Paper detail

Optimization of hyperparameters for SMS reconstruction

Simultaneous multi-slice (SMS) imaging accelerates MRI data acquisition by exciting multiple image slices simultaneously. Overlapping slices are then separated using a mathematical model. Several parameters used in SMS reconstruction impact the quality of final images. Therefore, finding an optimal set of reconstruction parameters is critical to ensure that SMS acceleration does not significantly degrade resulting image quality. Gradient-echo echo planar imaging (EPI) data were acquired with a range of SMS acceleration factors from volunteers. Images were collected using two head coils (a 48-channel array and a smaller 32-channel array). Data from these coils were reconstructed offline using a range of coil compression factors and reconstruction kernel parameters. A hybrid space (ky-x), externally-calibrated coil-by-coil slice unaliasing approach was used for image reconstruction. The resulting SMS images were compared with identical EPI data acquired without SMS. A functional MRI (fMRI) experiment was also performed and group analysis results were compared between data sets reconstructed with different coil compression levels. The 32-channel coil with smaller dimensions outperformed the larger 48-channel coil in our experiments. Generally, a large calibration region and small kernel sizes in ky direction improved image quality. Use of regularization in the kernel fitting procedure had a notable impact. With optimal selection of other hyperparameters in the hybrid space SMS unaliasing algorithm, coil compression caused small reduction in quality in SMS accelerated images. Similarly, group analysis of fMRI results did not show a significant influence of coil compression on resulting image quality. Hyperparameters used in SMS reconstruction need to be fine-tuned once the experimental factors such as the RF receive coil and SMS factor have been determined.