Abstract
Gradient-echo echo-planar imaging, GRE-EPI, is the most used imaging technique in functional magnetic resonance imaging, fMRI, based on the blood oxygenation level-dependent, BOLD, effect. BOLD fMRI exploits susceptibility differences between oxygenated and deoxygenated blood to detect neuronal activation. However, the sequence parameters which make GRE-EPI sensitive to the BOLD effect also make it sensitive to susceptibility artifacts such as geometric distortions and signal loss. Susceptibility gradients near air/tissue interfaces can severely compromise the detection power of the BOLD analysis in certain areas of the brain, especially in the orbitofrontal cortex and temporal lobes.
The objective of this project was to develop methods that could detect areas in the brain particularly sensitive to signal loss and geometric distortions and areas where the detection power of the BOLD analysis had been compromised as a result of susceptibility differences in the brain. The effect of susceptibility induced magnetic field inhomogenities on the image quality and on the BOLD sensitivity was investigated. Magnetic field maps were generated and used as a guide to detect areas at risk for artifacts. The field maps were also used to generate simulations of the image intensity and BOLD sensitivity.
BOLD fMRI has, since its inception, been a valuable part of brain research. The methods developed in this project were tested on data collected from the FRONT project. FRONT conducted BOLD fMRI analysis on patients with a known damage to the frontal lobe and on healthy controls. Magnetic field maps and simulated BOLD sensitivity maps were generated for a sample of the subjects.
BOLD fMRI has become an increasingly valuable tool for use in presurgical planning. Data from an epilepsy patient was collected. The BOLD analysis was conducted for the purpose of locating important centers, such as the language centre, in preparation of epilepsy focus resection. The patient had several metallic clips fastened in the scull, from previous operations, susceptibility differences from the clips, in addition to differences close to air/tissue interfaces, made the detection of BOLD signal challenging. Simulated BOLD sensitivity maps were generated to help in the presurgical planning.
It was found that the BOLD sensitivity maps together image distortion maps, generated from the magnetic field maps, are useful tools for identifying problem areas in in BOLD fMRI.