Date of Award
Master of Science (MS)
School of Computing
Dyslexia is a neurodevelopmental reading disability and is supposed to be affecting about 5-15 percent of the population in the United States alone. However, neuroimaging studies in dyslexia research involve relatively small sample sizes, thus limiting inference and the application of novel methods. Besides, the lack of standards among datasets shared makes the datasets useless and also raises questions about the privacy and security of individual subjects involved in the research. Hence, it is essential to develop a data-sharing platform that solves all of these issues. In this thesis, we develop and describe of the platform Dyslexia Data Consortium. The overarching goal of this project is to advance our understanding of a disorder that has significant academic, social, and economic impacts on children, their families, and society. In this platform, researchers can upload and share dyslexia datasets for collaboration. Furthermore, a deep learning-enabled data quality check ensures that the data shared has all the features needed for study and ensures that the subjects’ privacy is protected. Thus, researchers can access shared data to address fundamental questions about dyslexia, replicate findings, apply new methods, and educate the next generation of dyslexia researchers. Moreover, the platform generates secondary datasets from the shared datasets and provides them to researchers. These secondary datasets can be helpful to determine how much previous findings replicate in their samples. In addition, brain volume estimation and correlation analysis help researchers get answers to questions related to dyslexia.
Bhandari, Roshan, "Design and Development of Dyslexia Data Consortium" (2021). All Theses. 3582.