No description available.

Made available in DSpace on 2016-06-02T19:44:01Z (GMT). No. of bitstreams: 1 TeseCGP.pdf: 6995198 bytes, checksum: 2362893ce6d74661c57a00d4f64e85f1 (MD5) Previous issue date: 2007-02-27 / Universidade Federal de Sao Carlos / In the last few decades, it seems to be increasing the number of families who find difficulties to educate its children. The most frequent parents concerns are: the difficulty of establishing limits to the behavior of the child or adolescent; how to develop behaviors that they believe to be right; and how to proceed when behaviors problems appears. These difficulties do not restrict to the universe of the families who take care of its own children, but they are also focus of concern on institutions that foster care children victims of domestic violence and those caregivers. The lack of preparation of parents, educators or caregivers of institutions, many times may develop or aggravate problems of development and behavior of these children. For this reason, this study had the aim to develop, implement and evaluate a program of intervention with caregivers that work with sheltered children who had been victims of domestic violence, in order to present a reference on Educative practices, preventing or reducing possible behavior problems of these children. To evaluate the effectiveness of the intervention program, evaluations had been applied before the beginning (pre-test) of the intervention program and soon after its ending (postest), using the following instruments: IEP - Inventory of Educative Practices, CBCL - Child Behavior Checklist and the Script of interview of life and professional background of the caregivers. Also were collected continuous measures of the Interaction between the caregivers and the children in 19 sessions, defined as 3 of base line, 12 taken three days after each Intervention session and 4 of follow-up, applied five months after the ending of the intervention. The results demonstrated a significant reduction of the negative educative practices, mainly in Relaxed Disciplines and Negative Monitoring, with high scores at the pre-test, as well an increase of the positive educative practices. The evaluation using the CBCL demonstrated that the children had presented statistical significant behavior improvement. The continuous measures data had indicated an increase of positive interaction behaviors between caregivers and children, as well a reduction of negative interaction behaviors, and the most significant outcome was that the measured frequencies of good behaviors in follow-up sessions did not return to the measured levels of the base line. The considered intervention program presented can be considered efficient to generate changes in the interaction standards between caregivers and children; however it would need to be applied in different institutions to assure the applicability in distinct environments. / Nas últimas décadas parece vir aumentando o número de famílias que encontram dificuldades para educar seus filhos. Dentre as reclamações mais freqüentes, advindas dos pais, estão: a dificuldade em estabelecer limites ao comportamento da criança ou do adolescente, como desenvolver comportamentos que acreditam ser adequados e como proceder quando problemas comportamentais aparecem. Estas dificuldades não se restringem ao universo das famílias que cuidam de seus filhos, mas também são focos de preocupação das instituições que abrigam crianças vítimas de violência doméstica e das monitoras destas. A falta de preparo de pais, educadores ou monitoras de instituições que cuidam de crianças, muitas vezes faz surgir ou agravar problemas de desenvolvimento e de conduta destas. Por esta razão, este estudo teve como objetivo desenvolver, implementar e avaliar um programa de intervenção junto a monitoras que trabalham com crianças abrigadas, que foram vítimas de violência doméstica, visando apresentar um referencial de práticas educativas, que possa prevenir ou amenizar possíveis problemas de conduta destas crianças. Participaram deste estudo 14 monitoras e 13 crianças de um abrigo localizado em uma cidade litorânea do Estado de São Paulo. Para avaliar a eficácia do programa de intervenção foram feitas avaliações antes do início do programa de intervenção e logo após seu término, utilizando os seguintes instrumentos: IEP Inventário de Práticas Educativas, CBCL Child Behavior Checklist, Roteiro de entrevista sobre história de vida e profissional das monitoras. Além das medidas de pré e pós-teste foram também coletadas medidas contínuas de observação do comportamento de interação da díade monitoracriança, em 19 sessões, sendo três de linha de base, 12 três dias após cada sessão de intervenção e quatro de follow-up, cinco meses após o término da intervenção. Os resultados demonstraram uma redução acentuada das práticas educativas negativas, principalmente Disciplina Relaxada e Monitoria Negativa, com altos escores no pré-teste, bem como um aumento dos escores das práticas educativas positivas. A avaliação obtida com CBCL demonstrou que as crianças apresentaram melhoras comportamentais estatisticamente significativas. Os dados de medida contínua indicaram um aumento dos comportamentos de interação positiva, tanto das monitoras como das crianças, bem como uma diminuição dos comportamentos de interação negativa, sendo que o resultado mais significativo foi que as freqüências de comportamento medidas em follow-up não retornaram aos níveis medidos em linha de base. O programa de intervenção sobre práticas educativas proposto neste trabalho mostrou-se eficaz para gerar mudanças nos padrões de interação entre cuidadores e crianças, entretanto necessitaria de replicações em instituições diferentes para garantir a aplicabilidade em ambientes diversos.

