The goal of this project is to develop high resolution functional magnetic resonance imaging (fMRI) techniques to map human brain activity using standard clinical MR scanners. High resolution is needed for fMRI for three reasons: 1) to overcome partial volume effects in small, functionally-relevant areas of activation; 2) to reduce artifactual signal loss due to magnetic susceptibility effects in functionally important areas of the brain (e.g., near the frontal sinuses) and 3) to accurately determine location of activation relative to cortical structure. The fMRI techniques that have produced the most reliable results to date use-high- speed imaging devices, but these are costly, currently have limited availability and produce only moderate resolution images. The techniques currently available for use with standard clinical MR scanners are subject to artifacts due to physiological factors. The proposed work, if successful, will lead to spiral k-space rapid imaging methods, which are robust to artifacts, have high-spatial resolution, and can be used on standard clinical MR scanners. The fundamental hypothesis is that spiral k-space imaging will yield more consistent and reliable activation than other high-resolution fMRI methods. Preliminary results indicate that this technique is significantly less sensitive to subtle artifacts arising from physiological sources. In other preliminary results, this technique was used to measure activation in prefrontal cortex in normal subjects. In the proposed work, these artifacts and their sources will be explored in detail and methods will be developed to minimize their effects. A variety of techniques to achieve both high in-plane and through-plane resolution with spiral k-space imaging will be investigated. Finally, a formal comparison of the spiral k-space with other high-resolution techniques will be conducted. This will be done in a pilot study which examines activation of prefrontal cortex during a working memory task in both normal control subjects and schizophrenics. The major limiting factors for other functional neuroimaging modalities have been the cost and invasiveness of the scanning procedures, as well as the limited/poor ability for spatial localization. The development of high resolution fMRI methods that can be used on widely available clinical MR scanners promises to dramatically reduce both of these limitations. The techniques we propose will permit detailed mapping of the functional organization of the human brain, using equipment that is already accessible to a large number of investigators and clinicians. These techniques promise to have immediate application in research on neurologic and psychiatric disease, and could rapidly lead to clinical use in these areas.
Showing the most recent 10 out of 16 publications
