Understanding how intelligent systems can adaptively support human spatial cognition represents a critical challenge in human-computer interaction. We investigate a behaviour-driven adaptive feedback mechanism that leverages Large Language Models (LLMs) to provide contextually aware guidance during 3D spatial tasks. Our approach monitors real-time user behaviours - including hesitation patterns, repeated errors, and strategy changes - to trigger personalised natural language feedback aligned with users' cognitive states. In a within-subjects experiment (N = 30), participants completed three Soma cube assembly tasks of equivalent difficulty (3-star rating), one per feedback condition: no feedback, request-based feedback, and adaptive feedback. We measured user trust, perceived feedback appropriateness, reflective engagement, and cognitive load using validated scales, alongside behavioural indicators. Adaptive feedback significantly enhanced user trust (r = .370, p = .004), perceived appropriateness (r = .260, p = .046), and reflective engagement (r = .310, p = .017) compared to request-based feedback. Mediation analysis revealed that trust fully mediated the relationship between feedback strategy and feedback appropriateness (98.2% of total effect). Behavioural analysis showed reduced backtracking in feedback-enabled conditions. These findings demonstrate how LLMs can serve as intelligent intermediaries that enhance human cognitive processes while maintaining user autonomy, offering practical guidance for designing adaptive support systems in educational and training environments.
Inadequate sleep is a health concern and a common college student experience. Recent studies have found social media addiction to be increasingly associated with poor sleep. Bedtime procrastination (BP), an under-researched sleep-related behaviour in which individuals delay going to bed, has also been shown to impact sleep. However, research examining the association between maladaptive use of social media, sleep quality, and BP among minoritised college students in the U.S. is scarce. Thus, this cross-sectional study explored the degree to which problematic social media use (PSMU) is associated with sleep quality via BP among 233 undergraduate students (77% female; avg. age = 23.1; > 64% racial/ethnic minorities) from a U.S. minority-serving institution. Because prior studies have found BP to be associated with inattention, we examined whether mindfulness moderates the association between PSMU and BP. Mediation analysis indicated that BP had a significant role in the association between PSMU and sleep quality. Moreover, mindfulness and the mindfulness facet acting with awareness attenuated the relationship between PSMU and BP. Our results extend prior research by demonstrating similar trends among a diverse student sample comprised of mostly minoritised students and highlights mindfulness as a practice suitable for bringing attention to an individual's pre-sleep behaviours.
While virtual reality (VR) technology promises enhanced decision support through immersive visualisation and rich interaction, empirical evidence remains limited. Through a 2 & times; 2 between-subjects experiment with a control condition (N = 183), we examined whether 3D VR environments provide advantages over traditional 2D interfaces and how the core features of VR, immersion and embodiment, affect individuals' decision-making experiences in a financial context. Specifically, we analyzed decision satisfaction (process-focused) and self-esteem and anxiety (outcome-focused, from both approach and avoidance perspectives). In addition, both internal and external locus of control were tested as moderators. Key findings reveal: (1) the 3D VR environments did not significantly enhance or impair decision-making experiences compared to the conventional (non-immersive) 2D environments; (2) regarding VR, high immersion marginally reduced decision satisfaction and self-esteem; (3) embodiment alone had no significant effects on any dimensions of decision-making experiences, while high embodiment reduced the negative effect of external locus of control on satisfaction. These findings advance our understanding of how VR features interact with individual differences to empower decision-making requiring sophisticated information processing, filling a research gap by focusing on the features of VR rather than treating VR as a monolithic technology.
Assessing emotion regulation using traditional methods, such as questionnaires and observational tools, can be time-consuming, expensive, and prone to bias. AI-powered facial expression analytics holds promise in offering a more efficient and objective method for assessing emotion regulation - a cornerstone of mental wellbeing. However, the feasibility and reliability of using facial analytic software for treatment planning and patient monitoring remain unclear. This study addresses this knowledge gap by examining the clinical validity and accuracy of iMotions 10 AFFDEX SDK 5.1 in coding negative-expressed affect in autistic children. Fifty autistic children aged 7-12 years (44 males, 6 females) participated. Coding was performed on video footage of participants playing a video game in their homes for 5 min, followed by a 2-minute enjoyable task. Results indicated a strong correlation between the negative affect ratings of two human raters (r = .99, p < .001), and weak and non-significant correlations between the human raters and the AI algorithm (rs < .18, ps > .23). Analysis of signal data from AFFDEX SDK 5.1, when synchronised with participants' video footage, suggested the software lacked sensitivity and misclassified anxious or frustrated facial grimacing as positive affect (smiling). These findings highlight the need for further validation of AI-driven technologies with specific clinical populations.
Wearable fitness technology has transformed consumer engagement with health monitoring, yet factors influencing user satisfaction and dissatisfaction remain underexplored. This study employed a mixed-methods approach, which integrated text mining of 50,000 Fitbit user reviews with survey-based psychometric validation. Text analysis techniques, including sentiment analysis, topic modelling, co-occurrence networks, and word embeddings, identified key satisfiers (e.g. tracking accuracy, motivation) and dissatisfiers (e.g. syncing issues, software malfunctions). A structured questionnaire was developed and administered to 683 Fitbit users, with exploratory and confirmatory factor analyses ensuring measurement validity. The empirical findings contribute to wearable technology research by offering a validated, empirically grounded scale for assessing consumer experience. This study provides actionable insights for improving wearable device functionality and enhancing user retention through data-driven design and customer service improvements.
Privacy policies function as both legal documents and information sources for users, but their length and complexity often discourage engagement. In this paper, we investigate whether a personalised approach can address this issue by prioritising information that concerns individual users most while maintaining a policy's legal compliance on disclosure. We first explored whether personal characteristics can be used to predict a person's most concerned category and, hence, serve as a baseline for personalisation. We then conducted an eye-tracking experiment and interviews (n = 30) to understand the effectiveness of personalised reordering of privacy policies. In the interviews, many participants perceived personalised reordering as helpful, although others raised concerns about the invasion of privacy through this personalisation. The eye-tracking results indicate that personalised reordering leads to higher engagement for the first few sentences of a privacy policy. Based on our findings, we present design recommendations for creating legally compliant forms of privacy disclosures that encourage user engagement as well as discussions and implications on privacy disclosure compliance.
Automation Surprise (AS) initially refers to moments when automated systems behave unexpectedly to the user, causing confusion and potentially leading to catastrophic consequences. While previous research primarily studied AS with safety-critical considerations, little is known about how users encounter and respond to such surprises during non-critical automated driving. To address this gap, we adopt an experience-centred lens to understand AS that builds on the pragmatist approach to experience in HCI and the sensemaking model of AS. Following design research traditions, we conducted a probe-based inquiry with a Wizard-of-Oz vehicle as an experience prototype on real roads, inviting participants to imagine being in an automated vehicle while engaging in non-driving-related activities. Through reflexive thematic analysis of think-aloud during the ride and post-hoc interviews, we identified AS as Direct and Indirect, and developed a conceptual categorisation along two dimensions: the nature of the surprise (positive to negative) and in-situ response level (high to low). Viewing AS through a pragmatist perspective on experience further demonstrates how AS is dynamic, situated, and relational. Our findings expand current understandings of AS beyond safety concerns to capture its experiential aspects, to inspire future experience-centred research and design in non-critical automated driving.
