Intra- and inter-session reliability and repeatability of ¹H magnetic resonance spectroscopy for determining total creatine concentrations in multiple brain regions

Using proton magnetic resonance spectroscopy (¹H MRS) to determine total creatine (tCr) concentrations will become increasingly prevalent, as the role of creatine (Cr) in supporting brain health gains interest. Methodological limitations and margins of error in repeated ¹H MRS, which often surpass reported effects of supplementation, permeate existing literature. We examined the intra- and inter-session reliability and repeatability of ¹H MRS for determining tCr concentrations across multiple brain regions (midbrain, visual cortex and frontal cortex). Eighteen healthy adults aged 20-32 years were recruited (50% female; n = 14 intra-session; n = 15 inter-session). ¹H Magnetic resonance imaging and spectroscopy were completed at 3 T. Intra-session analyses involved repeated ¹H MRS of the midbrain, visual cortex and frontal cortex without participant or voxel repositioning, whereas inter-session analyses involved measurements of the same regions, but with participant and voxel repositioning between repeated measurements. The ¹H MRS data (174 spectra) were analysed using TARQUIN and OSPREY, and voxel fractions (grey/white matter and CSF) were determined using segmentation. Our findings show that tCr concentrations can be determined reliably and repeatably using ¹H MRS, within an error of <2%, and that large inter-regional differences in tCr concentration are present in the human brain. We provide new minimum detectable change data for tCr concentrations, a detailed discussion of the inherent error sources in repeated ¹H MRS, including the substantial effect of the analysis package on tCr quantification, and suggestions for how these should be managed to improve the interpretability and clinical value of future research. More studies are needed to determine whether our findings can be replicated in other centres and different populations.