Correction of aliasing artifact in accelerated echo-planar imaging.

PURPOSE: To develop a comprehensive reconstruction pipeline that simultaneously addresses 2D Nyquist and aliasing artifacts in echo-planar imaging (EPI) data acquired using various schemes, including single-shot, multi-shot, parallel, and multi-band EPI. METHODS: We introduce a novel 2D Nyquist artifact correction method that extends our previously reported phase-search reconstruction approach. A series of phase-corrected images are generated using a range of candidate phase values, and the corresponding coil sensitivity profiles are compared with known profiles to estimate an optimal 2D Nyquist phase correction map. In addition, we propose an integrated reconstruction procedure that corrects aliasing artifacts arising from 2D Nyquist effects, shot-to-shot motion-induced phase variations, and both in-plane and through-plane acceleration schemes. The proposed methods were evaluated using resting-state fMRI data from 30 healthy volunteers. RESULTS: The proposed method substantially reduced residual artifacts in EPI data, as measured by the ghost-to-signal ratio. The resulting default-mode network maps showed improved correspondence with known reference networks compared to those obtained using conventional 1D Nyquist artifact correction methods. CONCLUSION: The developed reconstruction pipeline effectively corrects multiple sources of aliasing artifacts in EPI data, offering improved image quality and functional sensitivity across a wide range of EPI acquisition schemes.