Abstract
Cognitive functioning is central to independence and quality of life across adulthood, yet the drivers of individual differences in cognition and trajectories of cognitive ageing remain incompletely understood. Identifying the factors that shape these differences and support cognitive resilience has therefore become a key goal of contemporary ageing research. Current evidence, however, is often constrained by a focus on isolated biological systems and cognitive measures, as well as by analytical approaches that generalise poorly beyond the samples in which they were developed. This thesis addresses these limitations by applying predictive modelling to derive markers of cognitive functioning from mental health, body physiology, lifestyle behaviours, environmental exposures, and neuroimaging data, and by quantifying the joint contributions of these domains to cognitive ageing.
The first study (https://doi.org/10.7554/eLife.108109) examined predictive relationships between twelve mental health domains and cognition, and quantified the extent to which these associations are reflected in brain structure and function measured across three neuroimaging modalities: diffusion-weighted MRI (dwMRI), resting-state functional MRI (rsMRI), and structural MRI (sMRI). Mental health predicted cognition at r = 0.30, and 48% of this association was captured by a composite brain marker integrating phenotypes across the three modalities. Age and sex shared substantial overlapping variance with both mental health and neuroimaging in explaining cognition, accounting for 43% of the variance in the cognition–mental health relationship.
The second study (https://doi.org/10.64898/2026.01.13.26343950) investigated the predictive utility of nineteen body physiology phenotypes for cognition and assessed the contribution of brain-based markers to the body–cognition relationship. A composite body marker integrating all body physiology phenotypes predicted cognition at r = 0.40, and 85.1% of this association was explained by a composite brain marker. Together, body and brain accounted for 96.8% of age-related variance in cognition.
The third study (https://doi.org/10.64898/2026.02.26.26347222) evaluated the predictive contribution of twelve lifestyle and environment domains and examined the extent to which lifestyle–cognition associations are captured by body and brain markers. A composite lifestyle–environment marker predicted cognition at r = 0.48, and a combined body–brain marker accounted for 55.9% of this association. Together, lifestyle–environment and body–brain markers explained 92.6% of age-related variation in cognition.
Overall, these findings support a systems-level view in which individual differences in cognition reflect interactions between neural pathways and broader physiological, behavioural, and environmental influences. By quantifying how mental health, bodily physiology, lifestyle behaviours, and environmental exposures relate to cognitive functioning, this thesis lays a quantitative foundation for future strategies aimed at maintaining cognitive health in ageing populations.