$BACKGROUND$: The promotion and maintenance of higher physical activity (PA) levels in the older population is an imperative for cognitive and healthy ageing but it is unclear what approaches are best suited to achieve this for the increasing number of older people living in the community. Effective policies should be informed by robust, multi-disciplinary and multi-dimensional evidence, which not only seeks what works, but in 'what context? In addition to evidence on the efficacy and effectiveness of PA for maintaining cognitive health, social contexts such as 'how do we actually get older people to partake in PA?' and 'how do we sustain that activity long-term?' also need highlighting. This review is part of a comprehensive evidence synthesis of preventive interventions in older age, with a focus on healthy behaviours to identify evidence gaps and inform policy relating to ageing well and cognitive health. An overview of systematic reviews of PA was conducted to explore three topics: (1) PA efficacy or effectiveness for primary prevention of cognitive decline in 55+; (2) Interventions efficacious or effective for increasing PA uptake and maintenance in 55+; (3) barriers and facilitators to PA in 55+.

$METHODS$: Multiple databases were searched for studies in English from OECD countries between 2000 and 2016. Quality of included reviews in questions (1) and (2) were assessed using AMSTAR. Review protocols were registered on PROSPERO (CRD42014015554, 42014015584, CRD42014015557) and reviews follow PRISMA guideline.

$FINDINGS$: Overall, 40 systematic reviews were included. Question 1 (n = 14). 8,360 participants. Evidence suggests that PA confer mild positive effects on cognition in older adults with and without previous cognitive impairment. However, there is insufficient evidence of a dose-response relationship. Evidence on the effects of PA on delay of dementia onset is inconclusive. Question 2 (n = 17). 79,650 participants. Evidence supports the effectiveness of a variety of interventions, including group delivered, centre-based and cognitive approaches on short-term uptake of PA behaviour. Question 3 (n = 9). 22,413 participants. Barriers include health status, previous PA habits and experiences, and cultural sensitivity, while facilitators include enjoyable activities and convenient scheduling.

$CONCLUSION$: PA can offer small benefits to brain health, but evidence on how much activity is required to produce this effect is lacking. Evidence on the effectiveness of PA for preventing dementia and cognitive decline is lacking. Behavioural (walking, exercise) and cognitive (counselling and motivational interviews) interventions are effective for short-term uptake of physical activity in older people. In order to maintain long-term participation in PA, individualised interventions modelled using behavioural theories may be required. Public health messages should be aimed at promoting acceptable levels of PA above normal daily activities in older people. Policy and strategies aimed at increasing PA in older people should be encouraged while considering barriers and facilitators to behaviour change.