BackgroundPrevious studies have confirmed that physical exercise may be beneficial for brain health, but there is little data on this among older Chinese.ObjectiveThe purpose of this study was to explore the relationship between physical exercise and cognitive impairment, and to explore the possible mechanism by which physical exercise prevents cognitive decline.Materials and methods192 older adults with dementia, 610 older adults with mild cognitive impairment (MCI), and 2,218 healthy older adults were included in the study. Through standardized questionnaires, we obtained their general demographic information (such as gender, age, education, etc.), disease-related information (hypertension and diabetes) and physical exercise information (such as whether they did physical exercise and the frequency of physical exercise, etc.). The mini-mental state examination (MMSE) and Montreal Cognitive Assessment (MoCA) were used to assess their overall cognitive function, while the Wechsler block diagram was used to assess their executive function. Moreover, 164 healthy, randomly selected older adults also underwent brain MRI scans at the same time, and the target brain regions included hippocampus, gray matter, and cerebellar cortex.ResultsBy using stepwise multiple logistics regression analysis, we found that physical exercise was associated with both MCI (p = 0.001*, OR = 0.689, 95%CI: 0.553–0.859) and dementia (p < 0.001*, OR = 0.501, 95%CI: 0.354–0.709), independent of gender, age, education, and other factors. The results of ROC curve showed that the area under the curve of physical exercise in predicting MCI and dementia was 0.551 (p < 0.001*, 95%CI: 0.525–0.577) and 0.628 (p = 0.001*, 95%CI: 0.585–0.671), respectively. The results of partial correlation analysis showed that physical exercise was associated with left cerebellar cortex (r = 0.163, p = 0.023), right cerebellar cortex (r = 0.175, p = 0.015) and Wechsler block diagram score (r = 0.235, p = 0.011). Moreover, the results of linear regression analysis mediation model showed that physical exercise may affect Wechsler block diagram score through influencing the thickness of right cerebellum cortex, and the latter may play a partial mediation effect (indirect B = 0.001, p = 0.045).ConclusionPhysical exercise might be a protective factor for mild cognitive impairment and dementia among the Chinese elderly, and there might be an association among physical exercise, executive function, and the thickness of the cerebellar cortex.