Analysis of the optimal time delay for Reconstruction of Phase Space of VCG signals to identify anterior wall infarction
Rafael Duarte de Sousa, José Barbosa, Ana Nogueira Rodrigues De Oliveira, Carlos Danilo Miranda Regis

DOI: 10.14209/sbrt.2021.1570724165
Evento: XXXIX Simpósio Brasileiro de Telecomunicações e Processamento de Sinais (SBrT2021)
Keywords: Reconstructed Phase Space Anterior Myocardial Infarction Deep Neural Network Vectorcardiogram
Acute Myocardial Infarction (MI) can be detected through Reconstructed Phase Space (RPS) parameters. The RPS is dependent on a time delay (τ) that can be a fixed or an optimal value based on the mutual information between a vectorcardiogram lead and its delayed version. This paper aims to assess the effect of an optimized time-delay in the detection of MI in the anterior wall using RPS parameters. For optimal τ values the sensitivity, specificity, accuracy and AUC obtained were: 64.7%, 63.2%, 57.8% and 62.3%, respectively; while for a fixed τ value were, respectively, 81.3%, 74.5%, 64.9% and 70.1%.