Getting Fit to Counteract Cognitive Aging: Evidence and Future Directions

Abstract

Physical activity has shown tremendous promise for counteracting cognitive aging, but also tremendous variability in cognitive benefits. We describe evidence for how exercise affects cognitive and brain aging, and whether cardiorespiratory fitness is a key factor. We highlight a brain network framework as a valuable paradigm for the mechanistic insight needed to tailor physical activity for cognitive benefits.

Keywords: aging; brain; cognition; fitness; networks.

FIGURE 1.

The cardiorespiratory fitness hypothesis posits that the benefits of physical activity on cognitive aging are mediated by physiological adaptations associated with cardiorespiratory fitness An extension of the hypothesis is that the benefits of fitness on the aging brain are mediated through a direct effect on brain cerebrovascular health, metabolism, structure, and neuronal function.

FIGURE 2.

A: cognitive domains as operationalized in two key meta-analyses summarizing effects of exercise training on cognition, as defined by a well-known neuropsychological model of cognition (31)B and C: forest plot of effect sizes for exercise effects on cognition domains, shown as point estimates with 95% confidence intervals [Smith et al. 2010 (29) and Northey et al. (30)]. D and E: effects of exercise program moderators that are known to influence cardiorespiratory fitness benefits from training, but showed relatively little effect on cognitive outcomes in adults ages 50 and older (30). WAIS, Wechsler Adult Intelligence Scale.

FIGURE 3.

With tools of network science, we can model brain network connectivity as a brain graph to identify the functional role of brain regions (nodes) interacting with each other through structural and functional connections (edges) in subnetwork communities. In the brain graph on the left, the color of nodes denote distinct brain subnetworks, as shown mapped to an anatomical brain on the right Nodes closer together in topological space on the graph are more strongly connected, which can occur across long anatomical distances in the brain. Connector hub nodes are distinguished by a high diversity of subnetwork connections, thought to facilitate their role as conductors for subnetwork interactions. Connector node function requires high metabolic demand because of the diversity and anatomical distance of their signaling. L, left; R, right. To the right are (row 1) brain regions whose functional connectivity has been shown to increase with cardiorespiratory fitness (CRF) or aerobic exercise training that improved CRF, including the left and right middle temporal gyrus (LMTG and RMTG), left parahippocampal gyrus (LPHG), and left superior and middle frontal gyrus (LSFG and LMFG). Rows 2–3: these nodes overlap with lateral frontoparietal and midcingulo-insular subnetworks implicated in cognitive control and medial frontoparietal networks implicated in long-term memory. Rows 3 and 4: network nodes color coded by their participation coefficient (PC). The young adult cohort PC map was published as a normative sample (n = 62; age 24 ± 5.2 yr; 44% males) (53) from the Nathan Kline Institute Rockland sample (54); the older adult cohort includes cognitively normal older adults from the University of Iowa (n = 134; age 67 ± 5.6 yr; 42% males). Brighter nodes have higher PC indicating they operate most strongly as connector hubs. Prefrontal regions with higher functional connectivity to other regions in the brain as a function of CRF are connector nodes.

FIGURE 4.

An illustration of the potential protective effect of increasing cardiorespiratory fitness (CRF) on cognition through reducing the burden of biological aging on network function A: negative effects of aging on vascular function, brain atrophy, and cognition are prevented or slowed through maintaining average to high CRF levels. Protective pathways of CRF are indicated by (-) with varying levels of existing evidence denoted by size of the circle. Network hub function and modularity are functional mechanisms linking vascular and structural change to cognitive aging. Individuals A, B, and C are 3 hypothetical individuals with the same cognitive status early in their lives, whom carry progressively more biological aging burden. Individual A has maintained CRF from young adulthood, attenuating biological aging; their CRF is at a personal best and they are not well-represented by exercise clinical trials due to their high physical activity (PA) levels. Individuals B and C have lower than average CRF, leading to more biological aging burden. Individual B has lower burden, presenting more opportunity for CRF to counteract structural and functional networks to slow cognitive aging. B: severity of burden in biological aging constrains the potential preventative pathways through which CRF slows cognitive aging. Individual A maintains protective pathways for CRF to delay cognitive aging. Individual B is predicted to gain more from PA that improves CRF than individual C because they maintain more protective pathways counteracting aging shown in A. Greater training-induced gains in cognition from improved CRF for individual B compared with individual C is illustrated by greater volume of blue than red shaded regions symbolizing delayed cognitive aging.