La démence à corps de Lewy (DCL) et la maladie de Parkinson (MP) sont des maladies neurodégénératives touchant des milliers de Canadiens et leur prévalence croît avec l’âge. La MP et la DCL partagent la même pathophysiologie, mais se distinguent par l’ordre de manifestation des symptômes : la DCL se caractérise d’abord par l’apparition d’un trouble neurocognitif majeur (démence), tandis que la MP se manifeste initialement par un parkinsonisme. De plus, jusqu’à 80% des patients avec la MP développeront une démence (MPD). Il est désormais établi que le trouble comportemental en sommeil paradoxal idiopathique (TCSPi) constitue un puissant prédicteur de la DCL et la MP. En effet, cette parasomnie, marquée par des comportements indésirables durant le sommeil, est considérée comme un stade prodromal des synucléinopathies, telles que la MP, la DCL et l'atrophie multisystémique (AMS). Ainsi, la majorité des patients atteints d’un TCSPi développeront une synucléinopathie. Malgré les avancées scientifiques, les causes du TCSPi, de la MP et de la DCL demeurent inconnues et aucun traitement ne parvient à freiner ou à arrêter la neurodégénérescence. De plus, ces pathologies présentent une grande hétérogénéité dans l’apparition et la progression des divers symptômes. Face à ces défis, la recherche vise à mieux cerner les phases précoces/initiales et les trajectoires évolutives de ces maladies neurodégénératives afin d’intervenir le plus précocement possible dans leur développement. C’est pourquoi le TCSPi suscite un intérêt majeur en tant que fenêtre d'opportunités pour tester l’efficacité des thérapies neuroprotectrices contre les synucléinopathies, permettant d'agir avant que la perte neuronale ne devienne irréversible. Le TCSPi offre ainsi une occasion unique d'améliorer la détection de la démence et le suivi des individus à haut risque de déclin cognitif. D'où l'importance cruciale de pouvoir généraliser les résultats issus de la recherche sur de petites cohortes à l'ensemble de la population. Sur le plan de la cognition, les études longitudinales sur le TCSPi ont montré que les atteintes des fonctions exécutives, de la mémoire verbale et de l'attention sont les plus discriminantes pour différencier les individus qui développeront une démence de ceux qui resteront idiopathiques. De plus, un grand nombre de patients TCSPi souffrent d’un trouble neurocognitif mineur ou trouble cognitif léger (TCL), généralement considéré comme un stade précurseur de la démence. Les recherches actuelles sur les données cognitives chez cette population offrent des perspectives prometteuses, mais reposent sur des approches statistiques classiques qui limitent leur validation et généralisation. Bien qu'elles offrent une précision élevée (80 à 85%) pour détecter les patients à risque de déclin cognitif, une amélioration est nécessaire pour étendre l'utilisation de ces marqueurs à une plus large échelle. Depuis les années 2000, l'accroissement de la puissance de calcul et l'accès à davantage de ressources de mémoire ont suscité un intérêt accru pour les algorithmes d'apprentissage machine (AM). Ces derniers visent à généraliser les résultats à une population plus vaste en entraînant des modèles sur une partie des données et en les testant sur une autre, validant ainsi leur application clinique. Jusqu'à présent, aucune étude n'a évalué les apports de l'AM pour la prédiction de l'évolution des synucléinopathies en se penchant sur le potentiel de généralisation, et donc d'application clinique, à travers l'usage d'outils non invasifs et accessibles ainsi que de techniques de validation de modèles (model validation). De plus, aucune étude n'a exploré l'utilisation de l'AM associée à des méthodes de généralisation sur des données neuropsychologiques longitudinales pour élaborer un modèle prédictif de la progression des déficits cognitifs dans le TCSPi. L’objectif général de cette thèse est d’étudier l’apport de l’AM pour analyser l’évolution du profil cognitif de patients atteints d’un TCSPi. Le premier chapitre de cette thèse présente le cadre théorique qui a guidé l’élaboration des objectifs et hypothèses de recherche. Le deuxième chapitre est à deux volets (articles). Le premier vise à fournir une vue d'ensemble de la littérature des études ayant utilisé l'AM (avec des méthodes de généralisation) pour prédire l'évolution des synucléinopathies vers une démence, ainsi que les lacunes à combler. Le deuxième volet vise à explorer et utiliser pour la première fois l'AM sur des données cliniques et cognitifs pour prédire la