ABSTRACTAgeing is a complex phenomenon affecting a wide range of coexisting biological processes. The homogeneity of the studied population is an essential parameter for valid interpretations of outcomes. The presented study capitalises on the MRI data available in the Human Connectome Project–Aging (HCP‐A) and, within individuals over 55 years of age who passed the HCP‐A section criteria, compares a subgroup of 37 apparently neurocognitively healthy individuals selected based on stringent criteria with 37 age and sex‐matched individuals still representative of typical ageing but who did not pass the stringent definition of neurocognitively healthy. Specifically, structural scans, diffusion weighted imaging and T1w/T2w ratio were utilised. Furthermore, data of 26 HCP‐A participants older than 90 years as notional ‘super‐agers’ were analysed. The relationship of age and several microstructural MRI metrics (T1w/T2w ratio, mean diffusivity, intracellular volume fraction and free water volume fraction) differed significantly between typical and healthy ageing cohort in areas highly relevant for ageing such as hippocampus, prefrontal and temporal cortex and cerebellum. However, the trajectories of the healthy ageing population did not show substantially better overlap with the findings in people older than 90 than those of the typical population. Therefore, caution must be exercised in the choice of adequate study group characteristics relevant for respective ageing‐related hypotheses. Contrary to typical ageing group, the healthy ageing cohort may show generally stable levels of several MRI metrics of interest.