Dementia is increasing recognized as a major problem to public health worldwide. Prevention and treatment strategies are in critical need. Nowadays, research for dementia usually featured as complex longitudinal studies, which provide extensive information and also propose challenge to statistical methodology. The purpose of this dissertation research was to apply statistical methodology in the field of dementia to strengthen the understanding of dementia from three perspectives: 1) Application of statistical methodology to investigate the association between potential risk factors and incident dementia. 2) Application of statistical methodology to analyze changes over time, or trajectory, in cognitive tests and symptoms. 3) Application of statistical learning methods to predict development of dementia in the future. Prevention of Alzheimer’s disease with Vitamin E and Selenium (PREADViSE) (7547 subjects included) and Alzheimer’s disease Neuroimaging Initiative (ADNI) (591 participants included) were used in this dissertation. The first study, “Self-reported sleep apnea and dementia risk: Findings from the PREADViSE Alzheimer’s disease prevention trial ”, shows that self-reported baseline history of sleep apnea was borderline significantly associated with risk of dementia after adjustment for confounding. Stratified analysis by APOE ε4 carrier status showed that baseline history of sleep apnea was associated with significantly increased risk of dementia in APOE ε4 non-carriers. The second study, “comparison of trajectories of episodic memory for over 10 years between baseline normal and MCI ADNI subjects,” shows that estimated 30% normal subjects at baseline assigned to group 3 and 6 stay stable for over 9 years, and normal subjects at baseline assigned to Group 1 (18.18%) and Group 5 (16.67%) were more likely to develop into dementia. In contrast to groups identified for normal subjects, all trajectory groups for MCI subjects at baseline showed the tendency to decline. The third study, “comparison between neural network and logistic regression in PREADViSE trial,” demonstrates that neural network has slightly better predictive performance than logistic regression, and also it can reveal complex relationships among covariates. In third study, the effect of years of education on response variable depends on years of age, status of APOE ɛ4 allele and memory change.

No description available.

Title: Borderline personality disorder: examining trajectories of development among adolescents Background: Borderline personality disorder (BPD) tends to be highly comorbid with other disorders. In adolescence, information about the classification and development of BPD is in its early stages. There is limited empirical research available that investigates predictors of clinically significant symptom trajectories of the disorder using data collected in childhood. Given the enormous personal and societal costs associated with BPD, early detection and prevention is important. Clinical implications of this research include an improved understanding of risk factors and possible mechanisms for development of BPD symptomatology. Objectives: To identify trajectories of BPD symptomatology in a Canadian sample of adolescents (N = 703) assessed at ages 13, 14, 15 and 16, while examining predictors of trajectory group membership assessed at age 12. Methods: Data from the McMaster Teen Study was used to examine trajectories of BPD symptoms using group-based trajectory modeling. The influence of gender, depression, ADHD, family functioning and various sociodemographic variables as predictors of an individual’s group membership was tested. Chi-square, analysis of variance and multinomial logistic regression was used to analyze the data. Results: A four-group trajectory model was most robust at describing BPD symptomatology in this age group. Univariate analyses supported female gender, depression and ADHD at baseline, parental age, marital status, education, and income as significant predictors of group membership. Female gender, depression and ADHD severity at baseline were significant predictors of group membership when adopting a multivariate approach. There is a greater prevalence of girls with higher depression and ADHD scores in the high-increasing features and BPD group. Conclusion: Findings demonstrate four various developmental trajectories of BPD features. Results further the understanding of the factors associated with development of the disorder across time. / Thesis / Master of Science (MSc) / Information about the classification and development of borderline personality disorder (BPD) in adolescence is in its early stages. While evidence for similar construct validity to the adult disorder exists for adolescents, major gaps in knowledge regarding the stability in course of BPD symptoms and predictors of clinically significant symptom trajectories in this age group remain. As most clinicians will assess youth already having significant features of the disorder, early detection requires knowledge of the indicators that precede an unfavourable trajectory. This dissertation will help address these gaps by modeling trajectories of BPD symptoms in youth across ages 13-16, whilst examining factors influencing trajectory group membership.

