IBN: 9634419 PI: Leonard Modern technical improvements have made it possible to image fine details of the human brain during life. Unfortunately, the great individual variability that characterizes all biological phenomena makes the study of the brain extremely difficult and time consuming. Since the images are digitized, it should be possible to extract features of interest automatically. In this project, a neuroscientist, an applied mathematician and an electrical engineer will initiate a collaboration designed to to develop such automation. The first step will be to identify prominent skull landmarks to triangulate the medial origin of the feature, using a technique that has been proved very effective in ultrasound images of the heart. Once the feature is located, neural network techniques originally developed for automatic target recognition will be utilized to measure its shape and volume. The long term goal is to develop automated methods for identifying major functional areas in the brain. Such work is a necessary prerequisite for the interpretation of functional imaging data. The project should produce techniques applicable in a wide variety of fields, including cognitive neuroscience, education, industrial, developmental and cognitive psychology, and medicine.
