Shorter Planning Depth and Higher Response Noise During Sequential Decision-Making in Old Age

Abstract

Abstract Forward planning is crucial to maximize outcome in complex sequential decision-making scenarios. In this cross-sectional study, we were particularly interested in age-related differences of forward planning. We presumed that especially older individuals would show a shorter planning depth to keep the costs of modelbased decision-making within limits. To test this hypothesis, we developed a sequential decision-making task to assess forward planning in younger (age < 40 years; n = 25) and older (age > 60 years; n = 27) adults. By using reinforcement learning modelling, we inferred planning depths from participants' choices. Our results showed significantly shorter planning depths and higher response noise for older adults. Age differences in planning depth were only partially explained by wellknown cognitive covariates such as working memory and processing speed. Consistent with previous findings, this indicates agerelated shifts away from modelbased behaviour in older adults. In addition to a shorter planning depth, our findings suggest that older adults also apply a variety of heuristical low-cost strategies.