hide
Free keywords:
-
Abstract:
Purpose/Introduction: Motion artifacts are a major problem for functional and anatomical MRI. The state of the art in head motion correction is prospective motion correction using a tracking modality of choice while updating slice positions and orientations in real-time during the acquisition.1 This work explores the possibility of using
simultaneous tracking with multiple modalities. They can be used to supplement each other or provide an alternative when the tracking update from one method is lost or erroneous.
Subjects and Methods: Two tracking methods were used to track and prospectively correct2 for head motion simultaneously: Optical tracking with Moire´ Phase Tracking (MPT)3 (Kineticor Inc, HI, USA) and motion tracking using four 19F NMR field probes4,5,6,7. The MR scanner used in this experiment was a 9.4 T human scanner (Siemens, Erlangen, Germany). The subject was scanned with a gradient echo sequence (Resolution 0.8 9 0.8 9 1.6 mm, TR 80 ms, TE 4 ms, FA 20). The MPT marker was attached to a subject specific bite-bar in order to have line of sight to the camera inside a shielded coil. The field probes (FP) were
attached to the nose bridge and the temples of the subject. To simulate tracking dropouts of the MPT system, the subject was asked to obscure the marker for short periods of time. To still achieve continuous motion correction the tracking source was switched to field probes in those time intervals.
Results: The in vivo measurements in Fig. 1 show a reference image (a) and the different types of prospective motion correction with small head motion (s. also Fig. 2). The image quality for single modality tracking (b,c) is comparable to the reference case for both modalities.
With induced tracking dropouts the quality is visibly reduced (d) but can be improved again when field probe tracking is enabled as a fallback method (e). The motion trajectories measured with both systems for the two
measurements with tracking dropouts are shown in Fig. 2. Motion range and pattern are very similar in both measurements.
Discussion/Conclusion: The quality of the prospectively corrected images is improved when the fallback is enabled compared to the case when there are tracking dropouts and only one system is used. However, the remaining difference to the images with single source correction has to be investigated further. Further applications might include averaging of the motion estimates of multiple tracking systems or using the other tracking source to
cross-validate the plausibility of measured motion.