|PAUL THOMPSON'S LATEST IMAGES|
[Article on Computational Challenges in Medical Imaging]
|MAPPING BRAIN VARIABILITY IN HUMAN POPULATIONS|
Mapping Brain Variability in Human Populations. Just as faces differ among
different individuals, brain structure varies even among normal individuals,
and even more so in disease. This map displays the magnitude and principal
directions of anatomical variability in the brain, based on a group of 40 normal subjects scanned with MRI.
Pink colors indicate brain regions
with large anatomical variability across subjects, blue colors low
variability. The ellipsoidal glyphs are elongated along directions in which
variation is greatest. The nested, transparent glyphs represent the
probabilities of different anatomical variations, and are used to detect
abnormal anatomical differences in Alzheimer's Disease. These maps help distinguish pathological changes from normal variations.
[First 3 images by Paul Thompson, Andrew Lee, Kiralee Hayashi, Agatha Lee, and Arthur Toga;
final 2 images generated by Paul Thompson, Colin Holmes, and Arthur Toga].
|MAPPING GROWTH RATES IN BRAIN TUMORS|
Mapping Growth Rates in Brain Tumors. Tensor maps, such as the one above, can reveal the complex profiles of growth in brain tumors. In a patient with a malignant brain tumor, 3D models of the tumor were made based on 3D MRI scans of the brain, after a patient was scanned 15 times in 104 days. Using an algorithm that executes in parallel on a networked cluster of 50 R10000 workstations, an 3D elastic deformation field was computed to represent the changes in the brain tumor over time. Both the magnitude and principal directions of growth can be visualized; the shape and orientation of the glyphs help illustrate the dynamics of growth in the tissue. These maps may be used to help determine whether chemotherapy is effective in stemming tumor growth. [Visualization by Colin Holmes, Paul Thompson, and Arthur Toga].
|MAPPING BRAIN TISSUE LOSS IN ADOLESCENTS WITH SCHIZOPHRENIA|
Mapping Brain Tissue Loss in Adolescents with Schizophrenia. This map reveals the 3-dimensional profile of gray matter loss in the brains of teenagers with early-onset schizophrenia, with a region of greatest loss in the temporal and frontal brain regions that control memory, hearing, motor functions, and attention. Using novel image analysis algorithms, dramatic reductions in the profiles of gray matter were detected, based on a database of 96 images from schizophrenic patients scanned repeatedly with MRI. The parallel extraction of anatomical models from all patients in the image database required 60 CPU hours, when running in parallel on an SGI RealityMonster with 32 internal CPUs. [Image by Paul Thompson, Christine Vidal, Judy Rapoport, and Arthur Toga].
|PUBLICATIONS |CONTACT ME |RESUME |HOMEPAGE|