progression vers la démence dans le TCSPi, dans un devis longitudinal. Enfin, le dernier chapitre de la thèse présente une discussion et une conclusion générale, comprenant un résumé des deux articles, ainsi que les implications théoriques, les forces, les limites et les orientations futures. / Lewy body dementia (LBD) and Parkinson's disease (PD) are neurodegenerative diseases affecting thousands of Canadians, and their prevalence increases with age. PD and DLB share the same pathophysiology, but differ in the order of symptom manifestation: DLB is characterized first by the onset of a major neurocognitive disorder (dementia), whereas PD initially manifests as parkinsonism. Moreover, up to 80% of PD patients will go on to develop dementia (PDD). It is established that idiopathic REM sleep behavior disorder (iRBD) is a powerful predictor of DLB and PD. Indeed, this parasomnia, marked by undesirable behaviors during sleep, is considered a prodromal stage of synucleinopathies, such as PD, DLB and multisystem atrophy (MSA). Therefore, the majority of patients with iRBD will develop synucleinopathy. Despite scientific advancements, the causes of iRBD, PD, and DLB remain unknown and no treatment has been able to slow or halt neurodegeneration. Furthermore, these pathologies display great heterogeneity in the onset and progression of various symptoms. Faced with these challenges, research aims to better understand the early/initial stages and the progressive trajectories of these neurodegenerative diseases in order to intervene as early as possible in their development. This is why iRBD garners major interest as a window of opportunities to test the effectiveness of neuroprotective therapies against synucleinopathies, enabling action to be taken before neuronal loss becomes irreversible. iRBD thus provides a unique opportunity to improve dementia detection and monitoring of individuals at high risk of cognitive decline. Hence the crucial importance of being able to generalize results of research on small cohorts to the entire population. In terms of cognition, longitudinal studies on iRBD have shown that impairments in executive functions, verbal memory, and attention are the most discriminating in differencing between individuals who will develop dementia from those who will remain idiopathic. In addition, many iRBD patients suffer from a mild neurocognitive disorder or mild cognitive impairment (MCI), generally considered as a precursor stage of dementia. Current research on cognitive data in this population offers promising prospects, but relies on traditional statistical approaches that limit their validation and generalizability. While they provide high accuracy (80 to 85%) for detecting patients at risk of cognitive decline, improvement is needed to extend the use of these markers to a larger scale. Since the 2000s, increased computational power and access to more memory resources have sparked growing interest in machine learning (ML) algorithms. These aim to generalize results to a broader population by training models on a subset of data and testing them on another, thus validating their clinical application. To date, no study has assessed the contributions of ML for predicting the progression of synucleinopathies, focusing on the potential for generalization, and hence clinical application, through the use of non-invasive, accessible tools and model validation techniques. Moreover, no study has explored the use of ML in conjunction with generalization methods on longitudinal neuropsychological data to develop a predictive model of cognitive deficit progression in iRBD. The general objective of this thesis is to study the contribution of ML in analyzing the evolution of the cognitive profile of patients with iRBD. The first chapter of this thesis presents the theoretical framework that guided the formulation of the research objectives and hypotheses. The second chapter is in two parts (articles). The first aims to provide an overview of the literature of studies that have used ML (with generalization methods) to predict the progression of synucleinopathies to dementia, as well as the gaps that need to be filled. The second part aims to explore and use for the first time ML on clinical and cognitive data to predict progression to dementia in iRBD, in a longitudinal design. Finally, the last chapter of the thesis presents a discussion and a general conclusion, including a summary of the two articles, as well as theoretical implications, strengths, limitations, and future directions.

