Real Time Motion correction for Brain MRI

大脑 MRI 的实时运动校正

• 批准号: 7575217
• 负责人: Truman R Brown
• 金额: $ 5.38万
• 依托单位: COLUMBIA UNIVERSITY HEALTH SCIENCES
• 依托单位国家: 美国
• 财政年份: 2008
• 资助国家: 美国
• 起止时间: 2008-03-01 至 2010-10-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/7575217
• 关键词: Address; Algorithms; Area; Blood; Brain; Cerebrovascular Circulation; Child; Collection; Computer software; Data; Data Set; Dementia; Detection; Development; Devices; Diagnosis; Disease; Effectiveness; Elderly; Epilepsy; Evaluation; Frequencies; Functional Magnetic Resonance Imaging; Head; Head Movements; Height; Human; Human Volunteers; Image; Inferior; Injection of therapeutic agent; Left; Magnetic Resonance Imaging; Measurement; Measures; Methods; Metric; Monitor; Morphologic artifacts; Motion; Motor; Movement; Neurons; Parkinson Disease; Patients; Physiologic pulse; Population; Positioning Attribute; RF coil; Recording of previous events; Relaxation; Research; Rotation; Scanning; Scheme; Schizophrenia; Series; Signal Transduction; Slice; Solid; Speed; Spin Labels; Subject Headings; Techniques; Testing; Time; Tissues; Translations; Weight; Width; base; data acquisition; deoxyhemoglobin; improved; millimeter; motor control; novel; older patient; patient population; prevent; programs; prospective; relating to nervous system; response; success; tool; visual stimulus; volunteer

DESCRIPTION (provided by applicant): Magnetic resonance imaging (MRI) of the head and brain is a powerful tool for research and diagnosis. During a MRI scan patients are asked to keep their head very still because slight movements can spoil the MRI data, but this can be difficult for young children, elderly people, and those who suffer from Parkinson's disease, schizophrenia, epilepsy, and dementia. Our research will let MRI better serve these patients by allowing accurate data to be collected even when head movements occur during scanning. The standard approach to correct motion artifacts in MRI is retrospective image-based motion detection and correction as implemented in popular analysis packages such as SPM and AIR. This approach is well suited to motion within the imaging plane, but cannot handle substantial through-plane motion which both cannot be described by a single rotation/translation and alters the spin magnetization history of the tissue in the imaging field of view (FOV). Prospective motion correction techniques which measure head position in real time and adjust the FOV prior to data acquisition thus offer a compelling advantage for through-plane motion. However, existing 'prospective techniques such as navigator echoes and PACE impose a delay in data acquisition rates. Our objective is to implement and validate a novel scheme for prospective correction of MRI motion artifact that operates in parallel with the acquisition of imaging data, preventing temporal delay. We have developed a tracking device for real-time monitoring of three- dimensional changes in head position using three RF tracking coils for spatial localization simultaneous with image data acquisition via a standard head coil. Our first specific aim is to implement dynamic motion detection using our tracking device and prospective re-alignment of the imaging plane on a Philips Achieva scanner. Our second specific aim is to create realistic motion artifacts in MRI data from phantoms. Four metrics will be used to evaluate the success of the correction algorithm. Our third specific will study twelve volunteers who have been instructed to turn their heads to track a moving visual stimulus. The metrics used to evaluate the algorithm consist of a) comparison with standard retrospective motion correction using AIR, b) evaluation of the high spatial frequencies present in the images collected with and without the motion correction scheme, c) comparison of line profiles through the images of corrected vs. uncorrected images and d) measurement of the width at = height of small cylinders in one of the phantoms. We expect that our scheme will be better able to address the degree of motion typically seen in patient populations.

描述（由申请人提供）：头部和大脑的磁共振成像（MRI）是研究和诊断的有力工具。在 MRI 扫描过程中，患者被要求保持头部静止，因为轻微的运动可能会破坏 MRI 数据，但这对于幼儿、老年人以及帕金森病、精神分裂症、癫痫和痴呆症患者来说可能很困难。我们的研究将使 MRI 更好地服务于这些患者，即使在扫描过程中发生头部运动时也能收集准确的数据。纠正 MRI 中运动伪影的标准方法是基于回顾性图像的运动检测和纠正，如 SPM 和 AIR 等流行分析包中所实施的那样。这种方法非常适合成像平面内的运动，但不能处理大量的平面运动，这种运动既不能通过单个旋转/平移来描述，又会改变成像视场 (FOV) 中组织的自旋磁化历史。前瞻性运动校正技术可以实时测量头部位置并在数据采集之前调整视场，从而为平面运动提供了引人注目的优势。然而，现有的“前瞻性技术”（例如导航回波和 PACE）会延迟数据采集速率。我们的目标是实施和验证一种新颖的方案，用于 MRI 运动伪影的前瞻性校正，该方案与成像数据的采集并行操作，防止时间延迟。我们开发了一种跟踪设备，用于实时监控头部位置的三维变化，使用三个射频跟踪线圈进行空间定位，同时通过标准头部线圈采集图像数据。我们的第一个具体目标是使用我们的跟踪设备和飞利浦 Achieva 扫描仪上的成像平面的前瞻性重新对齐来实现动态运动检测。我们的第二个具体目标是在来自体模的 MRI 数据中创建逼真的运动伪影。将使用四个指标来评估校正算法的成功。我们的第三个具体项目将研究 12 名志愿者，他们被指示转动头部来追踪移动的视觉刺激。用于评估算法的指标包括：a) 与使用 AIR 的标准回顾性运动校正进行比较，b) 对使用和不使用运动校正方案收集的图像中存在的高空间频率进行评估，c) 图像中线条轮廓的比较校正图像与未校正图像的比较，以及 d) 测量模型之一中小圆柱体的宽度 = 高度。我们期望我们的方案能够更好地解决患者群体中常见的运动程度。