Quantification of glutathione transverse relaxation time T2 using echo time extension with variable refocusing selectivity and symmetry in the human brain at 7 Tesla

[Display omitted] •Glutathione (GSH) T2 was measured in the human occipital cortex at 7 Tesla.•Variable TE extension pulse selectivity and symmetry maximized GSH quantifiability.•Signals were quantified by linear combination model fits of simulated basis sets.•T2 of GSH was calculated to be 145.0 ± 20.1 ms in nine healthy adults.•T2 correction stabilized decreases in apparent GSH concentration at high TE. Glutathione (GSH) is an endogenous antioxidant implicated in numerous biological processes, including those associated with multiple sclerosis, aging, and cancer. Spectral editing techniques have greatly facilitated the acquisition of glutathione signal in living humans via proton magnetic resonance spectroscopy, but signal quantification at 7 Tesla is still hampered by uncertainty about the glutathione transverse decay rate T2 relative to those of commonly employed quantitative references like N-acetyl aspartate (NAA), total creatine, or water. While the T2 of uncoupled singlets can be derived in a straightforward manner from exponential signal decay as a function of echo time, similar estimation of signal decay in GSH is complicated by a spin system that involves both weak and strong J-couplings as well as resonances that overlap those of several other metabolites and macromolecules. Here, we extend a previously published method for quantifying the T2 of GABA, a weakly coupled system, to quantify T2 of the strongly coupled spin system glutathione in the human brain at 7 Tesla. Using full density matrix simulation of glutathione signal behavior, we selected an array of eight optimized echo times between 72 and 322 ms for glutathione signal acquisition by J-difference editing (JDE). We varied the selectivity and symmetry parameters of the inversion pulses used for echo time extension to further optimize the intensity, simplicity, and distinctiveness of glutathione signals at chosen echo times. Pairs of selective adiabatic inversion pulses replaced nonselective pulses at three extended echo times, and symmetry of the time intervals between the two extension pulses was adjusted at one extended echo time to compensate for J-modulation, thereby resulting in appreciable signal-to-noise ratio and quantifiable signal shapes at all measured points. Glutathione signal across all echo times fit smooth monoexponential curves over ten scans of occipital cortex voxels in nine subjects. The T2 of glutathione was calculated to be 145.0 ± 20.1 ms (mean ± standard deviati