GIST and Generalization in Young and Older Adults' Causal Learning

Abstract

Objectives To investigate age differences in GIST-based causal learning. Method Young and older adults learned to predict whether eating foods from each of 6 categories was or was not followed by sickness. Experience with the categories was varied by presenting 5 exemplars of a category in only one of 2 training phases (small categories) or by presenting 5 exemplars of a category in the first phase and an additional 5 exemplars in the second phase (large categories). Trained and novel exemplars from all categories appeared in subsequent causal judgment and recognition tests. Results There were no age differences in causal judgments for old exemplars or in generalization of causal associations to novel exemplars. For both groups, causal generalization was more successful and recognition was less successful for exemplars from large categories. There was no age difference in recognition for large categories, but older adults performed more poorly than young adults for small categories. Discussion Older adults resemble young adults in their ability to induce unseen category features from presented exemplars, acquire causal associations for these GIST representations, and generalize this knowledge to new exemplars. However, they continue to rely on these GIST-based representations for memory discrimination, whereas young adults use both category and individuating features of cues.