At international border areas that suffer from poor quality, assessment of pollutant sources and transport across the border is important for designing effective air quality management strategies. As part of the Cal-Mex 2010 field campaign at the US-Mexico border in San Diego and Tijuana, we measured black carbon (BC) concentrations at three locations in Mexico and one in the United States. The measurements were intended to support the following objectives: to characterize the spatial and temporal variability in BC concentrations and emissions in the border region, to identify potential source areas of BC emissions, and to characterize the cross-border transport of BC and assess its impact on local and regional air quality. BC concentrations at Parque Morelos, the campaign's supersite, averaged 2.1 ?g m?? and reached a maximum value of 55.9 ?g m??. This average value is comparable to levels in large American cities like Los Angeles and similarly sized Mexican cities like Mexicali. The maximum value occurred near midnight, and similar incidents were observed on nearly half of the overnight monitoring periods. BC and carbon monoxide (CO) were strongly correlated at the Mexican sites. The BC/CO ratio was ~3 times higher in Tijuana than in Mexico City, suggesting that gasoline-powered vehicles in Tijuana emit more BC than is typical or that diesel vehicles comprise a relatively high proportion of the vehicle fleet. Tijuana's emissions of BC are estimated to be 380-1470 metric tons yr??. BC measurements were used in conjunction with modeled wind fields to simulate forward and backward particle trajectories. Generally, BC in Tijuana appears to originate locally, as backward simulations showed transport from the US into Mexico at only one site. The majority of the trajectory analyses indicate that there is often transport from Tijuana into the US, crossing the border in a northeasterly direction to the east of San Diego-Tijuana and sometimes as far east as Imperial County at the eastern edge of California. These results suggest that any air quality management strategies considering BC should account for contributions from the border region, as BC is chemically inert in the atmosphere and can travel up to thousands of kilometers. / Master of Science

Thesis advisor: Barbara E. Wolfe / Despite the chronicity and less than optimal outcomes of inpatient treatment (IPT) for anorexia nervosa (AN), treatment guidelines continue to reflect the common notion of one-size-fits-all and the process of weight restoration continues to be poorly understood. Weight restoration, a primary goal of IPT for AN, does not occur in isolation but rather reflects an adaptation process within internal and external environments. It is unknown whether or not there are unique patterns of change in body weight that are associated with factors identified in the existing literature as being predictors of weight gain. The purpose of this study was to explore the extent to which patterns of change in body weight existed among individuals during IPT for AN, and the relationship with factors identified in the existing literature as being predictors of weight gain (i.e., age at time of admission, admission caloric intake, percent of ideal body weight [IBW] at time of admission, body weight at time of discharge, body mass index [BMI] at time of discharge). Individuals who were diagnosed with AN and admitted to the inpatient unit of an eating disorder treatment facility in the Northeast between January 1, 2012 to December 31, 2015 were included in this retrospective, exploratory study (N = 500). Group-based trajectory modeling (GBTM) was used to identify distinct trajectories of change in body weight, and to determine the risk of being in a particular trajectory. Four distinct trajectories were identified: weight gain (n = 197), weight loss (n = 177), weight plateau (n = 82), and weight fluctuate (n = 44) groups. Significant predictors of trajectories were age, history of prior IPT for AN, admission caloric intake, body weight at time of admission and discharge, and length of stay. Results from this study suggest that a further understanding of patterns of change in body weight among individuals with AN, will help guide assessment and treatment interventions and consequently influence outcomes. Additionally, there is an opportunity to update treatment guidelines and recommendations for AN. / Thesis (PhD) — Boston College, 2016. / Submitted to: Boston College. Connell School of Nursing. / Discipline: Nursing.

Background For the 1.2 million older adults residing in U.S. nursing homes, little is known about their experience with physical frailty and cognitive impairment, two critical interrelated aging conditions. Methods Minimum Data Set 3.0 was used. Physical frailty was measured by FRAIL-NH and cognitive impairment by Brief Interview for Mental Status and Cognitive Performance Scale. Demographic and clinical characteristics were adjusted accordingly. Aim 1 described the prevalence of physical frailty and cognitive impairment and longitudinally examined the association between two conditions with the non-proportional odds model. Aim 2 used latent class analysis to identify physical frailty subgroups and estimated their association with cognitive impairment using multinomial logistic regression. Aim 3 fitted group-based trajectory models to identify physical frailty trajectories and cognitive impairment trajectories and quantified the association between the two sets of trajectories. Main Results Around 60% of older residents were physically frail and 68% had moderate/severe cognitive impairment, with improvement and worsening observed in both conditions, particularly in the first three months. Older residents with moderate/severe cognitive impairment were consistently and increasingly more likely to be frail. Three physical frailty subgroups were identified at admission. Greater cognitive impairment was associated with higher odds to belong to “severe physical frailty”. Five physical frailty trajectories and three cognitive impairment trajectories were identified over the first six months. One in five older residents were “Consistently Frail” and “Consistently Severe Cognitive Impairment”. Conclusion Findings emphasized the need for care management tailored to the heterogeneous presentations and progression trajectories of physical frailty and cognitive impairment.

