Orientadores: Li Li Min, Fernando Cendes / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Ciencias Medicas / Made available in DSpace on 2018-08-04T03:09:51Z (GMT). No. of bitstreams: 1 Bonilha_LeonardoFatorGouvea_D.pdf: 8420238 bytes, checksum: 016f0b30e07800517df57c5fc2672c09 (MD5) Previous issue date: 2004 / Resumo: A esclerose hipocampal (EH) é a alteração histológica mais comum em pacientes com epilepsia do lobo temporal medial (ELTM). A Ressonância Magnética (RM) de crânio possibilita a detecção in vivo de sinais associados à EH, permitindo que pacientes com EL TM reftatária à medicação sejam submetidos à ressecção cirúrgica do hipocampo para tratamento de crises epilépticas. As causas de reftatariedade à medicação e ao tratamento cirúrgico ainda são desconhecidas, porém supõe-se que um dos motivos seja a presença de lesão neuronal acometendo outras áreas cerebrais além do hipocampo. O uso da morfometria por RM permite avaliação do dano neuronal tanto no hipocampo como em outras estruturas cerebrais através da avaliação e quantificação da atrofia presente nestas estruturas. Para avaliação pormenorizada das estruturas cerebrais foi realizada a implementação e validação de um protocolo anatômico para mensuração da região mesial do lobo temporal, com uso de RM tridimensional de alta definição. Foi também definido um protocolo para volumetria automatizada baseada em voxel de todo o cérebro. Foi observado que o dano neuronal em pacientes com EL TM se estende além do hipocampo e acomete regiões que se conectam funcionalmente e anatomicamente ao hipocampo. Tál achado sugere que exista lesão abrangendo uma rede neuronal, o que pode ser responsável em conjunto pelas manifestações clínicas observadas nesses pacientes / Abstract: Hippocampal sclerosis (HS) is the most common histological finding in patients with media! temporal lobe epilepsy (MTLE). Magnetic resonance imaging (MRl) permits in vivo detection of signs that are associated to HS, permitting the surgical treatment for these patients. The causes of medical and surgical reftactoriness observed in patients with MTLE are still unknown. One possible explanation is the fact that the neuronalloss encountered in these patients spans over other brain areas beyond the hippocampus. The use of morphometric quantification of brain structures through MRI is a powerful tool to investigate the neuronalloss in the hippocampus and in other areas of the brain. In order to assess the neuronal damage in brain structures of patients with MTLE, we developed a protocol for manual MRI morphometry of the media! temporallobe structures. We also developed an automatic protocol to assess the concentration of gray matter in the whole brain of these patients through the use of Voxel Based Morphometry. We observed that patients with MTLE exhibit neuronal loss that is not restricted to the hippocampus, but affects di:fferent areas throughout the brain that are functionally and anatomica1ly connected to the hippocampus. These findings suggest that a lesion of a network of neural structures may be responsible for the clinical symptomatology exhibited by patients with MTLE / Doutorado / Neurologia / Doutor em Ciências Médicas

