Manoj K. Sarma1,
Rajakumar Nagarajan1, Jonathan Furuyama1, Jenny Li1,
Paul M. Macey2, Rajesh Kumar3, M. Albert Thomas1
1Radiological
Sciences, UCLA School of Medicine, Los Angeles, CA, United States; 2School
of Nursing, UCLA School of Medicine, Los Angeles, CA, United States; 3Neurobiology,
UCLA School of Medicine, Los Angeles, CA, United States
Compressed Sensing (CS) has revolutionized medical imaging with optimal encoding/reconstruction schemes for MRI and hyperpolarized MRSI. There have been no reports on combining 2 spectral dimensions with 2D or 3D spatial encoding applicable to human brain pathologies so far and the CS-based approaches are well-suited for MRSI combining 2 spectral and 2 spatial dimensions. We have evaluated a novel four dimensional (4D) echo-planar J-resolved spectroscopic imaging (EP-JRESI) sequence using non-uniform under-sampling (NUS) approaches for acceleration and CS for reconstruction in healthy human brain in vivo. The sequence has been implemented on a 3T MRI scanner equipped with a 12-channel head coil. Eight healthy volunteers and one patient with obstructive sleep apnea (OSA) have been evaluated. The 2D J-resolved spectra extracted from the CS-reconstructed 4D EP-JRESI data were processed using the prior-knowledge fitting (ProFit) algorithm to quantify cerebral metabolites in hippocampus and other locations.