Cardiorespiratory fitness is differentially associated with cortical thickness in young and older adults

Highlights

• Few studies have examined peak VO₂ and cortical integrity in young and older adults.

• Greater peak VO₂ was associated with thicker cortex in healthy older adults.

• Greater peak VO₂ was associated with thinner cortex in healthy younger adults.

• Multiple regions that show cortical decline in aging are also sensitive to fitness.

• Relation between fitness and cortical thickness appears to be age dependent.

Abstract

Aging is associated with reductions in gray matter volume and cortical thickness. One factor that may play a role in mitigating age-associated brain decline is cardiorespiratory fitness (CRF). Although previous work has identified a positive association between CRF and gray matter volume, the relationship between CRF and cortical thickness, which serves as a more sensitive indicator of gray matter integrity, has yet to be assessed in healthy young and older adults. To address this gap in the literature, 32 young and 29 older adults completed treadmill-based progressive maximal exercise testing to assess CRF (peak VO₂), and structural magnetic resonance imaging (MRI) to determine vertex-wise surface-based cortical thickness metrics. Results indicated a significant CRF by age group interaction such that Peak VO₂ was associated with thicker cortex in older adults but with thinner cortex in young adults. Notably, the majority of regions demonstrating a positive association between peak VO₂ and cortical thickness in older adults overlapped with brain regions showing significant age-related cortical thinning. Further, when older adults were categorized as high or low fit based on normative data, we observed a stepwise pattern whereby cortex was thickest in young adults, intermediate in high fit older adults and thinnest in low fit older adults. Overall, these results support the notion that CRF-related neuroplasticity may reduce although not eliminate age-related cortical atrophy.

Introduction

Cerebral volume loss is a well-documented correlate of the aging process, accompanied by a reliable pattern of regional cortical thinning with advancing age (Fjell et al., 2009b, Salat et al., 2004, Shaw et al., 2016). Brain regions most susceptible to age-related cortical degeneration largely coincide with heteromodal association areas that support higher-level processing, and include lateral temporal, inferior parietal, and frontal regions, as well as the precuneus, temporoparietal junction, and fusiform gyrus (Fjell et al., 2009a, Fjell et al., 2009b, Shaw et al., 2016). Despite a general trend of cortical atrophy in healthy aging, individual differences have been observed in the rate and extent of degeneration, with some adults maintaining relatively more intact structural integrity throughout late life than same-aged peers (Pfefferbaum and Sullivan, 2015). Thus, in an effort to promote optimal brain health within a society where the population distribution is increasingly shifting towards older ages (Vincent and Velkoff, 2010), identifying relevant lifestyle factors that may protect against age-related neurostructural decline is warranted.

One factor that may play a role in mitigating age-associated brain decline is cardiorespiratory fitness¹ (CRF). CRF reflects the efficiency of the circulatory and respiratory systems to provide oxygenated blood to musculature during sustained aerobic physical activity. As a modifiable health factor, CRF can be optimized through regular engagement in physical activity of moderate to vigorous intensity such as jogging, swimming, or biking. The gold standard of CRF assessment is treadmill-based graded maximal exercise testing, which yields a measure of peak volume of oxygen consumption (peak VO₂) at maximum intensity exercise. Compared to self-report or estimated CRF metrics, peak VO₂ provides an objective and reliable indicator of CRF. Whereas systemic benefits of enhanced CRF to both emotional and physical health are well documented (DiLorenzo et al., 1999, Myers et al., 2015, Papasavvas et al., 2016, Tozzi et al., 2016, Warburton et al., 2006), only recently have studies begun to explore the relation between CRF and brain structure, particularly in the context of aging.

Cross-sectional studies of older adults have generally reported a positive association of CRF with gray matter volume (Boots et al., 2015, Bugg and Head, 2011, Erickson et al., 2007, Gordon et al., 2008, Weinstein et al., 2012) as well as with cognitive performance in the domains of executive function and memory (Barnes et al., 2003, Hayes et al., 2016). Interestingly, reliable CRF effects have been observed primarily within lateral prefrontal and parietal gray matter regions (Erickson et al., 2014, Hayes et al., 2014), cortical areas most vulnerable to age-related atrophy (Fjell et al., 2009b). Given this apparent regional overlap between CRF and aging effects, it follows that enhanced CRF could lessen the degree of cortical decline in aging. In support of this concept, aerobic exercise intervention studies have provided evidence of CRF-related neuroplasticity, as exercising older adults show greater regional brain volume than age-matched controls (Colcombe et al., 2006, Erickson et al., 2011).

Although positive associations between CRF and gray matter volume have been previously reported, it is less clear how CRF may relate to cortical thickness in older adult populations. Given that the surface area component used to define volume captures additional variance related to head size and the degree of regional cortical folding, volumetric approaches have been shown to yield less sensitive estimates of age-associated cortical atrophy when compared to surface-based cortical thickness techniques (Hutton et al., 2009, Lemaitre et al., 2012, Panizzon et al., 2009). Studies of patient populations (mild cognitive impairment, heart failure, and schizophrenia) have demonstrated positive associations between cortical thickness and CRF (Alosco et al., 2013, Reiter et al., 2015, Scheewe et al., 2013), yet the association between CRF and cortical thickness in the context of healthy aging remains unknown. One recent study (Lee et al., 2016) demonstrated a positive relation between self-reported physical activity and cortical thickness in healthy older adults. However, given that subjective CRF measures do not align well with objectively quantified peak VO₂ (Tager et al., 1998), additional study is warranted.

In the current cross-sectional study, we investigated age-dependent associations between objectively quantified CRF (peak VO₂) and cortical thickness in healthy older and young adults using a whole-brain cortical surface-based approach. The primary goals of this study were 1) to assess whether CRF is differentially associated with cortical thickness in young and older adults, 2) to determine the spatial overlap between observed CRF and aging effects across the cortex, and 3) to examine whether higher CRF in late life may eliminate age-related cortical thickness decline. Based on the previously reported beneficial effects of CRF on gray matter volume in aging, as well as positive associations between cortical thickness and CRF in patient populations, we expected to observe similar positive associations between CRF and cortical thickness in our older adult cohort. In particular, we anticipated that regions where CRF is positively associated with cortical thickness in older adults would overlap with frontal and temporoparietal regions most susceptible to age-related cortical thinning. Given the dearth of studies in young adults, we had no strong prediction about the association between CRF and cortical thickness for this cohort. A recent study of early adolescents (aged 9–11 years) found that greater CRF was associated with thinner cortex within regions of superior frontal, superior temporal and lateral occipital cortex (Chaddock-Heyman et al., 2015), a finding that likely reflects successful maturational development and neural pruning. The fact that in some brain regions protracted maturational cortical thinning has been observed throughout young adulthood (Tamnes et al., 2013) raises the possibility that young adults may show a negative association between CRF and cortical thickness similar to that observed in adolescents. By examining young and older adults in the same study, we were able to directly assess age-dependent associations between CRF and cortical thickness, and additionally, whether age-related differences in cortical thickness are impacted by CRF.