VR & Anxiety Subcorpus: Research Insights

Updated 14 December 2025

VR and Anxiety Subcorpus is a comprehensive collection of studies detailing how immersive VR environments are used to diagnose and treat diverse anxiety disorders.
Methodologies employ rigorous designs including randomized trials, psychometric scales, and physiological metrics to measure intervention efficacy.
Emerging trends emphasize personalized, adaptive VR scenarios with AI and biofeedback integration, enhancing clinical and subclinical anxiety management.

Virtual Reality and Anxiety Subcorpus

Virtual reality (VR)–mediated interventions have established a rich research area at the intersection of human–computer interaction, psychological science, and clinical practice. The “VR and Anxiety Subcorpus” comprises the technical, empirical, and discourse-oriented literature examining VR—both as a diagnostic tool and as an active intervention—for anxiety disorders, affect regulation, and anxiety-provoking scenarios. The corpus covers system architectures, experimental protocols, quantitative and qualitative outcomes, and methodological frameworks, with attention to both clinical and subclinical contexts, as well as user characteristics ranging from neurotypical to neurodivergent populations (Yamoah et al., 7 Dec 2025, Olatunji et al., 2022, Ppali et al., 3 Oct 2025, Brooks et al., 2 Oct 2025).

1. Scope and Breadth of the Subcorpus

The subcorpus encompasses studies utilizing VR for generalized anxiety disorder (GAD), social anxiety disorder (SAD), specific phobias (acrophobia, mysophobia, dental anxiety), public speaking anxiety, anxiety in neurodivergence (autism, ADHD), interview/test anxiety, and situational/acute stressors (e.g., earthquakes, MRI procedures). Both therapist-guided and self-guided VR therapies are represented (Graham et al., 29 Jan 2025, Hossan et al., 18 Nov 2025, Mao et al., 11 Oct 2025). Methodologies include randomized controlled trials (RCTs), experimental/within-subjects designs, qualitative co-design workshops, systematic and scoping reviews, and large-scale discourse/corpus-linguistics analyses (Yamoah et al., 7 Dec 2025, Olatunji et al., 2022).

A marked trend is the convergence of high-fidelity head-mounted display (HMD) setups (Meta Quest, Oculus Rift, HTC Vive Pro Eye) and consumer-grade/standalone platforms, with Unity3D/Unreal-powered environments and personalized, dynamically adjustable exposure scenarios. Research focus has shifted from simple graded exposure to encompassing relaxation, emotion regulation, educational priming, avatar embodiment, and AI-assisted decision support (Ppali et al., 3 Oct 2025, Skiers et al., 26 Aug 2025, Hawes et al., 2023).

2. Experimental Paradigms and Measurement Protocols

VR–anxiety studies apply rigorous psychometric and physiological instruments:

Psychometric Scales: State-Trait Anxiety Inventory (STAI), Positive and Negative Affect Schedule (PANAS), Depression Anxiety Stress Scales (DASS-21), Beck Anxiety Inventory (BAI), Social Phobia Inventory (SPIN), Fear of Negative Evaluation (FNE), and custom scenario-specific scales (MASI for interview anxiety, PRCA-24, SUDS).
Scoring Formulas:
- STAI-State: $S = \sum_{j=1}^{20} s_j\ (s_j \in \{1,2,3,4\})$
- DASS-Anxiety: $A=2\sum_{i=1}^7 a_i,\ a_i\in\{0,1,2,3\}$
- PANAS-PA: $PA = \sum_{i=1}^{10} x_i$ ; PANAS-NA: $NA = \sum_{j=11}^{20} x_j$
Physiological Metrics: Heart rate (HR), heart-rate variability (HRV), galvanic skin response (GSR), skin conductance (SCR), electroencephalogram (EEG β-power), fNIRS (prefrontal oxygenation), and eye-tracking (gaze fixation, pupil diameter)(Ni et al., 14 Oct 2025, Liu et al., 2023).
Statistical Analysis: Paired and independent t-tests, mixed ANOVA, repeated-measures ANOVA, Friedman’s test, regression modeling, effect size (Cohen’s $d$ ), generalized eta squared $(\eta^2_G)$ .
Design Features: Within- and between-subjects assignment (counterbalanced/Latin Square), randomization, control of potential order effects, and scenario manipulation (e.g., contamination cue, avatar realism, interview difficulty) (Hossan et al., 18 Nov 2025, Luo et al., 2023, Kim et al., 2023).

