Understanding the Effects of Interaction on Emotional Experiences in VR

Paper Title

Understanding the Effects of Interaction on Emotional Experiences in VR

Publication Info

• Topic area: Emotional experiences in Virtual Reality (VR) through interaction.
• Keywords: Virtual reality, emotional elicitation, interaction design, affective computing, physiological measures, self-assessment, immersive environments, emotion regulation, user engagement, object-level interaction.

Background and Problem

• Problem / challenge: Existing VR emotion elicitation research primarily focuses on passive stimuli (e.g., visual and auditory), with limited exploration of how interaction influences emotional experiences. Object-level interaction's role in modulating emotions remains underexplored.
• Significance: Understanding interaction's impact on emotions can guide the design of emotionally engaging VR environments, with applications in mental health, education, and training.
• Motivation and related work: Prior studies have shown that VR can elicit emotions effectively, but few have systematically investigated the effects of interaction. Object-level interaction has shown potential in enhancing immersion, ownership, and user satisfaction, but its emotional impact is unclear. This paper addresses these gaps.

Solution

• Proposed approach: Extend an existing VR emotion-elicitation dataset by adding a high-arousal, high-valence (HAHV) scene and embedding object-level interaction across six VR scenes.
• Novelty:
  1. Creation of a new HAHV scene (Surrounded by Elephants) to fill a gap in the affective space.
  2. Systematic integration of object-level interaction into VR scenes for controlled comparisons.
  3. Multimodal evaluation combining self-reports and physiological measures to analyze emotional responses.
• Procedure and key techniques:
  • Developed interactive and non-interactive versions of six VR scenes targeting different quadrants of the valence-arousal model.
  • Conducted two studies: one validating the HAHV scene and another analyzing the effects of interaction on emotional responses.
  • Measured emotions using the Self-Assessment Manikin (SAM) questionnaire, physiological signals (e.g., heart rate variability, skin conductance), and qualitative interviews.

Results

• Concrete findings:
  • The Surrounded by Elephants scene reliably elicited HAHV emotions, with higher valence and dominance compared to its 360° video counterpart.
  • Interaction increased dominance across all scenes, with scene-dependent effects on valence and arousal.
  • Physiological data showed higher heart rate variability (HRV) and lower heart rate (HR) in interactive conditions, indicating relaxation interspersed with moments of high arousal.
• Advantage over baselines:
  • Interactive scenes extended engagement time and encouraged broader spatial exploration compared to non-interactive versions.
  • Interaction provided coping mechanisms in negative scenes (e.g., Tunnel, Solitary Confinement) and enhanced enjoyment in positive scenes (e.g., Puppies, Jetty at Lake).
• Experiments / evaluation:
  • Study 1: Within-subject design (N=24) validated the HAHV scene.
  • Study 2: Mixed-design study (N=84) compared interactive and non-interactive conditions across six scenes using SAM, physiological data, and interviews.
  • Metrics included valence, arousal, dominance, engagement time, spatial exploration, and physiological measures (HRV, HR, SCR).
• Limitations and future work:
  • Focused on object-level interaction; more complex forms (e.g., social interaction) were not explored.
  • Relied on self-reports and basic physiological signals, limiting fine-grained insights.
  • Future work should integrate multimodal affect recognition and explore additional interaction modalities (e.g., haptic feedback).

Summary

This paper investigates how object-level interaction in VR modulates emotional experiences. By extending an existing dataset with a new HAHV scene and interactive elements, the authors demonstrate that interaction enhances user engagement, increases dominance, and modulates valence and arousal in context-dependent ways. Multimodal evaluations reveal that interaction fosters relaxation while enabling moments of high arousal, supporting both emotion elicitation and regulation. These findings advance affective computing research and provide practical insights for designing emotionally resonant VR applications in mental health, education, and training.