A magnetic resonance spectroscopy (MRS) procedure for in vivo measurement of lipid levels in mouse liver is described and validated. The method uses respiratory-gated, localized spectroscopy to collect proton spectra from voxels within the mouse liver. Bayesian probability theory analysis of these spectra allows the relative intensities of the lipid and water resonances within the liver to be accurately measured. All spectral data were corrected for measured spin-spin relaxation. A total of 48 mice were used in this study, including wild-type mice and two different transgenic mouse strains. Different groups of these mice were fed high-fat or low-fat diets or liquid diets with and without the addition of alcohol. Proton spectra were collected at baseline and, subsequently, Alcoholic (1, 2) and nonalcoholic fatty liver (NAFL) (3-7) are highly prevalent in human populations and may develop into steatohepatitis and in some cases into cirrhosis requiring liver transplantation. Animal models of both conditions have been developed. Mouse models are particularly useful because genetic manipulations are highly developed in inbred mice. However, longitudinal studies require the killing of animals because no noninvasive method for quantifying liver fat is available. Here, we describe a noninvasive, nondestructive method for quantifying the liver fat contents of the mouse using a NAFL mouse model. The overwhelming majority of NAFL cases are associated with obesity, dyslipidemia, hypertension, insulin-resistant type 2 diabetes mellitus, and atherosclerotic cardiovascular disease (8)(9)(10)(11)(12). This constellation defines the metabolic syndrome (13,14). One naturally occurring cause of fatty liver is familial hypobetalipoproteinemia (FHBL). FHBL is defined by less than fifth percentile plasma levels of LDL-cholesterol and/or total apolipoprotein B (apoB), segregating in families as an autosomal dominant trait (15-17). The mean liver triglyceride content in apoBimpaired FHBL subjects is 3-to 5-fold greater than that of controls (18,19).In an attempt to understand the cellular/molecular bases of hypobetalipoproteinemia, several recombinant mice mimicking human FHBL have been produced (19)(20)(21)(22)(23)(24)(25). The resulting mice closely resemble their human counterparts with respect to the fatty liver phenotype. Thus, these mice could serve as good models of one genetic form of fatty liver for studies of the progression of fatty liver and on the effects of metabolic, hormonal, and therapeutic perturbations over time.One current limitation of such studies is the absence of an accurate, quantifiable, noninvasive method for repeat assessments of liver fat in animals. In humans, the gold standard for quantifying liver fat is the liver biopsy, which is invasive and, at best, semiquantitative. The analogous procedure in animals is killing of the animal and direct analysis of liver fat by standard chemical methods after extraction. Obviously, direct sampling of livers for fat analysis is not ideal for longitudinal, fol...