Marika Mauti^1,2 (), Moreno I. Coco^1,2; ¹Sapienza University of Rome, Rome, Italy, ²IRCCS Fondazione Santa Lucia, Rome, Italy

Our memory processes undergo a continuous change, shaped by a dynamic interplay between the formation and decay of episodic representations. While memory formation has been extensively investigated, the forgetting mechanisms underlying memory degradation remain elusive. More importantly, forgetting has always been derived as changes in the rate of successfully recalled memory over time. However, as memory processes are intimately related to attentional mechanisms, it is conceivable to derive an explicit understanding of forgetting processes directly from them. Unlike most studies on forgetting, we employed naturalistic scenes rather than letters or isolated objects, which provide rich and meaningful contexts, returning a more ecologically valid explanation of its mechanisms. Participants (N = 15) studied 132 scenes encompassing diverse indoor and outdoor settings in preparation for three recognition sessions (30 minutes, 4 hours, and 8 hours) while being eye-tracked. The spatial distribution of attention during the study phase, measured as root-mean-square distances weighted by fixation duration, positively predicted memory accuracy independently of recognition sessions (i.e., no significant interaction). While the forgetting rate tends to decrease across recognition sessions, we observe it not to be perfectly linear. Specifically, we observed a rapid decline in memory accuracy between the initial two intervals (30 minutes and 4 hours), followed by a marked stabilization and even a slight improvement between the second and third intervals (4 and 8 hours). This pattern challenges the traditional linear forgetting curve and suggests a more complex decay process. Overall, these findings provide novel insights into the dynamic nature of forgetting and the role attention plays in preventing memory decay.

Acknowledgements: This work has been supported by NextGeneration EU under Piano di Ripresa e di Resilienza(PRIN 2022) with grant agreement (2022APAFFN; CUP B53D23014480001) awarded to MIC