HOME| Alzheimer's| HIV/AIDS| Schizophrenia| Drug Abuse| Development| Projects| Publications| People

(FOR MATH PROJECTS click here)

For incoming students, and anyone else who's interested, here's a summary of some of the projects I'm currently working on, with my research team at the UCLA Laboratory of Neuro Imaging.
I'm including some exciting areas where students coming to the lab briefly for a lab rotation may be interested in helping out. Some short videos describe some of the projects (see symbols below). Contact me if you have any questions! If you are a mathematician or an engineer, you may prefer to look at these projects.
[Updated - June 2007]

Brain Projects


My research focuses on developing new mathematical and computational approaches for analyzing human 3D brain image data. We use these approaches to investigate the major diseases of the human brain, to better understand brain structure and function in health and disease. Our research team consists of neuroscientists, medical doctors, mathematicians and engineers. We collaborate with around 50 labs worldwide (see here) and publish extremely actively.

Patient populations being studied include large numbers of subjects with Alzheimer's Disease, with mild cognitive impairment, and people at genetic risk for Alzheimer's Disease. Another project studies how HIV/AIDS damages the brain. Other active projects focus on understanding brain changes and drug effects in schizophrenia and bipolar disorder. We are studying chronically-medicated populations, first-episode patients, twins discordant for schizophrenia, and children and adolescents with early onset schizophrenia. Our interest in how schizophrenia develops has also led us to expand our studies of brain development . We also study bipolar disorder (in adolescents and adults) and drug effects on the brain (lithium, antipsychotics, anti-dementia drugs). We are creating new approaches to map how the brain grows in childhood and in the teenage years. Finally, we are especially interested in mapping genetic influences on brain structure. Even in normal individuals, it is intriguing to understand how our genes (and other factors) affect our brain structure and function. It can also help us investigate the genetic causes and inherited risks for disease. Part of this work involves constructing population-based brain atlases to encode and represent patterns of anatomic variation, and to detect structural differences in health and disease. These approaches often use some very interesting mathematics as well as high-performance computing techniques. We also have active research projects on methamphetamine abuse, autism, Williams syndrome, fragile X syndrome, brain asymmetry, and IQ (cognitive ability).


In one project, we are creating the first time-lapse maps of Alzheimer's disease spreading in the living brain. We are building a population-based atlas of the brain in Alzheimer's Disease. This is a framework for analyzing brain data in Alzheimer's patients. It combines images and 3D anatomical models with MRI, PET, SPECT and histologic data from large numbers of patients with dementia. Our goal is to determine 3D maps of structural differences, patterns of gray matter loss, anatomic variability, brain asymmetry, and regional atrophy in populations of normal elderly subjects, patients with mild cognitive impairment (MCI), and patients diagnosed with, or at genetic risk for, Alzheimer's Disease. In these projects, new mathematical and computational approaches are being developed, tested and used for analyzing MRI-derived structural models of the cortex, deep sulci, ventricular system, basal ganglia, corpus callosum, amygdala and hippocampal formation. We are especially interested in understanding the patterns of degenerative rates over time. In an MCI clinical trial, we are investigating how these profiles of degeneration can be detected early and slowed down by therapy. We are also developing imaging strategies for early detection and mapping of degenerative change. A new project is linking cortical thinning in Alzheimer's with amyloid burden measured in living patients using FDDNP-positron emission tomography.


Several exciting studies of schizophrenia are underway. Using novel brain mapping approaches that we developed for tracking subtle changes in the brain over time, we are analyzing chronically medicated patients, neuroleptic naive patients, twins discordant for schizophrenia, and children and adolescents with early-onset schizophrenia. We recently charted a dynamic wave of gray matter loss that occurs as schizophrenia develops. We are interested to see if this implicates some non-genetic trigger in initiating the disease. In a clinical trial, we are investigating how antipsychotic drugs such as haloperidol and olanzapine decelerate these brain changes. Using functional MRI (fMRI), we are also studying patterns of brain activation in schizophrenia, and how they relate to genetic polymorphisms and underlying changes in cortical structure.


In another project, we are analyzing growth patterns during normal and abnormal brain development. There are some exciting findings, which we reported recently in the journal Nature, exploring the extremely complex dynamics of brain growth in children ranging from 6 to 15 years of age. Some very nice articles were written about these findings in the news media, and these can be found here. Detailed maps of growth patterns in young children, who were scanned repeatedly with MRI, can be found here. Current work is creating time-lapse movies of brain development based on repeated MRI scans of children as they grow up. We are studying how these brain changes are accelerated or derailed in children and teenagers who develop schizophrenia, or bipolar disorder. We are also developing new brain mapping techniques to help study brain changes in developmental disorders. We are mapping cortical and subcortical structural differences in autistic children and Fragile X syndrome. Another project maps subtle cortical abnormalities in a genetic disorder of brain development: Williams syndrome, where we recently mapped the profile of cortical abnormalities.


This project studies the damage caused by the HIV virus in the brain. We are creating maps of the regions affected, and assessing how these changes relate to cognitive and immune changes in people who are HIV-positive or have AIDS. We are examining how anti-retroviral treatment influences the rate of brain degeneration in HIV patients. Specific projects focus on the cerebellum and examining cognitive correlates of progressive brain changes.


