The main objective of this thesis was to study the variability of the cardiac, respiratory and vascular systems through electrocardiographic (ECG), respiratory flow (FLW) and blood pressure (BP) signals, in patients with idiopathic (IDC), dilated (DCM), or ischemic (ICM) disease. The aim of this work was to introduce new indices that could contribute to characterizing these diseases. With these new indices, we propose methods to classify cardiomyopathy patients (CMP) according to their cardiovascular risk or etiology. In addition, a new tool was proposed to reconstruct artifacts in biomedical signals. From the ECG, BP and FLW signals, different data series were extracted: beat to beat intervals (BBI - ECG), systolic and diastolic blood pressure (SBP and DBP - BP), and breathing duration (TT - FLW). -Firstly, we propose a novel artifact reconstruction method applied to biomedical signals. The reconstruction process makes use of information from neighboring events while maintaining the dynamics of the original signal. The method is based on detecting the cycles and artifacts, identifying the number of cycles to reconstruct, and predicting the cycles used to replace the artifact segments. The reconstruction results showed that most of the artifacts were correctly detected, and physiological cycles were incorrectly detected as artifacts in fewer than 1% of the cases. The second part is related to the cardiac death risk stratification of patients based on their left ventricular ejection (LVEF), using the Poincaré plot analysis, and classified as low (LVEF > 35%) or high (LVEF = 35%) risk. The BBI, SBP, and IT series of 46 CMP patients were applied. The linear discriminant analysis and support vector machines (SVM) classification methods were used. When comparing low risk vs high risk, an accuracy of 98 12% was obtained. Our results suggest that a dysfunction in the vagal activity could prevent the body from correctly maintaining circulatory homeostasis Next, we studied cardio-vascular couplings based on heart rate (HRV) and blood pressure (BPV) variability analyses in order to introduce new indices for noninvasive risk stratification in IDC patients. The ECG and BP signals of 91 IDC patients, and 49 healthy subjects were used. The patients were stratified by their sudden cardiac death risk as: high risk (IDCHR), when after two years the subject either died or suffered complications, or low risk (IDCLR) otherwise. Several indices were extracted from the BBI and SBP, and analyzed using the segmented Poincaré plot analysis, the high-resolution joint symbolic dynamics, and the normalized short time partial directed coherence methods. SVM models were built to classify these patients based on their sudden cardiac death risk. The SVM IDCLR vs IDCHR model achieved 98 9% accuracy with an area under the curve (AUC) of 0.96. Our results suggest that IDCHR patients have decreased HRV and increased BPV compared to both the IDCLR patients and the control subjects, suggesting a decrease in their vagal activity and the compensation of sympathetic activity. Lastly, we analyzed the cardiorespiratory interaction associated with the systems related to ICM and DCM disease. We propose an analysis based on vascular activity as the input and output of the baroreflex response. The aim was to analyze the suitability of cardiorespiratory and vascular interactions for the classification of ICM and DCM patients. We studied 41 CMP patients and 39 healthy subjects. Three new sub-spaces were defined: 'up' for increasing values, 'down' for decreasing values, and 'no change' otherwise, and a three-dimensional representation was created for each sub-space that was characterized statistically and morphologically. The resulting indices were used to classify the patients by their etiology through SVM models achieving 92.7% accuracy for ICM vs DCM patients comparison. The results reflected a more pronounced deterioration of the autonomous regulation in DCM patients.
© 2001-2024 Fundación Dialnet · Todos los derechos reservados