Home < Collaboration:Workshop June 16

Georgia Tech (Core 1) & UC-Irvine (Core 3) Collaboration

Dates: June 16-19, 2005
Place: UC-Irvine
Lodging: Jim Fallon's house
Attendees:

GT gained hands on experience identifying key anatomical structures in the brain, especially relevant to schizophrenia
GT learned detailed rules for other prefrontal regions surrounding DLPFC, rule plans for entire cingulate region, and detailed landmark rules for caudate, putamen, GPe, GPi, septum, nucleus accumbens, thalamus, amygdala, portions of hippocampus, dorsal and ventral midbrain, pons
GT was instucted on the use of UCI's 28 schizophrenic data sets
UCI gained exposure to the new DLPFC Semi-automatic Segmentor and was able to offer suggestions for improvement
UCI gained knowledge of segmentation techniques and pros/cons of automatic segmentation algorithms

Validation of algorithmic success can be based upon: a neuroanatomist's segmentation (i.e. Fallon's Rules), comparison to segmentation of the same patient with a higher resolution MRI, comparison to the same patient with stained autopsy imagery
Algorithms should be centered on a systems approach (i.e. the segmentation is only a means to provide the basis for future DTI or other analysis, especially defining long tracks)

GT to provide UCI with Renderer program (Allen)
GT to explore enhancing Putamen, GPE/GPI images to get same definition as demonstrated in averaged imagery (Allen, John)
GT to build upon DLPFC Semi-automatic segmentor to include surrounding areas (Ramsey)
GT to build a Semi-automatic segmentor of the subcortical areas as included in the Fallon Rules packet (Ramsey, to be completed at a later date)
GT to explore automatic segmentation of distinct gyrii and sulci and the application of labels using modified Fallon Rules (John)
UCI to visit GT
UCI to provide T3 and T1 imagery from the same patients to establish another metric of truth
UCI to facilitate use of 28 schizophrenic cases