Examples of Brain Data Analysis as Motivations for a Data Sharing System

CS Dept. Head Candidate

Dr. Fillia Makedon
Computer Science Department
Dartmouth College

Monday, April 28, 2003
11 a.m. - 12 noon
Fuller Labs 320

Magnetic resonance imaging (MRI) of brain tissues can be used to detect many pathology-related differences (changes) in patients' brains, such as size or shape changes in affected structures. Similarly, functional magnetic resonance imaging (fMRI), an application of MRI to measuring oxygen usage in brain tissue, can reveal differences in activation patterns between patients and clinically normal subjects. Both types of imaging have diagnostic potential.

This talk will describe three examples of brain data analysis that can be supported and enhanced with a data sharing system we are building at Dartmouth. The first example involves experimental results from hippocampal morphometric analysis (considering both shape and volume) using the spherical harmonic shape description and linear classifiers to discriminate with over 90% accuracy between healthy controls and schizophrenics. The second example uses a Fisher linear discriminant classifier to differentiate Alzheimer's patient activation responses from controls with over 80% accuracy. The third example describes a way to integrate and track multimodal MRI Multiple Sclerosis (MS) lesions.

These three examples are being used to develop a secure, scalable and sustainable two-tier data sharing system that enables users to advertise published and unpublished work or to search for desired research data and services through a metadata representation. The negotiation component of the system is to enable agreement on the conditions of sharing prior to the actual data exchange between the data owners and the system users. The talk will describe features of the data sharing system that is currently under development.

Host

Computer Science Dept Head Search Committee

Refreshments will be served.

Maintained by webmaster@wpi.edu
Last modified: Sep 27, 2006, 16:05 EDT