Bispectrum and ECG Changes and Correlation in Ischemic and Dilated Cardiomyopathy

Abstract

Background: Ischemic cardiomyopathy (ICM) and dilated cardiomyopathy (DCM) are major causes of heart failure, with overlapping clinical features that complicate differentiation. Electrocardiography (ECG) is a cornerstone diagnostic tool, but its linear parameters may not fully capture the complex electrophysiological dynamics of cardiomyopathies. Bispectrum analysis, a higher-order spectral technique, quantifies non-linear interactions and phase coupling in ECG signals, potentially enhancing diagnostic precision. Methods: A prospective case-control study conducted at Malla Reddy Narayana Multispecialty Hospital, Hyderabad, from November 2024 to April 2025, involving 90 patients (30 ICM, 30 DCM, 30 healthy controls). ECG signals acquired at 500 Hz and processed using MATLAB for bispectral magnitude and bicoherence. Echocardiography assessed left ventricular ejection fraction (LVEF) and end-diastolic diameter (LVEDD), while MRI quantified myocardial scar in ICM. Statistical analyses included ANOVA, Pearson/Spearman correlations, and ROC analysis (AUC) using SPSS version 26 (p < 0.05). Post-hoc comparisons were performed where appropriate, and diagnostic performance was assessed using AUC, sensitivity, and specificity. Results: ICM patients showed higher high-frequency bispectral magnitude (177.6 ± 16.2 AU) and Q-waves/ST-segment changes, reflecting myocardial scarring and ischemia. DCM patients exhibited higher low-frequency bispectral magnitude (181.3 ± 15.7 AU) and greater LVEDD (65.8 ± 5.2 mm), indicating ventricular dilation. Both groups had reduced LVEF (ICM: 36.2 ± 7.1%; DCM: 34.8 ± 7.5%) compared to controls (60.1 ± 2.0%). Arrhythmias were prevalent in ICM (10/30) and DCM (12/30). Combined bispectrum-ECG analysis achieved the highest diagnostic accuracy (AUC = 0.88, sensitivity 86%, specificity 83%) compared to ECG alone (AUC = 0.76) or bispectrum alone (AUC = 0.81). Significant correlations were found between low-frequency bispectral power and LVEF (r = -0.68, p < 0.01) in DCM, and high-frequency bispectral power and myocardial scar (r = 0.64, p < 0.01) in ICM. Conclusion: Bispectrum analysis enhances ECG’s ability to differentiate ICM and DCM by capturing non-linear electrophysiological dynamics. The combined bispectrum-ECG approach offers superior diagnostic accuracy, potentially reducing reliance on invasive methods.

Keywords: Bispectrum analysis, electrocardiography, ischemic cardiomyopathy, dilated cardiomyopathy, non-linear dynamics, diagnostic accuracy