Conference on Cognitive Computational Neuroscience

Efficiently prioritizing important information is crucial for human memory function. Previous studies have demonstrated that the value of stimuli can selectively influence memory, with humans selectively remembering reward-relevant information. Here, we add to this understanding by decomposing reward-relevance to different compositional features, which collectively define the value of a stimulus with differing importance. Using combined reward learning and recognition memory tasks operating on the same set of stimuli, we investigate the impact of feature importance on memory. Our findings suggest that selective memory for the most rewarding feature is influenced by the depth of expertise during reward learning. This research adds to a growing body of research on the mechanism of value-based memory, with novel insights into how expertise influences selective memory for reward-relevant features.