Longitudinal Characterisation of Healthy Ageing in Rats using Multimodal Magnetic Resonance Imaging
Eilidh MacNicol
DOI: https://doi.org/10.31237/osf.io/3vpyg
2021-08-04
Abstract:Ageing is a universally heterogeneous process which is associated with a decline in certain cognitive functions and changes to brain structure, but a given individual’s ageing phenotype is complex and difficult to predict. Ageing trajectories are non-linear and influenced throughout the lifespan by a multitude of factors, including diet and lifestyle. Non-invasive neuroimaging, such as magnetic resonance imaging, facilitates the in vivo characterisation of brain structure and function and is widely used in the endeavour of identifying individual trajectories of the ageing brain.However, it is unknown how ageing, brain structure, and cognitive function are related, which exaggerates the difficulties in predicting whether an individual is likely to be susceptible to cognitive decline and to what extent. In fact, ageing studies, which are typically cross-sectional, are often not designed to achieve this end; longitudinal studies, which are the most pertinent design to account for the considerable degree of individual variability observed, are less common due to the cost and time restraints. Thus, animal models of healthy ageing are valuable because they develop analogous phenotypes within a shorter lifespan, which reduces the time necessary to complete a longitudinal study. Moreover, there is greater control of environmental and genetic factors in preclinical studies, which can be difficult to account for in human studies.The Resilient study followed a cohort of 48 male Sprague-Dawley rats who were behaviourally tested and underwent an extensive MRI protocol up to four times, to characterise changes in brain structure and function across the adult lifespan. Half of the rats were also subjected to an environmental enrichment protocol, which involves weekly rotation of toys in the home cage, and a caloric restriction in the form of intermittent fasting. Combined, both interventions are intended to mimic the beneficial lifestyle choices believed to promote healthy ageing. A clinical frailty battery, and behaviour tests, including novel object recognition and attentional set-shifting, were employed to characterise individual ageing trajectories.Brain structure was assessed with a variety of metrics including a comparison of brain volumes, and mass univariate voxelwise comparisons using voxel-based morphometry, tensor-based morphometry, relaxometry, and diffusion metrics. As structurally similar regions are more likely to be physically connected, this was complemented with the construction of structural connectome graphs, derived from morphometric similarity networks using the multi-modal metrics to characterise 154 brain regions. Changes in brain structure between the groups were most frequently observed at the 5-month time point, which was likely a consequence of the difference in tissue growth.Brain function was assessed by graph-theoretical analyses of adjacency matrices derived from the regional BOLD signal fluctuations, as measured by multi-echo resting-state fMRI, and regional comparisons of cerebral blood flow maps. Differences were observed between the groups’ overall functional connectivity and modularity, particularly at 5 months old. Although the groups did not differ in their overall functional connectivity of the default mode network, the previously observed result of declining connectivity within the default mode network with increasing age was replicated, with the EEDR group deferring their largest decline in DMN covariance later than the decline observed in the control group. Finally, an age-related increase in cerebral blood flow was observed, and variance was observed in the regional trajectories, indicating that the change of perfusion rate is not uniform across the brain.As there were fewer group differences than hypothesised a priori, post-hoc analyses explored the functional and structural changes using physiological and behavioural outcomes instead of experimental grouping. However, the metrics explored did not explain more of the individual variability than the original groups. Consequently, post-hoc exploratory analyses used weight, experimental group, and age to predict the observed age-related changes in brain structure and function, revealing substantial residual variance despite controlling for factors controlled by genetics, shared environment, and age. The effect of body weight on brain tissue volume and defaultmode network overall connectivity changed with age and is possibly reflective of the changes in metabolic processes that underpin brain growth and remodelling.Future directions for this work include larger replications, without restrictive timelines, which would better characterise the life outcomes of the subjects. Moreover, future work should measure both cerebral metabolic rate of oxygen and glucose, to measure the relative contributions of metabolic substrates that may underpin the observed differences in brain structure and function in healthy ageing.