Generative Multi-Sensory Meditation: Exploring Immersive Depth and Activation in Virtual Reality

TL;DR

The paper introduces MindfulVerse, an AI-generated, multimodal VR meditation framework designed to personalize mindfulness experiences across diverse users. By building a mindfulness agent and an end-to-end content pipeline, it dynamically generates environment, music, and guided prompts in response to user input, and evaluates neurophysiological and experiential outcomes with a within-subject fNIRS study (N=30). Results show integrative multimodal meditation robustly activates self-regulation brain regions and improves emotional regulation and immersion compared with unimodal or static content, highlighting the value of cross-modal coherence for VR wellness. Limitations include latency and fixed interaction, while future work aims to enable real-time neuroadaptive content and longitudinal efficacy assessments to scale personalized VR mindfulness.

Abstract

Mindfulness meditation has seen increasing applications in diverse domains as an effective practice to improve mental health. However, the standardized frameworks adopted by most applications often fail to cater to users with various psychological states and health conditions. This limitation arises primarily from the lack of personalization and adaptive content design. To address this, we propose MindfulVerse, an AI-Generated Content (AIGC)-driven application to create personalized and immersive mindfulness experiences. By developing a novel agent, the system can dynamically adjust the meditation content based on the ideas of individual users. Furthermore, we conducted exploratory user studies and comparative evaluations to assess the application scenarios and performance of our novel generative meditation tool in VR environments. The results of this user study indicate that generative meditation improves neural activation in self-regulation and shows a positive impact on emotional regulation and participation. Our approach offers a generative meditation procedure that provides users with an application that better suits their preferences and states.

Figures (7)

• Figure 1: (a) The MindfulVerse interaction pipeline. Users verbally express their mental state or intention, which drives the generation of personalized meditation content across three dimensions: environment, music, and guided prompts, forming an adaptive VR mindfulness experience. (b) Experimental protocol using fNIRS. Participants engage with the generated meditation in immersive VR while whole-brain hemodynamic activity is recorded.
• Figure 2: Procedure of the study. (a) User‑study workflow. (b) Initial setup/entry state. (c) In‑system questionnaire completion. (d) Task buffer/transition scene.
• Figure 3: Timeline and task sequence for participant (ID: M11) generative mindfulness tasks in this study. The session begins with a non-VR-based baseline (form a-1 to a-2), followed by (from a-3 to a-6) randomized generative mindfulness tasks interleaved with instrument calibration and rest phases.
• Figure 4: Surface projection of the main effect of four tasks in L/R mark hemispheres. The one-way repeated measures ANOVA was performed on the four meditation tasks.
• Figure 5: PANAS‑SF results. Panels (a) and (b) show PA and NA changes from baseline to each meditation task; panel (c) shows item‑level comparisons.