The present study focuses on developing an efficient and stable unmanned aerial system traffic management (UTM) system that utilizes a data-driven target tracking method and a distributed path planning algorithm for multiple Unmanned Aerial Vehicle (UAV) operations with local dynamic networks, which can provide flexible scalability, enabling autonomous operation of a large number of UAVs in dynamically changing environment. Traditional dynamic motion-based target tracking methods often encounter limitations due to their reliance on a finite number of dynamic motion models. To address this, data-driven target tracking methods were developed based on the statistical model of the Gaussian mixture model (GMM) and deep neural networks of long-short term memory (LSTM) model, to estimate instant and future states of UAV for local path planning problems. The estimation accuracy of the data-driven target tracking methods were analyzed and compared with dynamic model-based target tracking methods. A hybrid dynamic path planning algorithm was proposed, which selectively employs grid-free and -based path search methods depending on the spatio-temporal characteristics of the environments. In static environment, the artificial potential field (APF) method was utilized, while the $A^*$ algorithm was applied in the dynamic state environment. Furthermore, the data-driven target tracking method was integrated with the hybrid path planning algorithm to enhance deconfliction. To ensure smooth trajectories, a minimum snap trajectory method was applied to the planned paths, enabling controller tracking that remains dynamically feasible throughout the entire operation of UAVs. The methods were validated in the Software-in-the-loop (SITL) demonstration with the simple PID controller of the UAVs implemented in the software program. / Ph.D. / This dissertation focuses on developing data-driven models for tracking and path planning of Unmanned Aerial Vehicle (UAV) in dynamic environments with multiple operations. The goal is to improve the accuracy and efficiency of Unmanned Aircraft System traffic management (UTM) under such conditions. The data-driven models are based on Gaussian mixture model (GMM) and long-short term memory (LSTM) and are used to estimate the instant and consecutive future states of UAV for local planning problems. These models are compared to traditional target tracking models, which use dynamic motion models like constant velocity or acceleration. A hybrid dynamic path planning approach is also proposed to solve dynamic path planning problems for multiple UAV operations at an efficient computation cost. The algorithm selectively employs a path planning method between grid-free and grid-based methods depending on the characteristics of the environment. In static state conditions, the system uses the artificial potential field method (APF). When the environment is time-variant, local path planning problems are solved by activating the $A^*$ algorithm. Also, the planned paths are refined by minimum snap trajectory to ensure that the path is dynamically feasible throughout a full operation of the UAV along with controller tracking. The methods were validated in the Software-in-the-loop (SITL) demonstration with the simple PID controller of the UAVs implemented in the software program.

From 2001-2007, students from India have consistently comprised the largest ethnic group of international students on college campuses across the United States (Open Doors: Report on International Educational Exchange, 2007). Despite a number of studies that have researched the mental health of international students in the U.S., none have done so primarily with Indian graduate students. Theoretical and empirical literature regarding the psychological changes and acculturation patterns that international students undergo after their transition do not explore the possibility of multiple pathways of change. The current study identified four separate mental health trajectories for Indian international graduate students during their first year in the U.S. It also found three distinct patterns of acculturation for the Indian culture and four acculturation trajectories for the European American culture. The size of one's adjustment, feelings about transition, gender role attitudes, and availability of out-group support were all significant contributors to the variability among empirically derived mental health trajectories.

Turbine engine blades are subjected to extreme conditions characterized by significant and simultaneous excursions in both stress and temperature. These conditions promote thermo-mechanical fatigue (TMF) crack growth which can significantly reduce component design life beyond that which would be predicted from isothermal/constant load amplitude results. A thorough understanding of the thermo-mechanical fatigue crack behavior in single crystal superalloys is crucial to accurately evaluate component life to ensure reliable operations without blade fracture through the use of "retirement for cause" (RFC). This research was conducted on PWA1484, a single crystal superalloy used by Pratt & Whitney for turbine blades. Initially, an isothermal constant amplitude fatigue crack growth rate database was developed, filling a void that currently exists in published literature. Through additional experimental testing, fractography, and modeling, the effects of temperature interactions, load interactions, oxidation and secondary crystallographic orientation on the fatigue crack growth rate and the underlying mechanisms responsible were determined. As is typical in published literature, an R Ratio of 0.7 displays faster crack growth when compared to R = 0.1. The effect of temperature on crack growth rate becomes more pronounced as the crack driving force increases. In addition secondary orientation and R Ratio effects on crack growth rate were shown to increase with increasing temperature. Temperature interaction testing between 649°C and 982°C showed that for both R = 0.1 and 0.7, retardation is present at larger alternating cycle blocks and acceleration is present at smaller alternating cycle blocks. This transition from acceleration to retardation occurs between 10 and 20 alternating cycles for R = 0.1 and around 20 alternating cycles for R = 0.7. Load interaction testing showed that when the crack driving force is near KIC the overload size greatly influences whether acceleration or retardation will occur at 982°C. Semi-realistic spectrum testing demonstrated the extreme sensitivity that relative loading levels play on fatigue crack growth life while also calling into question the importance of dwell times. A crack trajectory modeling approach using blade primary and secondary orientations was used to determine whether crack propagation will occur on crystallographic planes or normal to the applied load. Crack plane determination using a scanning electron microscope enabled verification of the crack trajectory modeling approach. The isothermal constant amplitude fatigue crack growth results fills a much needed void in currently available data. While the temperature and load interaction fatigue crack growth results reveal the acceleration and retardation that is present in cracks growing in single crystal turbine blade materials under TMF conditions. This research also provides a deeper understanding of the failure and deformation mechanisms responsible for crack growth during thermo-mechanical fatigue. The crack path trajectory modeling will help enable "Retirement for Cause" to be used for critical turbine engine components, a drastic improvement over the standard "safe-life" calculations while also reducing the risk of catastrophic failure due to "chunk liberation" as a function of time. Leveraging off this work there exists the possibility of developing a "local approach" to define a crack growth forcing function in single crystal superalloys.