The basal ganglia (BG) are a group of brain nuclei situated deep in the cerebral hemispheres. While BG were primarily associated with motor functions, in recent years there has been an increasing evidence that BG are also significantly involved in a wide range of non-motor functions. This work focused on some of the non-motor symptoms associated with two typical basal ganglia disorders: Parkinson's disease (PD) and Huntington's disease (HD). The first study concerned spatial navigation impairment in patients with HD. Their spatial navigation skills were tested using the Blue Velvet Arena, technique evaluating spatial navigation in real space, capable to selectively differentiate between two components of spatial navigation - allocentric (environment-oriented) and egocentric (self-oriented). Allocentric navigation is linked to hippocampal function, whereas egocentric navigation is usually associated with striatum, a structure predominantly affected in HD. We found that spatial navigation is not significantly affected in the early stages of HD and that in more advanced stages, when spatial navigation is already impaired, there is no significant difference between allocentric and egocentric navigation impairment. We speculate that the striatal involvement does not contribute to the impairment of the...

Le trouble comportemental en sommeil paradoxal idiopathique (TCSPi) précède de plusieurs années le diagnostic d’une maladie synucléinopathique. Dans cette étude, nous cherchions à déterminer si la puissance spectrale relative, les composantes rythmiques et arythmiques des spectres de puissance, ainsi que la connectivité fonctionnelle permettaient d’identifier à un temps de base les patients ayant un TCSPi qui développerait une synucléinopathie lors des suivis cliniques annuels. Un enregistrement EEG au repos et une évaluation neuropsychologique ont été conduits auprès de quatre-vingt-un participants atteints d’un TCSPi (66.89 ± 6.91 ans, 20 femmes) et des évaluations neurologiques annuelles étaient menées afin de définir si les patients montraient des symptômes d’une maladie synucléinopathique. La puissance spectrale standard ainsi qu’une estimation spectrale des composantes rythmiques et arythmiques ont été calculées. Ensuite, la connectivité fonctionnelle globale et entre chaque paire d’électrodes ont été estimée par le weighted Phase Lag Index. Après une durée de suivi de 5.01 ± 2.76 ans, 34 participants ont été diagnostiqués avec une synucléinopathie et 47 sont restés exempts de maladie. Comparativement aux participants avec un TCSPi n’ayant pas converti, ceux ayant converti montraient, lors de l’évaluation de base, une puissance spectrale relative plus élevée dans la bande thêta, une pente de la composante arythmique plus abrupte ainsi qu'une puissance rythmique plus élevée en thêta dans les régions occipitales et temporales ainsi qu’en en bêta1 dans les régions frontales. De plus, les patients TCSPi ayant converti présentaient une hyperconnectivité globale dans la bande bêta, mais une hypoconnectivité dans la bande alpha entre les régions temporo-occipitales gauches lors de l’évaluation de base comparativement à ceux n’ayant pas converti. Les altérations mesurables en EEG au repos lors de l’évaluation de base chez les participants avec TCSPi ayant converti vers une maladie synucléinopathique suggèrent une perturbation des réseaux à grande échelle affectés par la neurodégénérescence précoce des structures sous-corticales. / Idiopathic REM sleep behavior disorder (iRBD) precedes the diagnosis of synucleinopathies by several years. In this study, we aimed to determine whether relative spectral power, rhythmic and arrhythmic components of power spectra, and functional connectivity at baseline could identify patients with iRBD who will develop a synucleinopathy at follow-up. Resting-state EEG recordings and neuropsychological evaluations were conducted on eighty-one participants with iRBD (66.89 ± 6.91 years; 20 women), and annual neurological assessments were performed to define the emergence of synucleinopathy symptoms. Standard spectral power and spectral estimates of rhythmic and arrhythmic components were calculated. Additionally, global and pairwise functional connectivity were estimated using the weighted Phase Lag Index. After a follow-up period of 5.01 ± 2.76 years, 34 participants were diagnosed with a synucleinopathic disorder, while 47 remained disease-free. Compared to patients who did not convert, patients who converted at follow-up exhibited higher relative spectral power in the theta band, steeper slopes of the arrhythmic component, and increased rhythmic power in theta in posterior regions and beta1 in frontal regions at baseline evaluation. Furthermore, participants who converted showed hyperconnectivity in the beta band and hypoconnectivity in the alpha band between left temporo-occipital regions at baseline compared to participants who did not convert. The measurable alterations in resting-state EEG at baseline in participants with iRBD who phenoconverted towards a synucleinopathy suggest disruption of large-scale networks affected by early neurodegeneration of subcortical structures.