Filters applied:
Did You Mean? ?
one result

 

  • Book
    Thomas Sushil John.
    Digital2011
    Magnetic resonance imaging (MRI) is a versatile imaging tool that can be used to detect changes in brain hemodynamics that accompany neuronal activity. This ability to observe both anatomy and the structures that participate in specific functions is often referred to as functional magnetic resonance imaging (fMRI). Echo planar imaging (EPI) is the most widely used data acquisition method for fMRI. Functional images are often distorted because EPI is highly sensitive to field inhomogeneities, eddy currents, and gradient delays. These distortions often lead to image misregistration, and this is an impediment to correctly interpreting functional studies, and to neurosurgical planning. Many methods have been developed to address this. However, none yields anatomical and functional images that may be directly overlaid onto each other with no post-processing. This dissertation demonstrates that automatic, inherent self co-registration is achieved when optimized flyback EPI with echo time shifting (ETS) is used for both functional and anatomical images. fMRI experiments validate the proposed method, and the result is high-quality functional and high-resolution anatomical images that are inherently self co-registered. EPI acquisitions are also prone to timing and phase errors that give rise to image ghosting. ETS is an effective ghost correction method in EPI. Although ETS corrects for ghosting in a robust, non-iterative, and automatic manner, it does so at the expense of increasing total scan time. This dissertation proposes a simple, yet effective scheme to increase the efficiency of ETS by acquiring partial echoes during the idle wait times present in ETS. Using the proposed technique, higher resolution images are acquired when total scan time is fixed. Alternately, shorter scan times are possible when the in-plane resolution is fixed. Equations that predict the expected improvement in resolution are derived and shown to be accurate through simulations and experiments. In vivo images obtained using conventional ETS and the proposed method show that the time savings obtained with the proposed method do not result in any visible degradation in image quality. Signal changes caused by cardiac pulsation and respiratory fluctuation are also a major confound in fMRI. These time-locked signal changes, also referred to as "physiological noise, " decrease the sensitivity of fMRI to neuronally-induced signal fluctuations. To date, no method exists that compensates for these detrimental effects in 3D fMRI acquisitions. This dissertation presents a novel algorithm that compensates for physiological noise in 3D echo-planar fMRI. The proposed method utilizes composite property mapping and efficient root-solving to estimate physiologically-induced signal fluctuations in 3D k-space data. These signal fluctuations are subsequently regressed from k-space data. Results show that the estimated signal fluctuations have significant power in the cardiac and respiratory frequency bands. Further, resting-state fMRI experiments show an increase in the number of activated voxels when the proposed compensation algorithm is employed. This is accompanied by an increase in the average z-score of activated voxels, and a more Gaussian distributed voxel time-course. In conclusion, this dissertation describes advances in data acquisition and post- processing in 3D fMRI that are easy to implement. Further, promising experimental results suggest the efficacy of all the proposed methods.