Research Appraisals
Evidence-based critical appraisals of the latest medical research, systematically evaluated using Oxford CEBM methodology.
Showing 23 appraisals
Proceedings of the National Academy of Sciences of the United States of America
Large language models accurately identify decision reasons in verbal reports.
Understanding the reasons behind human choices under risk is a central goal of decision scientists, but traditional methods relying on behavioral data are limited by strict invariance assumptions. We introduce a scalable analytical framework using large language models (LLMs) to analyze verbal reports and identify articulated reasons for choice between monetary lotteries. A validated LLM accurately identified predefined decision reasons in participants' free-text reports, aligning with their actual choices in 95% of trials. Our analysis reveals that the reasons behind people's decisions vary systematically and are driven more by the structure of the choice problem than by individual differences. Crucially, reasons identified from verbal reports yield more parsimonious and informative representations of decision processes compared to those inferred from choices alone; furthermore, problem-specific reason profiles achieve out-of-sample prediction accuracy that is competitive with established computational models. This work demonstrates that verbal reports are a rich data source and our analytical framework can unlock their potential, delivering results that challenge the field's foundational invariance assumptions and pave the way for more context-sensitive and interpretable models of human decision making.
1 July 2026
Read appraisal →Proceedings of the National Academy of Sciences of the United States of America
AI agents are sensitive to nudges.
Large language models (LLMs) are increasingly deployed as autonomous agents that make choices and use tools on behalf of users. Yet, we have limited evidence about how their decisions are shaped by their environment. We adapt a human decision-making task to test leading LLMs under four forms of choice architecture: defaults, suggestions, information highlighting, and "optimal" nudges derived from a resource-rational model of human choice. We treat human behavior as a baseline for predictable sensitivity to such interventions. Across models and prompting strategies, LLMs often depart substantially from this baseline. They sometimes pay excessive costs to acquire information, sometimes ignore available information, and, most crucially, are far more responsive to nudges than humans, such that weak cues that slightly shift human behavior have larger effects on model choices, toward both better and worse payoff outcomes. Chain-of-thought prompting and in-context human data do not reliably stabilize behavior. Recent reasoning-optimized LLMs can, in some configurations, restore more human-level sensitivity to nudges, but do so inconsistently and at substantial computational cost. These results point to an important and largely neglected safety concern: LLM agents can be behaviorally brittle under subtle changes in choice architecture, even in the absence of adversarial settings.
23 June 2026
Read appraisal →Blood advances
A pilot randomized clinical trial of a peer support intervention for hematopoietic stem cell transplantation
Peer support has been associated with improved patient-reported outcomes (PROs), including reduced psychological distress, among patients undergoing hematopoietic stem cell transplantation (HSCT), but structured interventions are limited. This study assessed the feasibility, acceptability, and preliminary effects of a scalable peer support intervention to reduce psychological distress and enhance quality of life (QOL) post-HSCT. We conducted a pilot randomized clinical trial evaluating the Supporting Transplant Experiences with Peer Program (STEPP), a 5-session, telephone-based intervention delivered by an HSCT survivor offering psychoeducation and coping strategies. Participants were recruited 1 to 2 weeks pre-HSCT admission and randomized to STEPP or usual care. Feasibility was defined as ≥60% enrollment and ≥60% completing ≥3 sessions. Acceptability was assessed using the Client Satisfaction Questionnaire (CSQ; ≥3/4 mean benchmark). PRO measures of anxiety, depression (Hospital Anxiety and Depression Scale), posttraumatic stress (Post-Traumatic Stress Disorder Checklist-Civilian Version), QOL (Functional Assessment of Cancer Therapy-Bone Marrow Transplant), social support (Social Support and Effectiveness Questionnaire), and self-efficacy (Cancer Self-Efficacy Scale) were assessed pre-HSCT and at 30 and 60 days post-HSCT. Mixed-effects models explored preliminary effects. We enrolled 77% (90 of 117) of eligible patients (STEPP, n = 45; usual care, n = 45); 12 became ineligible due to HSCT cancellation or postponement. Among 78 active participants (mean age, 58.9 years [standard deviation, 12.3]; 57.7% women), all STEPP participants completed ≥3 sessions, and the mean CSQ score was 3.5. STEPP showed small-to-moderate improvements in day 30 anxiety (d = -0.36) and day 60 anxiety (d = -0.58), depression (d = -0.25), posttraumatic stress (d = -0.28), QOL (d = 0.48), social support (d = 0.36), and self-efficacy (d = 0.46). STEPP exceeds feasibility and acceptability thresholds, showing promising psychosocial benefits warranting a fully powered multisite efficacy trial. This trial was registered at www.clinicaltrials.gov as NCT06010017.
