Design of neuro-fuzzy models by evolutionary and gradient-based algorithms

All systems found in nature exhibit, with different degrees, a nonlinear behavior. To emulate this behavior, classical systems identification techniques use, typically, linear models, for mathematical simplicity. Models inspired by biological principles (artificial neural networks) and linguistically motivated (fuzzy systems), due to their universal approximation property, are becoming alternatives to classical mathematical models. In systems identification, the design of this type of models is an iterative process, requiring, among other steps, the need to identify the model structure, as well as the estimation of the model parameters. This thesis addresses the applicability of gradient-basis algorithms for the parameter estimation phase, and the use of evolutionary algorithms for model structure selection, for the design of neuro-fuzzy systems, i.e., models that offer the transparency property found in fuzzy systems, but use, for their design, algorithms introduced in the context of neural networks. A new methodology, based on the minimization of the integral of the error, and exploiting the parameter separability property typically found in neuro-fuzzy systems, is proposed for parameter estimation. A recent evolutionary technique (bacterial algorithms), based on the natural phenomenon of microbial evolution, is combined with genetic programming, and the resulting algorithm, bacterial programming, advocated for structure determination. Different versions of this evolutionary technique are combined with gradient-based algorithms, solving problems found in fuzzy and neuro-fuzzy design, namely incorporation of a-priori knowledge, gradient algorithms initialization and model complexity reduction.

The activity of mouse Cerberus like 2 during cardiogenesis - genetic and morphogenetic studies

Cardiogenesis is a delicate and complex process that requires the coordination of an intricate network of pathways and the different cell types. Therefore, understanding heart development at the morphogenetic level is an essential requirement to uncover the causes of congenital heart disease and to provide insight for disease therapies. Mouse Cerberus like 2 (Cerl2) has been defined as a Nodal antagonist in the node with an important role in the Left-Right (L/R) axis establishment, at the early embryonic development. As expected, Cerl2 knockout mice (Cerl2-/-) showed multiple laterality defects with associated cardiac failure. In order to identify the endogenous role of Cerl2 during heart formation independent of its described functions in the node, we accurately analyzed animals where laterality defects were not present. We thereby unravel the consequences of Cerl2 lossof- function in the heart, namely increased left ventricular thickness due to hyperplasia of cardiomyocytes and de-regulated expression of cardiac genes. Furthermore, the Cerl2 mutant neonates present impaired cardiac function. Once that the cardiac expression of Cerl2 is mostly observed in the left ventricle until around midgestration, this result suggest a specific regulatory role of Cerl2 during the formation of the left ventricular myoarchitecture. Here, we present two possible molecular mechanisms underlying the cardiac Cerl2 function, the regulation of Cerl2 antagonist in activation of the TGFßs/Nodal/Activin/Smad2 signaling identified by increased Smad2 phosphorilation in Cerl2-/- hearts and the negative feedback between Cerl2 and Wnt/ß-catenin signaling in heart formation. In this work and since embryonic stem cells derived from 129 mice strain is extensively used to produce targeted mutants, we also present echocardiographic reference values to progressive use of juveniles and young adult 129/Sv strain in cardiac studies. In addition, we investigate the cardiac physiology of the surviving Cerl2 mutants in 129/Sv background over time through a follow-up study using echocardiographic analysis. Our results revealed that Cerl2-/- mice are able to improve and maintain the diastolic and most of systolic cardiac physiologic parameters as analyzed until young adult age. Since Cerl2 is no longer expressed in the postnatal heart, we suggest that an intrinsic and compensatory mechanism of adaptation may be active for recovering the decreased cardiac function found in Cerl2 mutant neonates. Altogether, these data highlight the role of Cerl2 during embryonic heart development in mice. Furthermore, we also suggest that Cerl2-/- may be an interesting model to uncover the molecular, cellular and physiological mechanisms behind the improvement of the cardiac function, contributing to the development of therapeutic approaches to treat heart failures.