[EN] Automated temporal planning deals with the construction of plans for problems specified with durative actions of possibly different duration. The goal of temporal planning is to select and schedule the actions so as to achieve the problem goals, and the optimality criteria is the plan makespan. However, many real-word application define goals with time constraints which may not be satisfied with the plan of optimal makespan. The 2006 International Planning Competition introduced the PDDL3.0 language and organized the first and only track in planning with state trajectory constraints, including time restrictions, soft constraints and preferences. Particularly, four domains that feature deadline constraints were tested at the IPC and two planners participated in such competition although no one exhibited a good performance or fully-correction in the fulfillment of deadlines. An early detection of unsolvability in temporal planning problems, specifically those with strict deadlines, is crucial for avoiding an unfruitful exploration of the search space. And this is precisely the objective of this PhD dissertation. This work contributes with a temporal landmark-based model embedded into a temporal planner that allows for a rapid and eficient identification of unsolvable problems with deadline constraints. Our model incorporates the basic model operators of PDDL3.0 for defining temporal constraints and extends the STRIPS landmark concept to a temporal context. A temporal landmark is associated to three intervals that denote the time frame of the generation, validity and necessity of the landmark in the plan, respectively. The set of temporal landmarks of a problem along with their ordering relations and temporal restrictions define a temporal landmark graph which comprises the relationships that must exist between the literals of a solution plan and, therefore, between the corresponding actions. As long as new data is incorporated in the graph, the landmarks intervals are updated and propagated accordingly, thus reflecting a more accurate picture of a solution plan. The temporal landmark model is integrated in a domain-independent temporal planner named TempLM. During the search process, TempLM uses the temporal landmark graph to prune those partial plans of the search tree that are not compliant with the information of the graph. Additionally, we present a further improvement by which a feedback routine between the landmarks graph and the plan of a node tree is established. This process is used to refine the information of the graph and likewise narrow the search. In the experimental evaluation, we show the effectiveness of the proposed approach for detecting unsolvability and solving temporal planning problems with tight deadline constraints. / [ES] El objetivo de la planificación temporal automática es la construcción de planes con acciones de diferente duración que necesitan ser programadas adecuadamente con el fin de conseguir los objetivos del problema. En planificación temporal, la optimalidad se mide como la duración del plan más corto. Sin embargo, en muchos problemas del mundo real es necesario gestionar restricciones temporales asociadas a los objetivos del problema que pueden no satisfacerse con el plan de menor duración. En la Competición Internacional de Planificación del año 2006 se presentó el lenguaje PDDL3.0 y se realizó la primera y única competición de planificadores con gestión de restricciones de trayectorias de estado y preferencias. Concretamente, en esta IPC se probaron cuatro dominios con restricciones temporales donde los objetivos deben satisfacerse dentro de un límite de tiempo establecido o deadline. Dos planificadores participaron en esta competición aunque ninguno presentó un buen comportamiento respecto al cumplimiento de los deadlines. En este tipo de problemas, especialmente cuando se trata de problemas altamente restringidos, es crucial detectar la irresolubilidad de los mismos lo antes posible, y en este punto es donde se establece nuestro objetivo. Este trabajo de tesis presenta un modelo de landmarks temporales que permite identificar rápida y eficientemente la irresolubilidad de problemas de planificación con restricciones. Nuestro modelo incorpora las restricciones temporales del lenguaje PDDL3.0 y extiende el concepto de landmark STRIPS al contexto temporal. A cada landmark temporal se le asocian tres tipos de intervalos que se actualizan y propagan de acuerdo a las relaciones de orden y restricciones temporales entre ellos. Los landmarks temporales junto con sus relaciones de orden y restricciones temporales forman un grafo de landmarks temporales donde se sintetiza las relaciones que existen entre los literales de un plan solución y, consecuentemente, entre sus correspondientes acciones. Cuando se añade nueva información al grafo, se actualizan y propagan los intervalos de los landmarks, reflejando una imagen más precisa del plan solución. Posteriormente, el modelo de landmarks temporales se integra en un planificador temporal heurístico independiente del dominio denominado TempLM. TempLM utiliza toda la información extraída del problema para podar los planes parciales en el árbol de búsqueda que no son compatibles con la información del grafo de landmarks temporales. Además, se dispone de un proceso de retroalimentación entre el grafo de landmarks temporales y el proceso de búsqueda de un plan solución que permite enriquecer el grafo y, asimismo, acotar el espacio de búsqueda. Los resultados experimentales muestran que esta aproximación detecta rápidamente problemas irresolubles y también es muy efectiva para resolver problemas con restricciones muy ajustadas. / [CAT] L'objectiu de la planificació temporal automàtica és la construcció de plans amb accions de diferent durada que necessiten ser programades adequadament amb la finalitat d'aconseguir els objectius del problema. En planificació temporal, l'optimitat es mesura com la durada del pla més curt. No obstant açò, en molts problemas del món real és necessari gestionar restriccions temporals associades als objectius del problema que poden no satisfer-se amb el pla de menor durada. En la Competició Internacional de Planificació de l'any 2006 es va presentar el llenguatge PDDL3.0 i es va realitzar la primera i única competició de planificadors amb gestió de restriccions de trajectòries d'estat i preferències. Concretament, en aquesta IPC es van provar quatre dominis amb restriccions temporals on els objectius han de satisfer-se dins d'un límit de temps establit o deadline. Dos planificadors van participar en aquesta competició encara que cap va presentar un bon comportament respecte al compliment dels deadlines. En aquest tipus de problemes, especialment quan es tracta de problemes altament restringits, és crucial detectar la irresolubilitat dels mateixos el més prompte possible, i en aquest punt és on s'estableix el nostre objectiu. Aquest treball de tesi presenta un model de landmarks temporals que permet identificar ràpida i eficientment la irresolubilitat de problemes de planificació amb restriccions. El nostre model incorpora les restriccions temporals del llenguatge PDDL3.0 i estén el concepte de landmark STRIPS al context temporal. A cada landmark temporal se li associen tres tipus d'intervals que s'actualitzen i propaguen d'acord a les relacions d'ordre i restriccions temporals entre ells. Els landmarks temporals juntament amb les seues relacions d'ordre i restriccions temporals formen un graf de landmarks temporals on se sintetitza les relacions que existeixen entre els literals d'un pla solució i, conseqüentment, entre les seues corresponents accions. Quan s'afig nova informació al graf, s'actualitzen i propaguen els intervals dels landmarks, reflectint una imatge més precisa del pla solució. Posteriorment, el model de landmarks temporals s'integra en un planificador temporal heurístic independent del domini denominat TempLM. TempLM utilitza tota la informació extreta del problema per a podar els plans parcials en l'arbre de cerca que no són compatibles amb la informació del graf de landmarks temporals. A més, es disposa d'un procés de retroalimentació entre el graf de landmarks temporals i el procés de cerca d'un pla solució que permet enriquir el graf i, així mateix, limitar l'espai de cerca. Els resultats experimentals mostren que aquesta aproximació detecta ràpidament problemes irresolubles i també és molt efectiva per a resoldre problemes amb restriccions molt ajustades. / Marzal Calatayud, EJ. (2016). Modelo de landmarks temporales para problemas de planificación temporal altamente restringidos [Tesis doctoral no publicada]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/61460 / TESIS

