Accelerating diffusion tensor imaging using multi-reference image constrained reconstruction

Introduction: MR diffusion tensor imaging (DTI) [1] has been increasingly used to quantitatively characterize the microstructure of ordered tissues such as the brain white matter. However, practical applications of DTI are hampered by the low SNR, long acquisition times, and low spatial resolution. Since in DTI essentially the same image (except for diffusion encoding) is acquired repeatedly, reduced encoding and single-reference constrained reconstruction have been shown to improve the DTI scan efficiency [2]. Separately, more accurate capturing of the image contrast has been achieved by constrained reconstruction using two reference images [3]. The goals of the current study are to extend multi-reference constrained reconstruction to DTI, and evaluate whether the methodology can further improve the scan time efficiency of DTI.