Abstract : The impact of age is crucial and must be taken into account when applying a voxel-based quantitative analysis on brain images from [(18)F]-fluorodeoxyglucose Positron Emission Tomography (FDG-PET). This study aimed to determine whether age-related changes in brain FDG-PET images are more accurately assessed when the conventional statistical parametric mapping (SPM) normalization method is used with an adaptive template, obtained from analysed PET images using a Block-Matching (BM) algorithm to fit with the characteristics of these images. Age-related changes in FDG-PET images were computed with linear models in 84 neurologically healthy subjects (35 women, 19 to 82-year-old), and compared between results provided by the SPM normalization algorithm applied on its dedicated conventional template or on the adaptive BM template. A threshold P value of 0.05 was used together with a family-wise error correction. The age-related changes in FDG-PET images were much more apparent when computed with the adaptive template than with the conventional template as evidenced by: (1) stronger correlation coefficients with age for the overall frontal and temporal uptake values (respective R (2) values of 0.20 and 0.07) and (2) larger extents of involved areas (13 and 5 % of whole brain template volume, respectively), leading to reveal several age-dependent areas (especially in dorsolateral prefrontal, inferior temporal/fusiform and primary somatosensory cortices). Age-related changes in brain FDG uptake may be more accurately determined when applying the SPM method of voxel-based quantitative analysis on a template that best fits the characteristics of the analysed TEP images.