No description available.

Temporal context memory is memory for the timing of events. People can make temporal judgments based on strategies such as assessing the relative familiarity of events or inferring temporal order from the semantic associations among events. The purpose of present study is to investigate the brain regions that support temporal context retrieval in the absence of such non-temporal strategies (i.e. pure temporal context memory). We used three word familiar phrases (triplets) as stimuli. In study phase, three words were presented quickly one after another in either familiar or scrambled order. Participants were instructed to read aloud each word and try to remember the order of the words. Then they were tested on their memory for the order of the words in each triplet. We propose that memory for the scrambled triplets reflects primarily temporal retrieval for two reasons. First, participants were prevented from using semantic strategies during encoding. Second, the relative familiarity of the words in each triplet was similar and not diagnostic of the order of the words during encoding. Neuroimaging results indicate that temporal context retrieval, memory for the order of words in scrambled triplets, was associated with the hippocampus, parahippocampal cortex, ventromedial prefrontal cortex, retrosplenial cortex, and posterior cingulate, which are consistent with the retrieval of non-temporal context in episodic memory. The results also suggest that temporal context retrieval could rely on familiarity, which was demonstrated by the higher accuracy and greater activation of PRc in familiar phrases and scrambled triplets presented in studied order in the test phase. / Master of Science

No description available.

No description available.

No description available.

Automatic test case generation is a process that starts with text based functional requirements which are converted to a formal system requirements model. Once the formal system requirements model is created the automatic test case generation software creates a set of test scenarios that will verify that the requirements are all met. The automatic test case generation software accomplishes the conversion in a four step process: create base scenarios, identify unverified requirements, enhance scenarios to cover all requirements and allow black box testing, and then combine the scenarios into a single scenario tree. The automatic test case generation system outputs a set of scenarios by walking the final scenario tree. This dissertation expands on automatic test case generation for embedded systems in two major ways. The first is to extend functional automatic test case generation to allow for time based requirements as first class objects. The second is to use the automatic test case generation system to enable system partitioning decisions. The addition of time based requirements to the automatic test case generation system allows more complex systems to be developed. By providing a partitioning recommendation based on the test cases generated from the system requirements, the scope and capabilities of a single designer can be expanded to more complex systems. The resulting upgrades to the theory of automatic test case generation could be applied to the existing tools or incorporated in modern UML/SysML based design tools.

No description available.

No description available.