Magnetic resonance imaging (MRI) is an imaging technology with the advantages of non-invasive scanning and high resolution in the field of medical imaging, in which magnetic resonance spectroscopy (Mrs) is the only non-invasive means to detect metabolites in living tissues. Bootstrap method has been widely used in biomedicine, finance, medicine, foreign trade and other fields in recent years since American scientist Efron systematically introduced Bootstrap method in 1979 to deduce the standard error of arbitrary estimation. In this paper, the residual self-help method and Wild self-help method of self-help method are applied to the MRS measurement of γ -aminobutyric acid (gaba) and T 1 atlas of human brain respectively. Based on the study of residual bootstrap magnetic resonance spectroscopy, the residual bootstrap method is combined with the uncertainty estimation of GABA content measurement for the first time, and strong robustness is obtained. Firstly, the anterior cingulate cortex (anterior cingulate cortex) of 12 healthy subjects was detected by Siemens 3.0 T Verio magnetic resonance scanner combined with MECHER-Garwood point-resolved spectroscopy (MEGA-Press). The concentration of gaba in ACC) and occipital cortex (OCC) was collected twice by each person, and the time between the two collections was one week or more. Then GABA concentrations of ACC and OCC in the brains of individuals and groups were analyzed by Gannet and model-based residual self-help technology. Residual Bootstrap resamples the residual of Gaussian fitting model, and then carries out Gaussian fitting on the resampled residual and collected data again. In individual research, the uncertainty and coefficient of variation (CV) of OCC GABA+ are less than ACC. In population studies, the uncertainty of GABA+ in ACC is less than that in OCC. However, compared with the traditional Gaussian fitting analysis, the residual self-help method obviously reduces the CV value and uncertainty in OCC and ACC regions. Residual bootstrap can provide robust uncertainty estimation when detecting GABA concentration in different regions of the brain of individuals and groups.