Motion correction for MRI using external tracking devices

Subject motion is a major problem for Magnetic Resonance Imaging (MRI), both clinically and in research. Typical MRI scans can last several minutes, during which substantial subject movements can occur. These movements negatively affect image quality resulting in loss of data or the need for repeated scans. There are currently a large number of potential solutions to subject motion. Some of these solutions involve altering the scan to make it more resistant to subject motion, others involve tracking the subject and either retrospectively correcting data or prospectively updating the acquisition sequence to follow the subject as they move. The method by which the subject is tracked varies depending on the users needs; MR based methods such as navigators are often the simplest but may increase scan times, hardware based methods such as NMR probes or optical tracking are the most accurate methods but require additional set up times and are more expensive. In this thesis the use of a Moiré phase tracking (MPT) camera is investigated for its use in prospective motion correction. When using an external tracking camera, data is acquired in the camera’s frame of reference and must be converted to the magnet’s frame of reference (cross calibration) before it can be used for motion correction. This causes an additional source of error when determining the subject position. The impact of this error was investigated to determine the acceptable accuracy of the calibration transform. A novel method of measuring calibration stability is demonstrated to increase confidence in the calibration transform. Using the calibration transform, a software patch (Geocam) was written to enable prospective motion correction on a Philips 7 T MRI scanner. This patch was compared to existing motion correction techniques available on the scanner which correct for inter-volume motion. The new functionality (intra-volume motion correction) introduced by the Geocam patch was demonstrated with phantom and in-vivo data by performing large movements during a volume imaging sequence. The Geocam patch was then demonstrated in combination with a novel method of determining B0 maps to correct geometric distortions in EPI data. The proposed method was able to improve the registration of EPI data to an anatomical images compared to standard methods. Despite the advantages of optical tracking cameras, they still suffer from significant limitations due to the requirement of line of sight between the marker and camera. Commercial optical tracking cameras are also relatively expensive. A small electronic sensor is proposed that could potentially address these limitations. The proposed sensor performed comparably to existing methodologies in terms of overall accuracy. Future work to enable to the sensor to be used for prospective motion correction is proposed. This thesis addresses the practical implications of employing optical tracking camera technology for prospective motion correction at 7 T and proposes alternative devices which could be used