Brain Injury Detection and Monitoring through fMRI Time Series Data Mining

摘要

The fundamental purpose of functional Magnetic Resonance Imaging (fMRI) based detection and monitoring is to provide tangible benchmarks for neuronal activity via hemodynamic response characteristics. Voxel intensity maps are calculated from normalized images, the maps are then stored in as a single dataset in order to gain knowledge from their collective analysis. At this point data mining techniques are employed to control for certain levels of treatment as well as demographic conditions. The model presented involves three stages; data collection, preprocessing, and data mining/statistical analysis. The general idea for the proposed fMRI based detection and monitoring system is to leverage high precision hidden knowledge from medical images captured for other purposes.