Trina Kok1,
Navin Michael2, Sendhil S. Velan, 13
1A*STAR-NUS
Clinical Imaging Research Center, Singapore, Singapore; 2Singapore
Institute for Clinical Sciences, Singapore, Singapore; 3Singapore
Bioimaging Consortium, Singapore, Singapore
Detection of metabolite signals are often hampered by spectral overlap arising from J-coupling effects. Existing techniques for improved separation include 2D spectroscopy methods such as 2DPRESS, which carries a penalty in scan-time and is often prohibitive when combined with spectroscopic imaging. In this work, we apply compressed sensing to a randomly under-sampled data set, collecting 30 t1 increments of the original 100 t1 increments. The reconstructed 2D spectrum is subsequently fitted with ProFit and we show that metabolites were recovered with reasonable CRLB values, with 30% of the original data.
Keywords