Adaptive Motion Correction in Functional Magnetic Resonance Imaging of Brain Activity

功能磁共振成像（fMRI）是一种生成大脑活动图像的方法。自20世纪90年代被发现以来，功能磁共振成像已经成为研究人类思维的神经科学研究的关键。医学功能磁共振成像的应用也在发展，比如在神经外科，神经外科医生在计划切除脑肿瘤时使用功能磁共振成像，以尽可能多地保留周围的正常组织。

Functional magnetic resonance imaging (fMRI) is method for generating pictures of activity within the brain. Since its discovery in the 1990s, fMRI has become essential in neuroscience research to investigate the human mind. Medical fMRI applications are also developing, such as in neurosurgery, where neurosurgeons use fMRI when planning to remove brain tumours, to spare as much normal surrounding tissue as possible. Unfortunately, fMRI signals of brain activity are very small and are easily corrupted if the patient does not keep their head still. A typical fMRI exam requires approximately 5 to 10 minutes for collection of a series of images while the patient performs a task (for example, finger movements) designed to activate specific brain regions. Patients must not move their head more than about a millimetre during this process. Although a variety of established fMRI methods reduce the head motion problem, such as use of light restraints, fast imaging (like time-lapse photography, which "freezes" motion without blurring), and computer processing methods that can rotate and translate images back into alignment, collectively these methods are simply not effective enough. Recently, studies have started to show the potential of "adaptive fMRI" for substantially reducing the head motion problem. Adaptive fMRI uses a position tracking system to measure head motion during imaging, and to adjust the scanning position and orientation continuously in compensation. Over the next 5 years, research is proposed that involves using an optical position tracking system to develop adaptive fMRI as a robust method. Specific goals will be to perform motion correction rapidly for individual image slices, while ensuring that other signal errors arising from the interaction between head motion and magnetic field uniformity, and between head motion and signal receiver hardware are also compensated. The newly developed methods will be validated in moving test objects and in representative fMRI exams of healthy adults and patients, and will enable practical fMRI examinations of much broader patient populations in the future.