Reduced age differences in semantic memory networks: Evidence from semantically diverse free associations.

A hallmark of successful aging is increased life experiences and knowledge. Yet how this additional information is incorporated into semantic memory is unclear. Network science has proven to be a useful tool for modeling semantic memory networks in younger and older adults. Previous research suggests that although vocabulary and knowledge are largely stable across adulthood, older adults may have semantic memory networks that are less efficient, less interconnected, and more segregated. However, prior work, including our own, has largely focused on semantic memory networks derived from highly salient, physical concepts (e.g., animals). Though words essential for natural conversation vary greatly in terms of their psycholinguistic characteristics. In the present study, we examine age-related differences in semantic memory networks derived from a free association task, using both abstract and concrete cues that varied in semantic diversity - the number of unique contexts in which they could appear. Across several analytic approaches, we found that including abstract words in semantic memory networks minimized age-related differences: there were no age differences in network efficiency, but older adults had more interconnected and less segregated semantic memory networks compared to younger adults. Looking at word-level characteristics of the semantic memory networks suggested that for both younger and older adults, words that were high in semantic diversity and were more abstract had stronger connections to other words and were more interconnected. These results suggest that abstract and semantically diverse words are a cornerstone in maintaining older adults' semantic memory networks.