Motion during magnetic resonance imaging (MRI) is estimated to cost hospitals approximately $115,000 per scanner per year, which implies that the annual cost in the US is over $1 billion. Head movement during MRI affects clinical diagnosis, especially in the sickest patients and children. In brain imaging research, high resolution structural scans are used, and participants tend to move during these long acquisitions. Moreover, participants in patient groups may move systematically more than healthy controls, and this may introduce bias in the data with possibly erroneous conclusions from the research. In this project, we will address the problem of head motion MR neuroimaging, and validate the technology in routine clinical exams. The ideal motion tracking and correction system would require no external devices, operate at high temporal frequency to enable tracking of rapid motion, leave the contrast of the MRI sequence unchanged, and function properly in a broad array of MRI acquisition types. Unfortunately, state-of-the-art prospective motion correction requires an external device, so the impact of high-quality motion correction is limited. In this project we propose an array of innovative technical enhancements for ?navigator? methods that use the intrinsic motion information in the MR signal. The developments will result in a flexible, widely applicable high-frequency prospective motion correction (PMC) that radically reduces MRI motion artifacts. In order to achieve this, we will use ?cloverleaf? navigators (CLN), which have been shown to provide high- frequency motion information but unfortunately only in a limited set of ?3D steady-state? sequences. To enhance the flexibility of CLN we will use the MR signal from subcutaneous fat so that motion can be measured rapidly in non-steady-state and multi-slice sequences without affecting the water signal of interest. Furthermore, CLN will be enhanced using recent advances in coil-space motion detection. During development, an external camera will be used to evaluate motion measurement and PMC performance. The PMC-enabled 3D GRE, MPRAGE, and Fast-Spin-Echo sequences will be validated in clinical brain MRI exams. Such methods for sequence-universal, high-frequency prospective motion correction without any external camera equipment could be extended to other sequences, and would substantially broaden the impact of motion-robust brain MRI and reduce the financial burden for hospitals and research institutes world-wide.
MRI is a powerful and flexible imaging technology but head movement during MRI brain scans severely degrades image quality, necessitating repeated scans and sometimes the patient has to be sedated or anesthetized. Anesthesia is dangerous and carries risks to the developing brain. Poor image quality may lead to incorrect or missed diagnosis. In this project, we will develop new MRI scan software that can be used in existing MRI machines around the world to compensate for head motion in real-time, so that all of the clinical images required for diagnosis have high quality even if patients move their heads during the scans.
