This paper quantitatively investigates the feasibility of, and merit in, soliciting continuous subjective cybersickness ratings as participants passively engage in an immersive VR experience. The main research questions addressed are: (1) Feasibility: To what extent will participants successfully engage, unprompted, in continuous cybersickness reporting while engaging with a secondary task? and (2) Merit: To what extent do continuously reported subjective cybersickness ratings offer valuable insights beyond what can be obtained from less frequent querying? Participants used a physical slider device, in conjunction with discreet visual feedback, to continuously report their instantaneous motion sickness state as they rode nineteen consecutive rounds of a virtual roller coaster ride and performed a simple visual counting task. We analyzed the reported sickness ratings in the context of pre-post SSQ scores, #rounds endured before quitting, tonic skin conductance levels (SCL), optical flow of the visual stimulus, and rotational and translational velocity of the virtual viewpoint, as well as in comparison to previously-obtained data from different participants who underwent the same exposure but only verbally reported a single FMS score at the end of each round (every 65s). We found that most participants used the slider actively, and that, averaged across participants, the reported sickness scores not only increased over time but also varied up and down in conjunction with the intensity of the ride. We found a statistically significant positive correlation between instantaneous reported sickness levels and tonic electrodermal activity in 76% of our participants, as well as a statistically significant positive correlation with optical flow magnitude and viewpoint rotational velocity. We observed no significant differences in #rounds completed, Delta SSQ scores, and average maximum or average last-reported sickness levels between the continuous and discrete reporting groups. Altogether, our results (1) demonstrate the feasibility of collecting valid self-reported ratings of cybersickness on a continuous basis during a passive VR experience, and (2) suggest that such data has the potential to be useful for better understanding cybersickness evolution in the context of potentially transient triggers.
Virtual reality (VR) offers benefits such as enhanced interest, engagement, and vocabulary retention when learning another language. Due to the immersive nature of VR, users can enter a virtual setting that provides context clues as to what a particular vocabulary word in the target language represents. Users can also interact with the object to help increase the potential of being able to retain the term in the target language for the object through association with its surroundings. While the influence of using contextually-relevant virtual environments and providing object interaction have been studied separately, the aim of this research is to identify whether the combination of object interaction and a context-relevant environment is acceptable to users and effective at promoting short-term retention for basic vocabulary related to everyday objects.
This work presents FlowZone, an exploratory VR prototype that integrates yoga and mindfulness meditation. The system uses headset and controller data for lightweight pose tracking and provides real-time feedback in guided sessions. A preliminary user study (N=6) revealed common challenges in yoga practice, including uncertainty about pose correctness, difficulty maintaining consistency, and the importance of calming environments. These insights informed the design of FlowZone, which combines accessible pose guidance with immersive, meditation-oriented settings. While not a full evaluation, our early findings suggest that VR yoga and meditation can lower barriers to practice and support stress reduction, pointing toward promising directions for future research.
Notifications are a fundamental aspect of daily computing, whether on desktops, laptops, smartphones, or smartwatches. On average, adults receive around 200 notifications per day-approximately one every five minutes during waking hours. As Extended Reality (XR) headsets advance, they may become the primary medium for digital interactions, making notification management a crucial factor in their usability. While notifications are known to be disruptive on smartphones, their impact could be even more pronounced on headworn devices. To investigate this, we conducted an exploratory five-day study with eight participants wearing display-equipped smartglasses that delivered notifications from their smartphones. Participants used the glasses throughout their day for at least 2 hours receiving on average 62% of all notifications on the glasses, submitted daily journal entries, and participated in post-study interviews. We also logged notification sources and timestamps throughout the study. Our findings reveal both practical advantages and significant challenges of head-worn notification delivery. While participants appreciated the convenience and immediacy of glanceable alerts, concerns about privacy, social acceptability, and distraction emerged as key barriers to adoption.
