Decoding episodic autobiographical memory in naturalistic virtual reality

Abstract

Abstract Episodic autobiographical memory (EAM) is a long-term memory system of personally experienced events with their context – what, where, when – and subjective elements, e.g., emotions, thoughts, or self-reference. EAM formation has rarely been studied in a controlled, real-life-like paradigm, and there is no predictive model of long-term retrieval from self-rated subjective experience at encoding. The present longitudinal study, with three surprise free recall memory tests immediately, one-week and one-month after encoding, investigated incidental encoding of EAM in an immersive virtual environment where 30 participants either interacted with or observed specific events of varying emotional valences with simultaneous physiological recordings. The predictive analyses highlight the temporal dynamics of the predictors of EAM from subjective ratings at encoding: common characteristics related to sense of remembering and infrequency of real-life encounter of the event were identified over time, but different variables become relevant at different time points, such as the emotion and mental imagery or prospective aspects. This dynamic and time-dependent role of memory predictors challenges traditional views of a uniform influence of encoding factors over time. Current evidence for the multiphasic nature of memory formation points to the role of different mechanisms at play during encoding but also consolidation and subsequent retrieval.

Keywords

• Episodic autobiographical memory
• Episodic memory
• Virtual reality
• Prediction
• Emotion
• Self