Prospective motion correction with data reacquisition based on fat navigators for T1-weighted brain MRI at 3T

T1-weighted (T1w) sequence is widely employed in 3T clinical and research MRI. However, its image quality is vulnerable to motion artifacts. We aim to develop a robust prospective motion correction (PMC) method for T1w sequence based on fat navigators (FatNav). Since residual artifacts are often present after PMC, we also evaluated the effectiveness of data reacquisition in reducing such artifacts. Twenty healthy adults and six pediatric participants were scanned with PMC on and off while performing controlled or random head motion. To achieve PMC, the field of view was adjusted in real time based on motion parameters derived from fat navigators. Data acquired between two neighboring navigators that had relative motion score above a certain threshold were reacquired. Image quality was assessed using tissue boundary blurriness, SSIM, PSNR, and cortical and white matter volumes. For images acquired with controlled head movement, PMC with reacquisition significantly increased visual assessment scores from 1.42 ± 0.69 to 3.67 ± 0.76, indicating reduced visible artifacts, and significantly reduced blurriness (p = 0.001) with a mean reduction of 18.6% compared with PMC-off images. In parallel, cortical volume, SSIM, and PSNR significantly increased after PMC (p = 0.012, p = 0.027, and p = 0.043, respectively). Blurriness and cortical volume were fitted against the number of reacquired TRs using a linear model, which suggested an approximately linear trend under the tested conditions. However, the slope of the linear fit was significantly different (p = 0.019 and p = 0.030, respectively) from zero only when the motion scores of the reacquired TR were above a certain threshold. For images acquired with random head motion, visible artifact levels were also reduced with PMC and reacquisition compared to those with PMC off. PMC based on FatNav effectively reduces the impact of motion for T1-weighted MRI at 3 T. Reacquiring data with motion scores above a certain threshold can further improve image quality.