Effects of Social Contextual Variation Using Partner Avatars on Memory Acquisition and Retention

TL;DR

The paper investigates how partner avatar design in an immersive VR lecturer modulates memory through environmental context dependency, testing reinstatement and multiple-context effects using a Tagalog–Japanese word-pair task. Using a between-subjects design, participants learned with either a constant or varied set of partner avatars across six VR trials (IS and RP1–RP5) and then completed a real-world retention test a week later (FT). Results showed a robust reinstatement effect: constant avatars improved acquisition during RP trials, but there was no clear advantage for retention or forgetting in the final test, indicating no strong multiple-context effect. Correlations with social presence suggested a potential link between perceived presence and context-dependent memory, warranting further exploration of avatar design to maximize VR-based learning outcomes. The study demonstrates that instructor avatars in IVEs can influence memory processing, offering a foundation for optimizing avatar-driven VR education while highlighting the need for higher-powered studies to parse context, presence, and learning dynamics.

Abstract

This study investigates how partner avatar design affects learning and memory when an avatar serves as a lecturer. Based on earlier research on the environmental context dependency of memory, we hypothesize that the use of diverse partner avatars results in a slower learning rate but better memory retention than that of a constant partner avatar. Accordingly, participants were tasked with memorizing Tagalog--Japanese word pairs. On the first day of the experiment, they repeatedly learned the pairs over six sessions from a partner avatar in an immersive virtual environment. One week later, on the second day of the experiment, they underwent a recall test in a real environment. We employed a between-participants design to compare the following conditions: the varied avatar condition, in which each repetition used a different avatar, and the constant avatar condition, in which the same avatar was used throughout the experiment. Results showed that, compared to the constant avatar condition, the varied avatar condition resulted in significantly lower recall performance in the repeated learning trials conducted on the first day. However, the avatar conditions showed no significant differences in the final recall test on the second day. We discuss these effects in relation to the social presence of the partner avatar. This study opens up a novel approach to optimizing the effectiveness of instructor avatars in immersive virtual environments.

Figures (6)

• Figure 1: Participants learned Tagalog-Japanese word pairs from a lecturer avatar within an immersive virtual environment. These images depict first-person perspective views of the participant, showing the partner avatar's actions and facial expressions during each phase of a learning trial: (A) during the avatar familiarization phase, (B) during word presentation and recall tests, and (C) at the conclusion of the trial. Note that the avatar remained constant within each trial and could change between trials.
• Figure 2: Participants performed an initial study trial, in which they learned 20 Tagalog--Japanese pairs spoken by the avatar, and five retrieval practice (RP) trials, in which they were tested and received feedback. We compared the learning curves during RP1--5 and memory retention during a final test conducted in a real environment one week later between two conditions: the constant avatar condition, where the same partner avatar was used for all six trials, and the varied avatar condition, where a different partner avatar was used for each trial.
• Figure 3: (a) Proportion of correct responses for RP1--5 and FT. The constant avatar condition showed significantly better recall during RP1--5 than the varied avatar condition. Error bars represent standard errors. (b) Forgetting ratio from RP5 to FT. We found no significant differences between the conditions.
• Figure 4: Questionnaire results. No significant differences were found between the conditions.
• Figure 5: Proximity. Participants approached significantly closer to the partner avatar on RP4 than RP1 and RP2. Error bars represent the standard errors. $^{**}$: p$<$ .01.