Recently introduced model-based deep learning (MoDL) technique successfully incorporates convolutional neural network (CNN)-based regularizers into physics-based parallel imaging reconstruction using a small number of network parameters. We propose wave-encoded MoDL (wave-MoDL) combining the wave-encoding strategy with unrolled network constraints for highly accelerated 3D imaging while enforcing data consistency. My research interest is to develop rapid MR acquisition and high-fidelity physic-guided deep-learning based image reconstruction which may facilitate clinical and neuroscientific applications.

Quantitative MRI (qMRI) techniques aim to provide quantitative estimates of tissues’ relaxation parameters. These methods use Bloch simulations of the acquired signals to calculate T1, T2, T2*, or proton density (PD) maps in quantitative units. In clinical settings, qMRI can help identify physiological changes undetected by qualitative imaging, provide specific information to characterize pathologies, help assess treatment response and repair processes, and detect disease before morphological changes. High-resolution qMRI is often less suitable for a clinical MR exam due to the long acquisition times. My research interest is accelerating the acquisition while preserving the accuracy of the quantitative parameter maps.

Echo-planar imaging (EPI) is widely used for diffusion MRI due to its fast acquisition. Multi-shot EPI (msEPI) can mitigate distortion and T2/T2*-related voxel blurring and voxel pile-ups by increasing acceleration and reducing echo-spacing time significantly. However, msEPI suffers from magnetic field-related image distortion in the phase-encoding direction, the phase variation between multi shots, and motion-related artifacts. My research interest is to develop an efficient, motion-robust, distortion-free, and high-resolution msEPI acquisition method.