Cette thèse se concentre sur la structuration du modèle acoustique pour améliorer la reconnaissance de la parole par modèle de Markov. La structuration repose sur l’utilisation d’une classification non supervisée des phrases du corpus d’apprentissage pour tenir compte des variabilités dues aux locuteurs et aux canaux de transmission. L’idée est de regrouper automatiquement les phrases prononcées en classes correspondant à des données acoustiquement similaires. Pour la modélisation multiple, un modèle acoustique indépendant du locuteur est adapté aux données de chaque classe. Quand le nombre de classes augmente, la quantité de données disponibles pour l’apprentissage du modèle de chaque classe diminue, et cela peut rendre la modélisation moins fiable. Une façon de pallier ce problème est de modifier le critère de classification appliqué sur les données d’apprentissage pour permettre à une phrase d’être associée à plusieurs classes. Ceci est obtenu par l’introduction d’une marge de tolérance lors de la classification ; et cette approche est étudiée dans la première partie de la thèse. L’essentiel de la thèse est consacré à une nouvelle approche qui utilise la classification automatique des données d’apprentissage pour structurer le modèle acoustique. Ainsi, au lieu d’adapter tous les paramètres du modèle HMM-GMM pour chaque classe de données, les informations de classe sont explicitement introduites dans la structure des GMM en associant chaque composante des densités multigaussiennes avec une classe. Pour exploiter efficacement cette structuration des composantes, deux types de modélisations sont proposés. Dans la première approche on propose de compléter cette structuration des densités par des pondérations des composantes gaussiennes dépendantes des classes de locuteurs. Pour cette modélisation, les composantes gaussiennes des mélanges GMM sont structurées en fonction des classes et partagées entre toutes les classes, tandis que les pondérations des composantes des densités sont dépendantes de la classe. Lors du décodage, le jeu de pondérations des gaussiennes est sélectionné en fonction de la classe estimée. Dans une deuxième approche, les pondérations des gaussiennes sont remplacées par des matrices de transition entre les composantes gaussiennes des densités. Les approches proposées dans cette thèse sont analysées et évaluées sur différents corpus de parole qui couvrent différentes sources de variabilité (âge, sexe, accent et bruit) / This thesis focuses on acoustic model structuring for improving HMM-Based automatic speech recognition. The structuring relies on unsupervised clustering of speech utterances of the training data in order to handle speaker and channel variability. The idea is to split the data into acoustically similar classes. In conventional multi-Modeling (or class-Based) approach, separate class-Dependent models are built via adaptation of a speaker-Independent model. When the number of classes increases, less data becomes available for the estimation of the class-Based models, and the parameters are less reliable. One way to handle such problem is to modify the classification criterion applied on the training data, allowing a given utterance to belong to more than one class. This is obtained by relaxing the classification decision through a soft margin. This is investigated in the first part of the thesis. In the main part of the thesis, a novel approach is proposed that uses the clustered data more efficiently in a class-Structured GMM. Instead of adapting all HMM-GMM parameters separately for each class of data, the class information is explicitly introduced into the GMM structure by associating a given density component with a given class. To efficiently exploit such structured HMM-GMM, two different approaches are proposed. The first approach combines class-Structured GMM with class-Dependent mixture weights. In this model the Gaussian components are shared across speaker classes, but they are class-Structured, and the mixture weights are class-Dependent. For decoding an utterance, the set of mixture weights is selected according to the estimated class. In the second approach, the mixture weights are replaced by density component transition probabilities. The approaches proposed in the thesis are analyzed and evaluated on various speech data, which cover different types of variability sources (age, gender, accent and noise)