Search

Background: Knee osteoarthritis (OA) is a disease affecting the whole joint including cartilage, bone and soft tissues. These structures can be assessed quantitatively using magnetic resonance imaging (MRI) though assessment is time consuming. Quantitative MRI assessment holds promise as a reliable method of determining treatment response in clinical trials. The main aim of the thesis was to determine whether vitamin D therapy, compared to placebo, is linked with a reduction in synovial tissue volume (STV) and subchondral bone marrow lesion (BML) volume in men and women with symptomatic knee OA. The objectives were; 1) to apply a semi-automated approach to the assessment of STV on MRI and to examine performance compared with manual segmentation and 2) to assess the relationship between MRI structural outcomes including STV and BML types and symptoms in men and women with symptomatic knee OA. Methods: Objective 1: 32 sagittal T1-weighted (T1-w) fat suppressed (FS) contrast enhanced (CE) MRI scans, acquired from a clinical study of brace therapy in symptomatic knee OA (BRACE), were segmented for STV using manual segmentation and a semi-automated approach based on the application of a 3D mask developed using active appearance modelling (AAM). Aim and objective 2: Data were acquired on a subsample of participants recruited to a multicentre randomised double-blind placebo-controlled trial of vitamin D therapy in men and women with symptomatic knee OA (UK VIDEO). A subsample of participants had MRI performed at yearly intervals from baseline at one study site and completed the WOMAC questionnaire. In those with serial images including a baseline scan (N = 50), STV was segmented on sagittal T1-w FS CE scans using a semi-automated approach whilst BML volume was segmented manually. BMLs were segmented by location (subchondral / ligament-based) and by the presence/absence of a cyst-like component. Results: There was no significant difference in STV measured using the semi-automated approach compared to manual segmentation (mean difference: 207.2 mm3, 95% CI -895.2 to 1309.7) though the approach was quicker (18 vs. 71 minutes). Reliability was very good. Using data from UK VIDEO, an increase in knee pain was associated with an increase in STV (b = 2.2, 95% CI 0.6 to 3.7), the volume of subchondral BMLs with cyst-like components (b = 1.3, 95% CI 0.1 to 2.4) and the volume of the cyst-like component itself (b = 0.06, 95% CI 0.02 to 0.1). There was no significant difference at 2 years between the vitamin D and placebo groups in mean change from baseline for STV (b = 93.9 mm3, 95% CI -1605.0 to 1792.7) and subchondral BML volume (b = -313.5 mm3, 95% CI -4244.7 to 3617.7). Conclusions: A semi-automated approach to the assessment of STV is reliable and is more time efficient than manual segmentation. STV assessed using this approach and also subchondral BMLs with cyst-like components were associated with increasing knee pain. Vitamin D supplementation for 2 years had no significant effect on change in subchondral BML volume or STV in men and women with symptomatic knee OA.

This thesis examines the validity of school value-added measures and the validity of arguments for their interpretation and use. The opening chapters review the development of school value added measures, existing evidence on their properties and validity and their current use in research, policy and practice. The empirical results are based on four studies using English National Pupil Database data and a large, nationally-representative dataset of teacher-assessed attainment data for English pupils aged from 7 to 13. The findings all relate to the properties of school value-added measures and the seriousness of a number of threats to their validity. The four empirical studies examine the following issues: observable bias and error, inter-method reliability when compared to estimates from a quasi-experimental regression discontinuity design, stability of school value added scores and of specific cohorts over time, and consistency of school value-added scores within cohorts and between different school cohorts at a single point in time. The closing chapters discuss the validity of value-added measures in general and in relation to the areas of use identified. Individually and collectively, the results advance understanding of numerous threats to validity and have substantial implications for the use of value-added measures in research, policy and practice.

Programmers and engineers in the domains of high performance computing (HPC) and electronic system design have a shared goal: to define a structure for coordination and communication between nodes in a highly parallel network of processing tasks. Practitioners in both of these fields have recently encountered additional constraints that motivate the use of multiple types of processing device in a hybrid or heterogeneous platform, but constructing a working "program" to be executed on such an architecture is very time-consuming with current domain-specific design methodologies. In the field of HPC, research has proposed solutions involving the use of alternative computational devices such as FPGAs (field-programmable gate arrays), since these devices can exhibit much greater performance per unit of power consumption. The appeal of integrating these devices into traditional microprocessor-based systems is mitigated, however, by the greater difficulty in constructing a system for the resulting hybrid platform. In the field of electronic system design, a similar problem of integration exists. Many of the highly parallel FPGA-based systems that Xilinx and its customers produce for applications such as telecommunications and video processing require the additional use of one or more microprocessors, but coordinating the interactions between existing FPGA cores and software running on the microprocessors is difficult. The aim of my project is to improve the design flow for hybrid systems by proposing, firstly, an abstract representation of these systems and their components which captures in metadata their different models of computation and communication; secondly, novel design checking, exploration and optimisation techniques based around this metadata; and finally, a novel design methodology in which component and system metadata is used to generate software simulation models. The effectiveness of this approach will be evaluated through the implementation of two physical-layer telecommunications system models that meet the requirements of the 3GPP "LTE" standard, which is commercially relevant to Xilinx and many other organisations.

Marietta College has been offering graduate programs for more than 30 years, and since 2002 has added three graduate programs. However, during this time there has been minimal marketing of the programs. This research is designed to find an effective way to promote both the established programs and the newer programs to help increase enrollment at Marietta College.

The results of inelastic light scattering experiments on two different quasi-2-dimensional electron systems are presented here. In the first system we study the effects of spatial confinement on the energy spectrum of shallow (Si) donors. Resonant Raman experiments on GaAs(Si)-Al(,x)Ga(,1-x)As multiple quantum well heterostructures show new features attributed to the presence of the Si impurities. The transition from 1s to 2s donor states, derived from the lowest conduction subband, has been identified. There is good agreement between the measured energy of this excitation and recent calculations. The width and lineshape of the feature give an estimate of Si diffusion during growth. Another effect attributed to the presence of the donors is a splitting of free intersubband transition peaks. The low energy peaks of the doublets behave like free intersubband transitions. The high energy peak ((TURN)2 meV above the free intersubband energy) is studied as a function of power, temperature, and GaAs well width. These results are consistent with assigning this peak to a 1s donor ground state to 1s resonant impurity state excitation. This represents the first observation of resonant impurity states in semiconductor multiple quantum well heterostructures. The second system considered is that of semi-insulating GaAs codoped with Cr and a donor (e.g., Sn or Te). GaAs is a material that has undergone extensive study with Raman spectroscopy; however, none of the previous results have shown evidence of 2-dimensional plasmas. The data presented here show an electronic feature whose position, lineshape, and power dependence are inconsistent with bulk electronic scattering, but consistent with scattering from a photoexcited 2-dimensional electron gas. The confining potential which binds the 2-dimensional electron gas is created at the surface of the GaAs by a mismatch in the Fermi levels at the surface and in the bulk. Our results show both collective and single particle-like transitions of the photoexcited electrons analogous to intersubband transitions in Si MOSFETs and heterostructure systems. By comparing these results with calculations we are able to give an estimate of the free carrier density and Fermi level mismatch.