To optimize the MEGA-PRESS pulse sequence for the simultaneous detection of glutathione and lactate.In brain MR spectroscopy (MRS), spectral editing is often used to selectively detect signals from lower concentration metabolites that overlap with those of other, more abundant compounds. Usually, one compound is observed at a time, however there have been examples of editing sequences designed to simultaneously detect 2 compounds which can allow for shorter scan times [1,2]. An optimization of the ‘MEGA-PRESS’ editing sequence to simultaneously detect glutathione (GSH, a major anti-oxidant) and lactate (Lac, an indicator of non-oxidative glycolysis) in a single acquisition is presented. The impairment of the function of both metabolites has been implicated in the pathophysiology of various brain pathologies such as schizophrenia and bipolar disorder [3,4,5,6].J-difference editing involves the acquisition of two slightly different experiments, labeled ON and OFF, which are subtracted to resolve a signal of interest from overlying signals. In the ON acquisition, frequency-selective editing pulses are applied to refocus the evolution of a coupling within the molecule of interest. In the OFF acquisition, these editing pulses are not applied and the coupling is allowed to evolve. Subtracting the two gives a difference spectrum, which only contains those signals that are affected by the editing pulses. Typically, these editing pulses are applied on resonance: 4.56 ppm for GSH and 4.1 ppm for Lac. In this implementation of J-difference editing, however, a less-selective editing pulse is applied at 4.35 ppm in the ON acquisition (between the two resonant frequencies), to detect both with near maximal sensitivity in a single acquisition.Phantom experiments and numerical simulations of PRESS and MEGA-PRESS were performed to determine the optimum TE and editing pulse frequencies for simultaneous GSH and Lac editing. Phantom spectra were acquired in a phantom containing 50 mM GSH and 25 mM sodium lactate in phosphate-buffered saline (pH = 7.2) on a Philips Achieva 3T scanner using a 32-channel receive head coil. In the PRESS and MEGA-PRESS experiments, the optimal echo time (TE) was found by varying the echo time from 70 ms to 240 ms in 10 ms increments (TE1 = 13.4 ms, TE2 = TE – TE1). Frequency-modulated slice selective refocusing pulses (‘fm_ref07’, 2.2 kHz bandwidth) were used in order to minimize chemical shift displacement effects. At an echo time of 140 ms, the optimal editing pulse frequency in the MEGA-PRESS experiment was found by varying the editing pulse frequency of the sinc-Gaussian editing pulses from 4.1 to 4.56 ppm in increments of 0.05 ppm for both 20 ms (75 Hz bandwidth) and 10 ms pulses (150 Hz) (‘off’ pulse at 10 ppm). Spatially resolved simulations of the MEGA-PRESS experiment for all positions on a 19x19 two-dimensional array across a (3 cm)3 voxel were performed in MATLAB, using the chemical shifts and coupling constants given by (3) and the following parameters: 2-kHz spectral width, 2048 points, 3-Hertz exponential filter and zero-filling to 8192 datapoints. Three MEGA-PRESS implementations were performed in vivo in one healthy individual with a 4x4x3 cm3 voxel placed in the midline parietal region: one with 75 Hz editing pulses at 4.1 ppm (on resonance for Lac), one with 75 Hz editing pulses at 4.56 ppm (on resonance for GSH), and one with 150 Hz pulses at 4.35 ppm.Figure 1a shows that in both phantom experiments and simulations, the maximum signal intensity for both GSH and Lac in the PRESS experiments occurs at an echo time of 150 – 160 ms. This optimal echo time holds true for MEGA-PRESS GSH phantom experiments and simulations as well as for MEGA-PRESS Lac simulations. Figure 2 shows that the optimal editing frequency to detect GSH and Lac equally occurs at about 4.35 ppm for both simulations and phantom experiments. GSH/Lac editing efficiency is 65% with 75 Hz editing pulses and is increased to 90% with 150 Hz editing pulses. Figure 3 shows the in vivo spectra from one subject showing good detection of both GSH and Lac comparable to the two separate Lac and GSH acquisitions with editing pulses placed on resonance.Simultaneous editing of Lac and GSH is possible at 3T. Compared to sequential measurements, simultaneous editing results in a 50% reduction in the scan time while retaining nearly the same sensitivity. Note that the optimization performed here did not consider T2 relaxation effects in vivo, which will tend to shorten the optimal TE. Detection of both metabolites can be improved by using an editing pulse with a squarer profile to perfectly invert the GSH and Lac spins.