Locus coeruleus integrity is related to an exploitation-based decision-making bias in older adulthood

We first tested our core premise of an age-related exploitation bias on a foraging-based decision-making task. We derived measures of exploration and exploitation bias from a foraging task paradigm wherein participants foraged for berries among spatially distributed patches with varying densities of berry clusters (13). Exploitation was defined as staying in patches longer than optimal, while exploration was defined as leaving patches earlier than optimal. Optimal performance, defined here as Leave Time Difference (LTD), was calculated using marginal value theorem (ref. 13 and SI Appendix, Materials and Methods). Consistent with our model (1) and review (14), we observed a robust age-related exploitation bias on our foraging task (No bias = 0, positive score = exploration bias, negative score = exploitation-bias). Younger adults (N = 108): M = 6.7(SD = 13.8), range = −32.2 to 44.9; Older adults (N = 145): M = −3.1(SD = 15.1), range = −49.2 to 28.2; t(251) = −5.3, P < 0.001, Cohen’s d = 0.67).

Next, we examined MTsat associations with chronological age within our older adult sample who had LC MTsat data (N = 220). As predicted (9), microstructural integrity of LC was significantly lower with advancing age (Fig. 1B; r(218) = −0.25, P < 0.001; pr(215) = −0.25, P < 0.001). Critically, we included sex, education, and brain size in the partial correlation model, the latter of which was included here as a global measure of age-related differences in brain structure and suggests that MTsat is a specific marker of structural brain aging.

Finally, we tested our central prediction that LC microstructural integrity would be associated with an exploitation bias within our sample of older adults who had both task and LC MTsat data (N = 133). As predicted, lower LC integrity was robustly associated with an exploitation-bias during foraging in this older adult cohort (Fig. 1C; r(131) = 0.22, P = 0.01; pr(126) = 0.28, P < 0.01). In the partial correlation analysis, we included the following as covariates: sex, education, brain size, age and MTsat of the corpus callosum to demonstrate specificity (Materials and Methods).

Older age is associated with an exploitation bias as well as lower microstructural integrity of LC. Critically, lower LC integrity is robustly associated with greater exploitation on a foraging-based decision-making task. (i.e., foraging too long within, and suboptimally switching between, berry patches). Taken together, the results provide evidence that LC integrity is a neurobiological substrate for an exploitative decision-making bias in older adults.

We have argued that larger stores of prior knowledge, declining cognitive control, awareness of one’s mortality, and focus on positive close social bonds converge in later life; creating a prepotent tendency to exploit options with known rewards, while avoiding the ambiguity and risk associated with exploration (1). However, optimal decision-making depends on alternating between exploitation and exploration to effectively track shifting reward and risk probabilities over time. As this is mediated by firing patterns of LC neurons (2, 15), we reasoned that lower microstructural integrity in LC would exacerbate prepotent exploitation-biases in later life. Our findings provide strong support for this prediction. Moreover, the association remained reliable after controlling for brain volume and MTsat in a myelin-rich region of no interest, confirming specificity to LC integrity, over and above global changes in brain health.

Decisions to explore or exploit arguably sit at the very core of human volition (16). Subtle biases can fundamentally shape how we interact with and experience the world around us. Overreliance on prior knowledge and experience comes at the cost of new learning and discovery. In contrast, overreliance on novelty and information-seeking comes at the cost of stability and certainty. There is no single path. Understanding neurobiological changes that may shape this landscape, favoring one direction over another, is critically necessary to promote and preserve optimal decision-making abilities. Here, we provide evidence that subcortical brain changes impact the decision-making landscape in older adulthood, a crucial step toward revealing broader structural and functional brain changes that bias decisional capacities in later life. Future research integrating multimodal techniques, including high-field (7 Tesla MRI) structural and functional imaging, as well as neurochemical positron emission tomography imaging, will enable increasingly precise neurobiological models and unique insights into the neural mechanisms that may portend an exploitative decision-making bias in aging and brain disease.

We would like to thank the PREVENT-AD research team and study participants for their time and dedication in collecting the data used in this study. Full acknowledgments can be found in Supplemental Materials. This work was supported in part by a grant from the Canadian Institute of Health Research (CIHR) to G.R.T and R.N.S., and grants from the Healthy Brains for Healthy Lives, Alzheimer’s Association (AARG-22-927100) and NIA R01 AG068563 to R.N.S., who is supported by Fonds de recherche du Québec–Santé. The PREVENT-AD cohort is supported in part by grants from CIHR (J.P. and S.V.), FRQS (J.P. and S.V.) and the J. L. Levesque Foundation. Data used in preparation of this article were obtained from the Presymptomatic Evaluation of Novel or Experimental Treatments for Alzheimer’s Disease (PREVENT-AD) program (https://douglas.research.mcgill.ca/stop-ad-centre). A complete listing of the PREVENT-AD Research Group can be found in: https://preventad.loris.ca/acknowledgements/acknowledgements.php?date=[2023-07-01].