Predictability of Memorability Neural Pattern and Attention Network in Subsequent Memory Retrieval

This study delved into the roles of memorability and attention – two pivotal elements in perception – in contributing to memory encoding through their neural patterns. The study replicated the memorability effect demonstrating that memorable images are remembered better than forgettable ones. Utilizing neuroimaging data from one previous dataset, we extracted the neural pattern associated with memorability, and applied it to this dataset to derive a memorability estimator. We also incorporated the functional connectivity network of attention to serve as a real-time attention state estimator. To predict memory performance, a generalized linear model was utilized to incorporate these two factors. While the findings did not find significant contribution of the neural patterns of these perception factors to the memory process, they revealed significant variability in neural correlates across individuals. This variability suggests the need for future research to adopt more robust and diverse methodological approaches, including larger sample sizes and new analytical models, to capture the intricacies of these cognitive and neural processes more accurately.