To compare two different methods of obtaining the water reference for determination of quantitative water-scaled in vivo concentration estimates of Î³-aminobutyric acid (GABA). Water-scaled GABA estimates from localized J-difference edited MR spectroscopy experiments can be computed using standard values for tissue-specific water content and relaxation times. Water content and relaxation may, however, be altered in pathology. This work re-analyzed data from a recent study in healthy controls and patients with minimal (mHE) or grade I (HE 1) hepatic encephalopathy, a disease associated with slight elevation of brain water content. J-difference edited MR spectroscopy data were combined with quantitative brain water measures, which provided individual water density references and T1 relaxation times. Resulting GABA estimates were compared to concentration values obtained using standard tissue-specific water content and relaxation values. Occipital GABA concentration values obtained from individual water and T1 maps were 1.64Â±0.35mM in controls, and significantly higher (P<0.01) than in mHE (1.15Â±0.28mM) and HE 1 patients (1.18Â±0.09mM). Results from the tissue-dependent approach (1.58Â±0.30mM (controls), 1.10Â±0.27mM (mHE) and 1.12Â±0.12mM (HE 1)) were slightly lower (P<0.05 in each group). Water-scaled in vivo GABA estimates can be obtained with individual water density and T1 relaxation mapping. This approach may be useful for studying GABA levels in pathologies with substantial brain water content or relaxation changes.