3. System Architectures, Design Principles, and Scenario Construction

Technical evaluations in the subcorpus include:

Immersion: HMDs such as Meta Quest 2/3, HTC Vive Pro Eye, and Oculus Rift DK2; real-time rendering pipelines (Unity 2021.x, UE4/UE5), with explicit reporting of hardware (resolution, refresh rate), spatial audio (binaural, location-based), and device integration (motion controllers, eye tracking, HRV sensors) (Brooks et al., 2 Oct 2025, Ppali et al., 3 Oct 2025).
Virtual Environments: High-fidelity static and interactive scenes—nature, office/classroom, clinical setups, bespoke gamified exposures, and metaphorical “mind mansions.” Environments can be passive (seated, guided visuals) or active (navigation, object manipulation, avatar embodiment, graded trial design).
Avatar Design: Manipulation of realism (cartoon, stylized, photorealistic), gender swapping/anonymity for bias and self-consciousness reduction (Liu et al., 2023, Kim et al., 2023). Co-design and user stakeholder input (children, parents, clinicians) inform pediatric anxiety solutions (Mao et al., 11 Oct 2025).
Gamification and Biofeedback: Real-time performance feedback (e.g., shooting scores, anxiety state–triggered progression), physiological data–driven adaptation, and reward structures (Hossan et al., 18 Nov 2025, Skiers et al., 26 Aug 2025).

4. Quantitative Outcomes: Efficacy, Effect Sizes, and User Experience

Multiple studies demonstrate statistically significant and clinically meaningful reductions in state anxiety and related constructs:

Study / Condition	Pre–Post Effect Size (Cohen's d)	Main Results
Acrophobia VRET (Donker et al. 2019)	$d=1.14$ –$1.8$	STAI $\Delta \approx -20$ , $p<.001$ (Olatunji et al., 2022, Graham et al., 29 Jan 2025)
SAD (Zainal et al. 2021; Lee et al. 2021)	$g=-4.17$ (MASI), $d\approx 0.8$	Anxiety $\downarrow$ , sustained 6mo post (Graham et al., 29 Jan 2025, Saeed et al., 2024)
Knowledge Worker “Tranquil Loom” (STAI-S)	$d\approx 0.82$	$t(34)=4.88$ , $p<0.001$ (Ppali et al., 3 Oct 2025)
MRI Patient MIRAGE (STAI-S)	$d=2.0$	$\sim$ 20% anxiety drop, SUS=83.5 (Brooks et al., 2 Oct 2025)
Self-guided VR (systematic review)	$d=0.4$ –$2.7$	Anxiety reduced across phobia types (Graham et al., 29 Jan 2025)
Group VR Interview (MASI)	See regression, direct realism–anxiety link	Lower realism $\rightarrow$ higher anxiety (Ni et al., 14 Oct 2025)

Several studies report that brief exposures (1–5 sessions, 10–30 min each) suffice for acute state-anxiety reduction, especially in high-immersion, scenario-relevant, and personalized VR configurations. Gamified exposures and graded difficulty outperform generic relaxation scenes in effect magnitude (Hossan et al., 18 Nov 2025, Rasch et al., 2024).

Usability and presence scores, where reported (SUS, IPQ), are high, supporting real-world acceptability. Attrition rates remain low (mean 16%); low rates of cybersickness and adverse events are documented (Graham et al., 29 Jan 2025, Skiers et al., 26 Aug 2025). Qualitative response patterns emphasize agency, autonomy, playful exploration, and self-pacing as critical to engagement and relief (Ppali et al., 3 Oct 2025, Skiers et al., 26 Aug 2025).

