Image 3




Mohammad Abdishektaei


Personal Homepage: NA
E-mail: ma4bx{at}virginia{dot}edu
Phone: NA
Supervisor: Dr. Abbas Nasiraei Moghaddam
Dr. Arsh Amini




M.Sc. Biomedical Engineering
Amirkabir University of Technology (Tehran Polytechnic), Tehran, Iran
2012 - 2015
B.Sc. Electrical Engineering
Amirkabir University of Technology (Tehran Polytechnic), Tehran, Iran
2010 - 2014
B.Sc. Biomedical Engineering
Amirkabir University of Technology (Tehran Polytechnic), Tehran, Iran
2008 - 2012
Ph.D Student of Biomedical Engineering
University of Virginia
since 2017 


Research Subject

The scope of my research in AMIRLab was to explore the application of Compressed Sensing in Magnetic Resonance Spectroscopic Imaging (MRSI). Spatial encoding in MRSI is only done through phase-encoding of each voxel at a time leading to a longer acquisition times. Longer acquisition times make the MRSI trials more prone to motion artifacts as well as increasing the cost trials in clinical settings. This problem is more pronounced in 2D spectral MRSI where additional phase-encoding dimension is introduced to delineate the chemical shift resonances due to J-coupling effect from the ones due to nuclear interactions. In order to get around this problem, Compressed Sensing (CS) can decrease the acquisition times by randomly skipping all the necessary phase-encoding k-space samples. To accommodate the sparse nature of the the MRS spectra, the L1-norm of the each spectra can be taken into account in the CS reconstructions. In parallel, sparse model-based reconstructions was also investigated where input data are still the undersampled k-space data.



In vivo 2D and 1D spectral Magnetic Resonance Spectroscopic Imaging (MRSI) is a powerful technique for evaluation of living tissue through exploiting the spatial distribution of biomarkers. In some pathological situations, there are changes in concentration of biomarkers. The distribution and degree of these changes can be measured by MRSI in clinic. Successful application of MRSI to not only evaluate pathologies in brain, breast and prostate but also in psychiatric and neurological disorders has been reported previously. Despite its great potential, 1D and 2D spectral MRSI still suffers from long acquisition times that prevent its widespread utilization in clinic. Compressed Sensing is a new framework that proposes novel sampling scheme based on information rate instead of conventional Nyquist rate. CS exploits the spars nature of data in some known transform domain to remove incoherent artifacts induced by random subsampling. Its application in Magnetic Resonance community is well established in previous studies. In this study, two distinct sampling-reconstruction methods have been proposed in order to accelerate acquisition times in 1D and 2D spectral MRSI.



  • Abdi, M.; Raschke, F.; Howe, F.; Nasiraei Moghaddam, A, "Accurate Compressive Sensing of 1H MR Spectroscopic Imaging in Brain Tumors", International Society for Magnetic Resonance in Medicine (ISMRM), 2015.
  • Abdi, M.; Nasiraei Moghaddam, A; Nagarajan, R; Thomas, M. A., "Sparse Reconstruction in Localized Correlated Spectroscopy: From Subsampled Priors to Fast Acquisition", International Society for Magnetic Resonance in Medicine (ISMRM), 2015.





Advanced Medical Imaging Research Laboratoy
Department of Biomedical Engineering
Amirkabir University of Technology
424 Hafez Avenue, Tehran
Tehran, IRAN, P,O.BOX: 15875-4413