A new project maps the damaging effects of methamphetamine use on human brain structure. Chronic drug abuse induces dramatic changes in the basal ganglia, hippocampus, ventricular system, and cortex. We are beginning to link these changes with metabolic changes, to understand how addictive drugs impact the human brain. We are also mapping the pattern of progressive white matter deterioration in meth users and its relation to depression.


We are currently conducting a broad range of projects analyzing tumor growth and therapeutic response in patients with malignant gliomas and glioblastoma multiforme. Using a variety of brain mapping approaches, our goal is to understand how drug treatment, radiotherapy, and surgical resection modulate the rates of tumor growth, as well as other imaging indices of tumor progression, over time. We are also developing new types of brain mapping tools that combine maps of local growth rates, spectroscopic maps, parametric MRI imaging, and pathologic and genetic measures of tumor changes. We are also investigating patients' response to therapy, to understand which therapies are best, and how they affect tumor progression and recurrence in different patients. We are creating detailed maps of dynamic response to the chemotherapeutic agent Temodar (temozolamide). We are also uncovering the inter-relationships among the major imaging measures as tumors evolve. These projects involve a significant amount of computational modeling. Finally, we are beginning a new intraoperative imaging project linking detailed maps of tumor growth rates with differential expression of pathological and genetic markers in tissue biopsied during surgery. This key project will be the first to link genetic and molecular cascades implicated in tumor growth and response with 3-dimensional changes in imaging markers observable at high-resolution in individual patients. Since tumor growth is heterogeneous in individual patients, this project will clarify the relationships between the molecular and cellular targets of therapy and fine-scale changes detectable with imaging through repeated short-interval scanning.


We have a major funded effort to determine how our genes affect brain structure, function and fiber connectivity. Genetic brain maps, in particular, can show whether we inherit patterns of brain structure from our parents, and if so, to what degree. We especially want to understand which parts of the brain are most strongly determined by our genes. We are scanning 700 twins with high-angular resolution diffusion imaging (HARDI) at 4 Tesla to examine which genes influence fiber integrity and connectivity in the brain. (Read more about HARDI on the math page.) Parallel studies are designed to help understand familial risk for diseases that affect the brain, as well as their genetic transmission.


The remaining projects concentrate on the mathematical and engineering aspects of brain imaging. One important on-going project focuses on developing complex surface modeling and 3D warping algorithms for brain data. Differences in brain structure make it hard to compare data from different subjects, and to distinguish abnormal structure from normal anatomic variations. Transforming 3D brain data into the shape of a single target anatomy, or onto a neuroanatomic atlas, removes subject-specific shape variations, and allows us to compare and integrate 3D functional brain imaging data across subjects and groups. Applications of highly non-linear image warping algorithms include the transfer of multi-subject 3D functional, vascular and histologic maps onto a single anatomic template, and the mapping of 3D atlases onto the scans of new subjects. Local shape changes can also be detected in 3D medical images in disease, and during normal and abnormal growth and development. To tackle this problem, I have been developing and testing algorithms for calculating biologically-driven flow field transformations of extremely high spatial dimension, which warp one anatomic scan into structural correspondence with another. These warping transforms measure the shape differences between the anatomies of the different subjects. We are also developing approaches to compare these brain mapping algorithms, and to automatically find structures in brain images, based on the mathematics of deformable templates in diffused potential fields. Our newest work in this area uses level sets (implicit functions), 2D and 3D harmonic maps, sphere carving, and fluid PDEs for nonlinear image registration and cortical surface mapping.


As a more technical project, I am developing strategies and algorithms to create a comprehensive probabilistic atlas of the human brain based on high-dimensional random tensor field transformations. This new type of probabilistic, digital brain atlas is designed to detect and measure structural abnormalities throughout new subjects' 3D MRI scans. The ultimate goal is to determine (1) whether the detected anatomic variants are characteristic of certain disease states (e.g., Alzheimer's), (2) how they relate to genetic, therapeutic, and demographic risk factors, and (3) how we can distinguish them mathematically from normal patterns of brain variation. More information on probabilistic brain atlases, and the many thousands of subjects that underlie them, can be found here.

Often disease-specific patterns that are hard to see in an individual brain scan begin to emerge after averaging brain data from large numbers of subjects. Construction of a population-based brain atlas requires the warping of large numbers of brain images into structural correspondence, prior to the estimation of an anisotropic Gaussian random field, to represent the calculated structural variations, on a high-resolution image lattice. Given a 3D brain image of a new subject, the algorithms calculate a set of high-dimensional volumetric maps, typically with 384x384x256x3 (~0.1 billion) degrees of freedom, fluidly deforming this scan into structural correspondence with other normal scans, selected one by one from an anatomic image archive. The family of volumetric warps so constructed encodes statistical properties of local anatomical variation throughout the architecture of the brain. Probability maps can then be generated, which quantify the severity of structural abnormalities in the anatomy of the new subject. These new techniques for brain image analysis are reviewed here, and less mathematically here. Our development of supercomputing approaches to help solve these problems is reviewed here.


Results of these projects, as well as other exciting new projects on brain development, brain atlases, and analysis of functional, neurochemical and histologic maps in Alzheimer's Disease, may be found in the form of over 600 publications, with summaries on the Internet at: http://users.loni.usc.edu/~thompson/thompson_pubs.html

If any of these papers look interesting to you, please send me a message by e-mail, and I'll be glad to send you some free reprints of any papers you would like!