Caracterização das Hipsarritmias secundárias à Síndrome Congênita do Zika vírus e à Síndrome de West baseada em Momentos Conjuntos e Medidas de Entropia

Abstract

West Syndrome is a rare and severe form of childhood epilepsy characterized by triad: presence of spasms, cognitive developmental delay and the hipsarrhythmia pattern on electroencephalogram (EEG). Hipsarrhythmia is a specific chaotic morphology present in the interictal period of the EEG signal, studied and known by neurologists since 1841, through the description of West Syndrome (WS) and which has recently also been identified in the examinations of patients with Zika virus congenital syndrome (ZVCS). The hipsarrhythmia characterization in infants with microcephaly due Zika virus is still very superficial. Then, the question arises whether there is a difference between the hysarrhythmic pattern that occurs in those born with CZVS of those from WS. Since the emergence of ZVCS cases, many questions about the characterization of this disease are still open, among them, whether the hypsarrhythmia in ZVCS follows the same electroencephalographic pattern as the hypsarrhythmia in WS. In view of this, this work proposes the development of a computational methodology for analysis and differentiation, based on the time-frequency domain, between the chaotic hipsarrhythmia pattern found in EEG signals of patients with microcephaly caused by Zika virus and also found in patients diagnosed with West Syndrome. Analysis in the time-frequency domain is performed from the Wavelet Continuous Transform (CWT) which reveals the energy distribution of the EEG signal at different frequency scales over time. Three mother-wavelet functions are tested to determine the most appropriate function to represent EEG signals with hipsarrhythmia ZVCS and hipsarrhythmia WS. Considering the energy distribution profiles generated by CWT, four joint moments are obtained - joint mean - μ(t,f)

, joint variance -

σ 2 (t,f) , join skewness - λ(t,f)

, and join kurtosis - κ(t,f) - and four entropy measures - Shannon, Log Energy, Norm, and Sure - to compose the attributes vector that representing the hypsarrhythmic signals under analysis. The performance of five classical types of machine learning algorithms are verified in classification using the k-fold cross validation and leave-one-patient-out cross validation methods. Discrimination results provided 78,08% accuracy, 85,55% sensitivity, 73,21% specificity, and AUC = 0,89 for the ANN classifier.