Seminars/Colloquia/Invited Talks

Seminars

Paul Thompson

Analyzing Brain Images from Large Human Populations

PLACE:	Clark 110
EVENT:	CIS Seminar Series
DATE:	September 28, 2005
TIME:	4:15 - 5:00

Abstract

Drawing on large image databases (including 3D brain MRI and PET scans), powerful computer algorithms can now compute information on variations in brain structure and function in thousands of human subjects. We report recent progress in this area, reviewing some of the mathematics used, which includes differential geometry, continuum mechanics, random field theory, PDEs, level sets, and 3D curve and surface models that are used to represent structures in the brain. We use these techniques to create and analyze the first time-lapse movies of brain changes that occur over the human lifespan. These visualize in detail the anatomical patterns and rates of structural brain changes that occur in a variety of diseases. Mathematical algorithms are shown that measure and map growth patterns in the brains of normally and abnormally developing children. They can also track Alzheimer’s disease as it spreads in the living brain, and they can map dynamically changing deficits in schizophrenia, HIV/AIDS, and in drug users. These engineering methods provide exceptional power for understanding how the brain varies in health and disease, and how these changes are affected by medication (e.g., in drug trials), and how they correlate with age and gender, and with genetic and cognitive differences among individuals.

Brief biography

Paul Thompson, Ph.D. is an Associate Professor of Neurology at UCLA. His research focuses on mathematical, software engineering and clinical aspects of brain imaging. His research team develops algorithms for brain image analysis, collaborating with over 20 imaging centers worldwide. Their image analysis methods are now used to investigate a variety of diseases, including Alzheimer’s Disease, HIV/AIDS, schizophrenia, and disorders of brain development, such as Williams syndrome. Dr. Thompson's 300+ collaborative publications describe novel mathematical and computational strategies for analyzing brain structure in health and disease, for detecting pathology in individual patients and groups, and visualizing genetic and developmental effects on brain structure and function. Recent 4D time-lapse visualizations and maps of brain changes in development and disease were published in Nature and PNAS. Dr. Thompson has an M.A. in Mathematics (Oxford University, England), and a Ph.D. in Neuroscience (UCLA). After research as a Fulbright Scholar and Howard Hughes Investigator at UCLA, Dr. Thompson now directs the Image Analysis Core of an NIH-funded National Brain Imaging Resource at UCLA, and his research is funded by the NIA, NIBIB and NCRR. He also serves as Associate Editor for the journals IEEE Transactions on Medical Imaging, Neuroimage, Human Brain Mapping, and Medical Image Analysis.