5. Thematic and Contextual Findings: Societal Discourse, Subgroup Effects, and Design Implications

Corpus linguistics analysis shows that “VR,” “Oculus,” “headset,” and platform terminology dominate online anxiety–VR discourse, with prepositional phrase analysis marking points of experience (“in virtual reality”), design/process (“of virtual reality”), and application (“for virtual reality”). Device ergonomics and immersive content remain central concerns; user agency and system accessibility surface as priorities for clinical and non-clinical adoption (Yamoah et al., 7 Dec 2025).

Studies targeting children (dental anxiety), non-native English speakers (SAD), women and non-binary neurodivergents, and multilingual cross-cultural users highlight:

The importance of customizability and safety (e.g., session duration limits to prevent over-immersion in pediatric populations).
Multisensory, agency-supportive environments that balance low sensory load with optional, user-driven interactivity (Skiers et al., 26 Aug 2025, Mao et al., 11 Oct 2025).
Caution against avatar designs triggering the “uncanny valley”; cartoon-like avatars minimize state anxiety in ESL users, while high realism elevates anxiety markers (HRV, GSR) (Liu et al., 2023).

Design guidelines from recent studies converge on (1) adaptive sensory controls, (2) minimized locomotion, (3) graded, scenario-relevant exposure, (4) biofeedback integration, and (5) personalization at both content and avatar/embodiment levels (Skiers et al., 26 Aug 2025, Ni et al., 14 Oct 2025, Hossan et al., 18 Nov 2025).

6. Methodological and Reporting Considerations

Research in the subcorpus reveals several methodological gaps and recommendations:

Underrepresentation of children/adolescents and neurodevelopmental populations remains prevalent (Graham et al., 29 Jan 2025, Olatunji et al., 2022).
Long-term follow-up (beyond 6–12 months) is rare, with limited data on durability of therapeutic effects (Graham et al., 29 Jan 2025).
Variability in standardized user experience reporting and absence of physiological monitoring in many trials (Graham et al., 29 Jan 2025, Saeed et al., 2024, Olatunji et al., 2022).
Calls for pre-registered RCTs, comprehensive statistical reporting (t/F values, $d$ , confidence intervals), use of active comparators, and hybrid clinician/self-guided protocols to optimize efficacy assessment and generalization (Saeed et al., 2024, Graham et al., 29 Jan 2025, Olatunji et al., 2022).

Open questions include protocol optimization for specific subgroups (e.g., trauma history, LGBTQ+ neurodivergence), harmonization of hardware/software standards, ethical management of anonymity and data privacy, and precise calibration of scenario realism to user sensitivity.

7. Future Research Directions and Open Issues

Convergent evidence positions immersive, adaptive VR as a robust medium for anxiety reduction via both exposure and relaxation/emotion regulation paradigms. Outstanding research tasks in the subcorpus include:

Algorithmic and LLM-driven personalization of scenario difficulty and motivational priming (Ppali et al., 3 Oct 2025, Hawes et al., 2023).
Integration of real-time physiological and behavioral analytics to guide exposure and optimize individual outcomes (Skiers et al., 26 Aug 2025, Brooks et al., 2 Oct 2025).
Comparative effectiveness trials vs. gold-standard CBT, internet interventions, and bibliotherapy (Graham et al., 29 Jan 2025, Olatunji et al., 2022).
Expansion into longitudinal monitoring, booster sessions, and real-world transfer of anxiety management skills.
Inclusive design for ostracized groups (neurodivergent, gender-diverse, multilingual) and sustained evaluation in non-clinical settings (workplace, public health, education).

By synthesizing technical, empirical, and discourse-layer findings, the VR and Anxiety Subcorpus delineates a maturing field prioritizing rigorous measurement, user-centered design, and practical translation of immersive systems into diverse psychoeducational and clinical contexts.