On traditional pattern separation tasks, older adults perform worse than younger adults when identifying similar objects but perform equally well when recognizing repeated objects. When objects are superimposed on semantically related scenes, older adults are influenced by the context to a greater degree than younger adults, leading to errors when identifying similar objects. However, in everyday life, people rarely need to differentiate between two perceptually similar objects. Therefore, we developed a task using short stories to represent similar events people may experience in daily life. Our goal was to investigate the influence of context, detail-type, and age on memory performance.

Twenty-one older and 18 younger adults listened to 20 short stories taking place in either a coffee shop or library, each paired with a unique picture (i.e., context). Participants were asked to imagine the story taking place within the picture. Approximately 20 minutes later, participants answered a yes/no question about a detail from a story superimposed on different contexts. The different context conditions were (1) the same picture from the original story, (2) a similar picture (i.e., a different library or coffee shop picture), (3) a dissimilar picture (i.e., a library picture instead of a coffee shop picture), or (4) a control using a Fourier-transform (FT) image without any spatial-context information. Questions either asked about an identical or similar detail from the story.

Correct answers were analyzed using a 4x2x2 repeated measures ANOVA including context (same, similar, dissimilar, and FT), detail type (identical and similar), and age (younger and older adults). Overall, younger adults were more accurate than older adults, F(1,37)=23.4, p<0.001. However, surprisingly, the context and detail-type made no difference in accuracy, (F’s<1.1) A similar model was used to analyze reaction times. Younger adults were faster than older adults, F(1,37)=23.4, p<0.001. Participants of both ages were faster at correctly responding to the identical detail than the similar detail, F(1,114)=62.87, p<0.001. Context also impacted reaction time, F(3,114)=7.97, p<0.001. All participants were faster while viewing same and similar contexts compared to both the dissimilar and FT contexts (t(39)’s>2.20, p’s<0.05).

We did not find the kinds of age-related effects normally observed on traditional pattern separation tasks. Although younger adults performed better overall, older adults were not any worse when responding to a similar detail compared to an identical detail, which is inconsistent with performance on pattern separation tasks where older adults perform worse when identifying similar objects compared to younger adults. Additionally, older and younger adults were influenced by context in the same way. Previous studies from our laboratory demonstrated that older adults are biased toward the context when recognizing similar objects, but the context in this paradigm did not differentially influence accuracy for either older or younger adults. Potentially, this task relies on more semantic similarity rather than the perceptual similarity of objects. Semantic similarity from the short stories may incorporate more information to better orthogonalize similar memories, rendering retrieval less susceptible to interference.