Ghost peaks simulation and reduction in high-resolution echo-planar spectroscopic imaging with flyback readout

Abstract The undesirable aspect of echo-planar spectroscopic imaging (EPSI) is the occurrence of Nyquist ghost spectral lines. This phenomenon reduces the intensity of each true spectral line due to energy transfer into one or more ghost peaks. It complicates spectra and decreases the signal-to-noise ratio of the genuine peaks. This study aims to simulate ghost spectral lines for EPSI with interleaved flyback gradient echo trains and to propose methods for reducing ghost peaks. Simulations utilized the fact that the spectra reconstructed from a single flyback gradient echo train do not contain ghost spectral lines. The acquired raw data were reorganized into interleaved two echo trains before processing. The simulations aimed to demonstrate the effects of temporal shifts, magnitude, and phase variations of echoes in the interleaved second echo train on the formation of ghost peaks. Iterative zero-order phase corrections were proposed to reduce ghost peaks. Spectra obtained from a water-vegetable oil phantom and a rat demonstrate that the zero-order phase corrections significantly decreased ghost peak magnitudes while increasing the magnitudes of the true peaks. The methods described for reducing ghost peaks are suitable for water-fat spectroscopy and imaging with high spatial and spectral resolution.