (ECG) traces

• Main Authors: Santos, Marc Ericson C., Libatique, Nathaniel Joseph C., Ramos, Catherine Manuela L., Tangonan, Gregory L.
• Format: text
• Published: Animo Repository 2011
• Subjects: Electrocardiography; Cardiovascular system—Diseases—Diagnosis—Equipment and supplies; Biomedical; Electrical and Computer Engineering
• Online Access: https://animorepository.dlsu.edu.ph/faculty_research/5751
• Institution: De La Salle University

The electrocardiogram (ECG) are analyzed in the frequency domain to reveal identifying signatures of different heart conditions. Two approaches to spectrum analysis are proposed. The first one constructed a feature vector for normal sinus rhythm (NSR), supraventricular arrhythmia (SVA) and sudden cardia death (SCD) using the QRS complex and the PR interval of ECG traces. These feature vectors were tested and they were successful in identifying SCD from NSR at a rate of 90%, and SVA from NSR at a rate of 80%. The second approach is the construction of constellation maps based on peaks in their spectrogram of the ECG traces. Constellation maps reveal that NSR have peaks varying from 3.25-3.75, 4.25-4.75 and 5.25-5.75 Hz whereas SVA signals have peaks occurring only at 4.25-4.75 Hz. Spectral properties expressed as either feature vector or constellation map are reliable basis for possible classification of heart conditions.