Perceptual memorability and reward valuation make parallel contributions to successful recognition memory

Salman Ehtesham Qasim,Ignacio Saez,Xiaosi Gu

crossref(2022)

引用 0|浏览1
暂无评分
摘要
Our experiences can be memorable for different reasons. Prior research has identified the importance of intrinsic perceptual features to memorability. However, recent work suggests that value-driven learning (i.e. via reward prediction errors [RPEs]) also influences memorability. Here, we examined the joint contribution of perceptual memorability and value-based computations in influencing memory strength. Participants performed a two-arm bandit reversal learning task, in which they learned the value associated with images with different levels of baseline perceptual memorability. Next, participants performed a recognition memory task. Importantly, both the model-estimated RPEs and the perceptual memorability associated with images contributed to recognition memory. Value-driven individuals, who earned more rewards on the learning task, also prioritized these extrinsic RPEs over intrinsic perceptual features during recognition memory. Over-reliance on RPE was advantageous to memory performance in individuals with low trait anxiety, with this advantage diminishing in highly anxious individuals. Together, these findings reveal a parallel relationship between intrinsic (perceptual memorability) and extrinsic (learned association with RPE) sources of information that both contribute to recognition memory; this relationship is dynamically modulated by both individual decision making and mental health traits.
更多
查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要