The number of older individuals with Alzheimer’s disease (AD) reached 6.5 million in 2022 in the United States (Alzheimer’s Association, 2022). The high prevalence of cognitive impairment among older adults is an important source of burden for caregivers, public health and medical systems. Therefore, it is crucial to investigate factors that can protect the cognitive and brain health of older individuals. Being physically active in later-life is a strong candidate for protecting cognitive function (e.g., attention, memory, executive functions) and brain health (e.g., grey matter volume, white matter integrity, task-relevant activity or resting-state functional connectivity), as summarized in previous reviews (Bherer et al., 2013, Domingos et al., 2021, Erickson et al., 2019, Kramer and Colcombe, 2018, Tyndall et al., 2018). Meanwhile, human cross-sectional and longitudinal observational studies have also revealed an association between increased voluntary physical activity (PA) engagement and greater grey matter volumes and functional activity in a large swath of brain regions that are vulnerable to cognitive decline including the prefrontal cortex, temporal lobes and hippocampus (Domingos et al., 2021). Most of the evidence for PA benefits was obtained in the context of intervention studies and improved fitness. Although observational studies lack experimental control and cannot establish causality, they provide a more ecologically valid approach to examining the association of PA with health outcomes and an important complement to intervention research. It is equally important for observational findings to link habitual PA to cognition and brain health. This could further support the potential benefit of promoting a physically active lifestyle among adults.

Previous studies have found that a composite low-risk lifestyle (e.g., increased PA, and, alcohol and cigarette smoking avoidance) in midlife was associated with better physical and cognitive health status (Atallah et al., 2018, Kesse-Guyot et al., 2014), and preserved brain structure in older age (Chan et al., 2018, Franz et al., 2021). The importance of midlife lifestyle behaviors in dementia prevention was also emphasized in a comprehensive meta-analysis (Livingston et al., 2020). Therefore, midlife might be an important period in which a healthy lifestyle can minimize cognitive decline and increase brain reserve in later-life. Understanding the relationship between an individual midlife lifestyle behavior (e.g., PA) and later-life cognitive health would build the knowledge basis necessary to guide health policy and best practices around lifestyle engagement. With respect to physical activity, a number of studies have investigated the association between a physically active lifestyle in midlife and cognitive health in later-life. On the one hand, greater midlife PA engagement was associated with better global cognition (Carty et al., 2022), memory and executive function in later-life (Chang et al., 2010), and a lower risk of developing cognitive impairment, dementia, and AD pathology (Iso-Markku et al., 2016, Morgan et al., 2012, Rovio et al., 2005, Zotcheva et al., 2018). On the other hand, there are also studies that have reported the absence of an association between midlife PA level and cognitive health or dementia risk (Greendale et al., 2021, Gross et al., 2017, Kunutsor et al., 2021). Given the inconsistency of the relationship between midlife PA and cognitive health, it is possible that there are other risk factors that moderate this relationship. That is, the effect of midlife PA might be different across specific sub-populations, and studying sub-groups at high-risk (e.g., high AD risk) may provide insight into interventions that target these specific groups.

A deeper understanding of the relationship between midlife PA and later-life brain health would provide neurobiological evidence supporting midlife PA as a strategy for early prevention to delay cognitive impairment in later-life. A longitudinal observational study found that healthy individuals who were active in midlife had greater grey matter density and total volume in later-life, especially in the prefrontal cortex (Rovio et al., 2010); this indicates that the benefits of physical activity engagement in midlife may extend to brain structural integrity in older adulthood. Meanwhile, the variability of resting-state functional connectivity, a separate measure of brain function that represents the intrinsic and coordinated dynamics of brain activity (Deco et al., 2011), is also closely linked to brain aging and pathology (Sheline and Raichle, 2013). Functional connectivity is also modifiable after exercise interventions in later-life (Moore et al., 2022, Voss et al., 2010). Motivated by these prior findings, the present study focuses on the relationship between midlife PA and multimodal brain measures including structural integrity and functional connectivity in at-risk aging.

It is currently unknown whether PA engagement in midlife has dissociable contributions to cognition and brain health independent of later-life PA. It is tempting to hypothesize that PA engagement at both life stages may share similar neurobiological effects on later-life brain health. Consistent with the evidence from later-life exercise studies (Erickson et al., 2014), prior research has shown that midlife PA also provides cognitive benefits on memory and executive function (Chang et al., 2010) and grey matter integrity in prefrontal cortex in later-life (Rovio et al., 2010). However, given the mixed findings on the relationship between midlife PA and cognition, it is also worth asking whether becoming physically active or inactive in later-life moderates the neurobiological association with being physically active in midlife. Meanwhile, a report showed that most physically active older adults were also physically active earlier in their lifespan (Castillo et al., 2010). This suggests the possibility that individuals who were physically active in midlife might be more likely to maintain physical activity engagement in later-life, which is often not addressed in previous reports. Hence, we have not only considered the confounding role of later-life PA but also tested its potential interacting role in the association between midlife PA and later-life cognition and brain health. By taking account of later-life PA, we will be able to clarify the potentially unique contributions of midlife PA on later-life cognition and brain health.

The current study aimed to further clarify the association between midlife PA engagement, and its potential interaction with later-life PA, cognition, grey matter integrity and resting-state functional connectivity in a population at a high risk of AD from the Pre-symptomatic Evaluation of Experimental or Novel Treatments for Alzheimer’s Disease (PREVENT-AD) longitudinal cohort (Tremblay-Mercier et al., 2021). First, we examined the association between PA measures in midlife and multidomain cognitive function changes in this high-risk population. Next, given that prior hypotheses on midlife PA and brain measures were controversial in the literature, we examined grey matter integrity and functional connectivity in later-life on a whole-brain scale with a data-driven approach. Lastly, we also examined brain functional connectivity by focusing on regions of interest (ROIs) derived from prior observational studies on later-life PA engagement. This hypothesis-driven approach was made under the assumption of overlapping neurobiological associations between midlife and later-life PA engagement. In the present study, we focused on bilateral prefrontal cortex, specifically dorsolateral prefrontal cortex (DLPFC), and hippocampus given the prior literature that implicates these regions (Domingos et al., 2021, Erickson et al., 2014). This analytic approach revealed the potentially independent contributions of midlife PA engagement on cognition and brain health in older adults at high AD risk. This evidence will inform midlife intervention research to promote the prevention of cognitive impairment and AD progression.