Context and purpose: The mechanisms of ageing of the healthy brain are not entirely clarified to date. In recent years several authors have tried to elucidate this topic by using F-18-FDG positron emission tomography. However, when correcting for partial volume effects (PVE), divergent results were reported. Therefore, it is necessary to evaluate these methods in the presence of atrophy due to ageing. In this paper we first evaluate the performance of two PVE correction techniques with a phantom study: the Rousset method and iterative deconvolution. We show that the ability of the latter method to recover the true activity in a small region decreases with increasing age due to brain atrophy. Next, we have calculated age-dependent recovery factors to correct for this incomplete recovery. These factors were applied to PVE-corrected F-18-FDG PET scans of healthy subjects for mapping the age-dependent metabolism in the brain. Results: Many regions in the brain show a reduced metabolism with ageing, especially in grey matter in the frontal and temporal lobe. An increased metabolism is found in grey matter of the cerebellum and thalamus. Conclusion: Our study resulted in age-dependent recovery factors which can be applied following standard PVE correction methods. Cancelling the effect of atrophy, we found regional changes in F-18-FDG metabolism with ageing. A decreasing trend is found in the frontal and temporal lobe, whereas an increasing metabolism with ageing is observed in the thalamus and cerebellum.