Enhancing user experience and performance, including task load in immersive environments, requires accurate prediction of user gaze point, reaction time, and mental and physical load uptake. Current gaze prediction approaches focus primarily on motionbased information, lacking physiological data, which leads to poor prediction accuracy in highly dynamic virtual reality (VR) environments. Traditional cognitive load measurements rely on posttask analysis without proper multimodal data integration and fail to capture the real-time dynamics of user states during interaction. Likewise, reaction time or attention load are often assessed only after the interaction, without using real-time immersive sensor data, which limits adaptive responsiveness. To tackle these limitations, we leveraged a comprehensive multimodal dataset VRWalking, which recorded timestamped eye-tracking metrics, physiological signals (heart rate and galvanic skin response), and behavioral performance data during real-time engagement in a VR environment. We developed a unified multitask model based on the MultiPatchFormer architecture, which processes multimodal VR signals through dual patch projection branches for gaze and classification inputs. The model employs multiscale patch embeddings, cross-attention between gaze and classification pathways, channel attention, and transformer encoders to jointly predict continuous user gaze and classify reaction time, cognitive load (mental load and physical load). Our methodology achieved excellent predictive performance: 95.64% for reaction time, 98.01% for mental load, and 97.45% for physical load, with a MAPE (Mean Absolute Percentage Error) of 15.24% for gaze prediction. We applied Shapley Additive explanations (SHAP) analysis to interpret the model's behavior across all features, including eye-tracking, head-tracking, and physiological signals. The analysis revealed which features most influenced the predictions of user gaze, reaction time, mental load, and physical load. Our methods, while based only on the VR-Walking dataset, demonstrated strong performance across all tasks, suggesting promising potential for real-world VR applications such as interactive training systems that respond to user attention lapses, educational platforms that adapt to cognitive load, and performance assessments that consider physiological indicators.
Virtual reality (VR) often induces perceptual distortions, for instance, underestimates of perceived distance which may distort perceived size. However, other depth cues are available in these immersive environments that may help users perceive size accurately. Here, we examined how binocular and monocular depth cues, as well as object interaction, influenced the perceived size of 3D shapes in VR. The results showed that binocular cues play an essential role in size judgements, whereas dynamic monocular cues, such as motion parallax, did not fully compensate for their absence, underscoring the importance of rendering high-quality binocular cues for fine size discrimination tasks in virtual environments.
Virtual Reality (VR) enables users to embody avatars with vastly different appearances and anatomies. Embodying virtual spiders with their alien morphology could offer exciting experiences for immersive VR in gaming or education. While prior research has explored embodiment of human-like avatars and even non-human forms such as animals, it still remains unclear how best to control anatomically distinct avatars such as spiders. In this exploratory study, we systematically compared four control methods-standard VR controller, hand control, half-body control, and full-body control-while embodying a spider in VR. Using a repeated-measures design with 20 participants, we assessed each control method in terms of embodiment, usability, and perceived exertion. Results indicate that half-body control offered the best overall balance, with the highest usability and lowest exertion, while still maintaining a comparable level of embodiment to other methods. Full-body control was rated significantly lower in usability and higher in perceived exertion. These findings suggest that half-body control may provide a good balance between realism and usability for embodying spiders in VR.
Studying bee behavior and diagnosing hive issues in their natural habitats has often been a challenge due to limited visibility, observer interference, and limitations of conventional recording methods. A teleoperated immersive 360(circle) view offers an unobstrusive, holistic observation within the hive. In this work, we present a framework for capturing and analyzing bee activity using commercially available 360(circle) cameras and hardware to create an immersive VR experience. A video demo is available at: https://youtu.be/96pqv9AyRlo.
