Abstract
OBJECTIVE: Age-related hippocampal atrophy is associated with memory loss in older adults, and certain hippocampal subfields are more vulnerable to age-related atrophy than others. Cardiorespiratory fitness (CRF) may be an important protective factor for preserving hippocampal volume, but little is known about how CRF relates to the volume of specific hippocampal subfields, and whether associations between CRF and hippocampal subfield volumes are related to episodic memory performance. To address these gaps, the current study evaluates the associations among baseline CRF, hippocampal subfield volumes, and episodic memory performance in cognitively unimpaired older adults from the Investigating Gains in Neurocognition Trial of Exercise (IGNITE) (NCT02875301). METHODS: Participants (N = 601, ages 65-80, 72% female) completed assessments including a graded exercise test measuring peak oxygen comsumption (VO(2peak)) to assess CRF, cognitive testing, and high-resolution magnetic resonance imaging of the hippocampus processed with Automated Segmentation of Hippocampal Subfields (ASHS). Separate linear regression models examined whether CRF was associated with hippocampal subfield volumes and whether those assocations were moderated by age or sex. Mediation models examined whether hippocampal volumes statistically mediated the relationship between CRF and episodic memory performance. Covariates included age, sex, years of education, body mass index, estimated intracranial volume, and study site. RESULTS: Higher CRF was significantly associated with greater total left (B = 5.82, p = 0.039) and total right (B = 7.64, p = 0.006) hippocampal volume, as well as greater left CA2 (B = 0.14, p = 0.022) and dentate gyrus (DG; B = 2.34, p = 0.031) volume, and greater right CA1 (B = 3.99, p = 0.011), CA2 (B = 0.15, p = 0.002), and subiculum (B = 1.56, p = 0.004) volume. Sex significantly moderated left DG volume (B = -4.26, p = 0.017), such that the association was positive and significant only for males. Total left hippocampal volume [indirect effect = 0.002, 95% CI (0.0002, 0.00), p = 0.027] and right subiculum volume [indirect effect = 0.002, 95% CI (0.0007, 0.01), p = 0.006] statistically mediated the relationship between CRF and episodic memory performance. DISCUSSION: While higher CRF was significantly associated with greater total hippocampal volume, CRF was not associated with all underlying subfield volumes. Our results further demonstrate the relevance of the associations between CRF and hippocampal volume for episodic memory performance. Finally, our results suggest that the regionally-specific effects of aging and Alzheimer's disease on hippocampal subfields could be mitigated by maintaining higher CRF in older adulthood.