Uncertainty assessment of gamma-aminobutyric acid concentration of different brain regions in individual and group using residual bootstrap analysis

Abstract

The aim of this work is to quantify individual and regional differences in the relative concentration of gamma-aminobutyric acid (GABA) in human brain with in vivo magnetic resonance spectroscopy. Spectral editing Mescher–Garwood point resolved spectroscopy (MEGA-PRESS) sequence and GABA analysis toolkit (Gannet) were used to detect and quantify GABA in anterior cingulate cortex (ACC) and occipital cortex (OCC) of healthy volunteers. Residual bootstrap, a model-based statistical analysis technique, was applied to resample the fitting residuals of GABA from the Gaussian fitting model (referred to as GABA₊ thereafter) in both individual and group data of ACC and OCC. The inter-subject coefficient of variation (CV) of GABA₊ in OCC (20.66 %) and ACC (12.55 %) with residual bootstrap was lower than that of a standard Gaussian model analysis (21.58 % and 16.73 % for OCC and ACC, respectively). The intra-subject uncertainty and CV of OCC were lower than that of ACC in both analyses. The residual bootstrap analysis thus provides a more robust uncertainty estimation of individual and group GABA₊ detection in different brain regions, which may be useful in our understanding of GABA biochemistry in brain and its use for the diagnosis of related neuropsychiatric diseases.

http://ift.tt/2dtPT5k