We propose SnapSteer, a bimanual 3D manipulation interface using common VR controllers, which allows restriction of motion degrees of freedom (DoFs) as needed. This interface is based on the conventional one-handed 6-DoF manipulation interface called Robot Telekinesis, and assigns the other hand the role of controlling whether and in which direction the DoFs are restricted. This enables users to quickly switch between unconstrained 6-DoF operation and precise 1-DoF operation according to the task. We designed and implemented a prototype of this interface in VR, and conducted a user study (N=12) comparing its performance in a straight 3D steering task with two baseline interfaces (i.e., a 6-DoF individual control interface and Robot Telekinesis). The results showed that our interface outperformed the other two in task efficiency. On the other hand, there was no significant difference in subjective workload or usability compared to Robot Telekinesis, and which derives discussion of improvements to visual feedback during the direction adjustment phase.
Providing users with realistic sensations of object stiffness in virtual environments remains challenging due to the intricacies of our haptic sense. We investigate the use of a visuo-haptic illusion to alter the perceived stiffness of hand-held objects in virtual reality. We manipulate the Control-to-Display ratio of the index finger and thumb movements during pinching to make virtual objects feel softer or harder. We evaluated this approach on a variety of haptic representations and visualizations we selected through a pre-study survey (N=24). Results of our user study (N=20) demonstrate that this method effectively and reliably modifies stiffness perception, bridging gaps of 50% in physical stiffness without adversely affecting the visuo-haptic experience. Our findings offer insights into how different visual and haptic presentations impact stiffness perception, contributing to more effective and adaptable future haptic feedback systems.
This paper presents an AI-assisted VR conference application with multilingual translation and navigation agent capabilities. A pilot study with 18 participants (11 females, 7 males) was conducted to assess the system's usability. AI-assisted navigation worked smoothly, but the AI translation had issues that prevented the users from having a good experience, nonetheless, participants expressed positive attitudes toward the system, and future work will focus on achieving better user experience.
Visually searching for objects is an everyday task. In many contexts, people must visually search for multiple objects at the same time while avoiding distractor objects, such as triage during a mass casualty incident. While many prior augmented reality (AR) and virtual reality (VR) studies have investigated cues to aid in visual search tasks, few have investigated cues in contexts involving multiple targets and distractors with a full 360 degrees effective field of regard (EFOR). Individually, multiple targets, distractors, and a full 360 degrees EFOR each add complexity to visual search; when combined, they compound the difficulty even further. In this paper, we present such a study that compares three common types of visual cues (2D Wedge, 3D Arrow, and Gaze Line) to a baseline condition with no cueing for a 360 degrees visual search task. Our results reinforce the importance of providing some type of cue, with the Gaze Line design being particularly beneficial. We discuss the potential implications of these findings for designing cues specifically for such complex visual search tasks.
The goal of this study is to support efficient and consistent three-dimensional (3D) painting in virtual reality (VR) space. For this, we propose the use of deformable primitives as a construction guide in the early stages of 3D painting and introduce FlexiPrim, consisting of two primitive types, cuboid and ellipsoid. The user can deform FlexiPrim by manipulating the control points in VR space. We conducted a user study to compare FlexiPrim with traditional Guide-Objects in OpenBrush, where the latter has an automatic snapping function. As a result, we found that the proposed method enables the creation of construction guides consisting of curved elements, and participants provided positive feedback on the subsequent rough sketch using the FlexiPrim as guide.
Inspired by behavioral biometrics for keystroke and touch-based systems, a large body of work has emerged over the past decade on using user behavior in VR applications as a signature of the genuine user. Recent work on forecasting approaches for behavioral biometrics for VR helps address a key challenge in existing approaches where complete user movement signatures are needed to authenticate the user. Forecasting-based approaches enable VR authentication systems to use limited user behavior data and forecast future movement trajectories. However, forecasting-based approaches present a new concern where malicious users can exploit the predictability of user motions to launch an attack. In this paper, we present the first forecasting-based attack model against VR authentication systems that rely on behavioral biometrics. We propose a two-phase approach to assess authentication performance and adversarial risk. Phase 1 develops a Fully Convolutional Network for authentication using VR motion data, evaluating stochastic gradient descent (SGD) and Adam optimizers with Equal Error Rate (EER) as the primary metric. Phase 2 introduces a forecasting attack, where partial motion sequences of an impostor's motion are fed to a Transformer model to generate future trajectories that represent genuine user behavior for an authenticator enabling an impostor to deceive the authentication system. Experimental results demonstrate the attack's effectiveness, achieving an EER as low as 0.0346, exposing security risks in motion-based authentication. These findings underscore the urgent need for robust countermeasures to defend against predictive motion attacks in VR environments. Our code is shared at: https://bit.ly/4n1GtxG.
