PURPOSE: To detect motion-corrupted measurements in multi-average turbo-spin-echo (TSE) acquisitions and reduce motion artifacts in reconstructed images. MATERIALS AND METHODS: An average-specific phase encoding (PE) ordering scheme was developed for multi-average TSE sequences in which each echo train is assigned a unique PE pattern for each pre-averaged image (PAI). A motion detection algorithm is developed based on this new PE ordering to identify which echo trains in which PAIs are motion-corrupted. The detected PE views are discarded and replaced by uncorrupted k-space data of the nearest PAI. Both phantom and human studies were performed to investigate the effectiveness of motion artifact reduction using the proposed method. RESULTS: Motion-corrupted echo trains were successfully detected in all phantom and human experiments. Significant motion artifact suppression has been achieved for most studies. The residual artifacts in the reconstructed images are mainly caused by residual inconsistencies that remain after the corrupted k-space data is corrected. CONCLUSION: The proposed method combines a novel data acquisition scheme, a robust motion detection algorithm, and a simple motion correction algorithm. It is effective in reducing motion artifacts for images corrupted by either bulk motion or local motion that occasionally happens during data acquisition. (c) 2007 Wiley-Liss, Inc.
PURPOSE: To detect motion-corrupted measurements in multi-average turbo-spin-echo (TSE) acquisitions and reduce motion artifacts in reconstructed images. MATERIALS AND METHODS: An average-specific phase encoding (PE) ordering scheme was developed for multi-average TSE sequences in which each echo train is assigned a unique PE pattern for each pre-averaged image (PAI). A motion detection algorithm is developed based on this new PE ordering to identify which echo trains in which PAIs are motion-corrupted. The detected PE views are discarded and replaced by uncorrupted k-space data of the nearest PAI. Both phantom and human studies were performed to investigate the effectiveness of motion artifact reduction using the proposed method. RESULTS: Motion-corrupted echo trains were successfully detected in all phantom and human experiments. Significant motion artifact suppression has been achieved for most studies. The residual artifacts in the reconstructed images are mainly caused by residual inconsistencies that remain after the corrupted k-space data is corrected. CONCLUSION: The proposed method combines a novel data acquisition scheme, a robust motion detection algorithm, and a simple motion correction algorithm. It is effective in reducing motion artifacts for images corrupted by either bulk motion or local motion that occasionally happens during data acquisition. (c) 2007 Wiley-Liss, Inc.