With great interest we read the paper by Iwasa et al. describing the longitudinal relationship between leisure activities and cognitive decline among Japanese community-dwelling older adults. With dementia reaching epidemic proportions worldwide research which may help elucidate ways to delay or even prevent it are crucial. The strengths of Iwasa et al.’s study are clear its longitudinal nature (5year follow-up) and the large sample population (over 500) of community‐dwelling men and woman aged 70years and over. That the authors attempted to include baseline covariates (both demographic and clinical) can also be seen as a strength; however, more details on these variables would have been useful. Vascular risk factors (not just smoking but also alcohol and medication usage) should also have been included. On the whole though, this is a relatively well-designed study. Nevertheless we have two main suggestions for possible improvements which may help future work in this field, namely, on the issue of which outcome measure/s to use to assess cognitive functioning, and on how to specifically categorize leisure activities.