Facial and vocal expressions are vital to conveying emotions and regulating social interaction. In online interactions, avatars are used as proxies for emotional expression, but most avatars do not allow for facial expressions to change in line with the user's conveyed emotions, creating a challenge for naturalness and conversational satisfaction. This demonstration presents a virtual reality application developed for the Meta Quest headset where participants control avatars with four distinct facial expressions (happiness, sadness, anger, and fear) created based on the Facial Action Coding System. Real-time voice communication allows for the coordination of vocal tone with the avatar's facial expression, enabling the study of matched and mismatched emotion cues in conversation. This demonstration highlights how streamlined emotional cues can impact conversational satisfaction of virtual environment interactions without requiring highly detailed avatars.
With the emergence of consumer-grade mobile Augmented Reality (AR) head-mounted displays (HMDs), it is increasingly important to study interaction modalities in mobile, outdoor contexts to explore user interactions under real-world conditions and ensure ecological validity. We present findings from a user study examining how state-of-the-art interaction techniques, originally developed for stationary indoor use, perform in outdoor walking contexts. We compare three techniques using a Fitts' law task: (1) eye gaze for target selection combined with a pinch gesture for activation, (2) hand-controlled raycasting for selection with pinch activation, and (3) a go-go-hand-inspired technique that enables users to reach distant targets via virtual touch with a virtual hand representation. We report both performance metrics and subjective user feedback for these techniques.
XR headset users often lose environmental awareness, resulting in unintentional collisions and social boundary violations. We present XR Marionette, a wearable system that applies progressive constraints on arm movement (none -> elbow-to-waist -> elbow-to-waist + wrist-to-shoulder), while switching XR interaction methods (grab, ray-casting, and microgestures) based on proximity to surrounding obstacles and people. The system adapts to both solo and social contexts with context-dependent distance thresholds, promoting safety, social harmony, and sustained XR engagement through proactive constraints rather than reactive warnings.
Indigenous Virtual Learning Spaces (IVLS) is a VR platform that applies Indigenous pedagogy through immersive, guided learning experiences. Developed using Unity and 3D models from the Ganondagan State Historic Site, it enables learners to explore historically informed Haudenosaunee environments led by a guide. User testing with faculty and students highlighted the ecological realism, cultural depth, and curricular potential of the system, while also noting areas for usability improvement, such as navigation and audio balance. IVLS shows strong promise as a cross-disciplinary educational tool, fostering deeper connections to land, culture, and history.
The memory orb is a cylindrical input device for virtual reality, inspired by a science-fiction artifact. This paper extends previous work by presenting a use case, interactions, and a user study directly grounded in a movie scene, guiding the design of hovering, selecting, translating, rotating, and scaling 3D objects. Results show performance comparable to traditional virtual reality controllers and positive user reception, supporting further exploration of this novel, fiction-inspired interface.
Cybersickness remains a persistent challenge in immersive media. Prior work suggests that rendering a virtual nose can reduce cybersickness by providing a stable visual reference. However, if it introduces unwanted perceptual side effects such as changing odor perception or reducing the sense of presence is currently unknown. We conducted a study with 22 participants to examine whether a virtual nose influences odor perception or sense of presence in mixed reality. Except for perceived familiarity of the odor, results provide moderate evidence that rendering a virtual nose does not differ from not rendering one. This indicates that a virtual nose does not significantly influence presence and odor sensitivity. However, future work should further explore how a virtual nose influences perceived familiarity of odors to learn more about perceptual side effects.
