Genevera I. Allen, Ph.D.
Assistant Professor, Department of Pediatrics-Neurology, Baylor College of Medicine and the Jan and Dan Duncan Neurological Research Institute, Texas Children's Hospital
Assistant Professor, Department of Statistics, Rice University
Research focus: modeling dependence structures in complex data to better understand their biological relevance to conditions of the brain
Dr. Genevera Allen develops mathematical and statistical models for analyzing enormous, complex data sets such as those that are produced by the financial markets or, closer to Dr. Allen’s interests, advanced techniques in neuroimaging, genetics, metabolomics and proteomics. Dr. Allen models the dependence structures in data to better understand their true biological relevance.
For example, functional magnetic resonance imaging (fMRI) can track changes in metabolism of different brain regions in response to cognitive or behavioral tasks in different experimental conditions, or to compare the brain in healthy and disease states. The images are then analyzed to determine whether apparent differences actually reveal changes attributable to the particular experimental activity or abnormality under consideration, as opposed to extraneous patient movements or physiological processes. This analysis is rendered difficult by the complicated dependence structure of the data. Each entity in the image, the voxel, represents a specific location in the brain, and each voxel is measured every couple of seconds. The voxels thus have both spatial and temporal dependencies. By mathematically modeling the spatio-temporal structure directly, one can better understand the biological regions of interest in the brain.
Dr. Allen's specific goal is to develop and apply mathematical and statistical tools in the areas of convex optimization, multivariate analysis, and machine learning to high-dimensional biological data. These tools allow researchers to better separate the biological truth ("signal") from extraneous information ("noise”).