9 June 2026
Read appraisal →JMIR research protocols
Creation of an mHealth Infrastructure to Support the Development and Delivery of mHealth Interventions: Protocol for Demonstration Projects Addressing Smoking Cessation in Cancer Care
BACKGROUND: Cancer remains a leading cause of morbidity worldwide. To reduce this burden, scalable, effective approaches are needed to address modifiable risk factors for cancer and support behavioral self-management. With smartphone ownership now nearly ubiquitous, mobile health (mHealth) interventions offer a powerful means to extend the reach, accessibility, and sustainability of evidence-based treatments for a variety of modifiable risk factors (eg, excessive alcohol use, physical inactivity, poor diet, and smoking). Moreover, the flexibility of mHealth platforms enables efficient delivery of novel interventions, supports innovative study designs, and facilitates real-time data collection to advance public health research. OBJECTIVE: Despite the great potential of mHealth interventions, developing high-quality mHealth tools is complex, time-consuming, and resource-intensive. To address these challenges, we are developing a coordinated, accessible, research-grade infrastructure for mHealth app development, testing, and dissemination. METHODS: The mHealth Florida infrastructure (mFLi) will provide a comprehensive, low-code software platform that enables researchers to build apps compatible with major mobile operating systems, namely, Apple iOS and Google Android. Through a modular interface, users will select from a menu of prebuilt features to tailor functionality to specific study needs. The platform will include 3 integrated environments (development, testing, and production), allowing researchers to prototype, evaluate, and deploy mHealth interventions. This infrastructure will be developed and maintained by a multidisciplinary team, ensuring that the platform is technically robust and usable and adheres to institutional and regulatory standards. To demonstrate the platform's functionality, utility, and adaptability, a multisite study comprising three initial projects focused on smoking cessation among patients with cancer is being conducted: (1) participant screening and enrollment, (2) randomization and treatment delivery, and (3) data processing using machine learning methods with on-device and cloud-based approaches. RESULTS: This study was funded in May 2023, and ethics approval was obtained from all involved sites' institutional review boards between February 2024 and October 2025. Recruitment began in March 2025 and enrollment is ongoing. As of January 2026, 41% (37/90) of the target sample have been enrolled and 21% (19/90) have completed their 6-month assessment. Data collection will be completed once the final participant completes their 6-month assessment (expected May 2027), with analyses commencing thereafter. Study findings are anticipated to be published in a peer-reviewed journal in 2027. CONCLUSIONS: Collectively, these projects will illustrate how mFLi can streamline app development, facilitate rapid translation of research into practice, and reduce barriers for researchers and developers. Ultimately, mFLi is designed to accelerate innovation in mHealth research, enhance access to behavioral interventions, and improve health outcomes among diverse populations. TRIAL REGISTRATION: ClinicalTrials.gov NCT06909357; https://clinicaltrials.gov/study/NCT06909357. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): PRR1-10.2196/92288.
1 June 2026
Read appraisal →JMIR research protocols
Telehealth Versus In-Person Caregiver-Mediated Behavioral Treatment for Challenging Behaviors in Children With Autism Spectrum Disorder: Protocol for COACH (Caregiver Outreach for Autism Coaching at Home) Randomized Controlled Trial
BACKGROUND: Approximately 40%-60% of children diagnosed with autism spectrum disorder (ASD) exhibit challenging behaviors such as aggression, elopement, self-injury, and property destruction that can endanger the health and safety of the child or others, often cause significant distress for families, and hinder the child's developmental progress. Although behavioral treatments grounded in the principles of applied behavior analysis have been shown to effectively reduce or eliminate these behaviors, significant gaps remain in how to deliver these interventions equitably. OBJECTIVE: This study aims to compare the efficacy of delivering in-person versus telehealth caregiver-mediated behavioral treatment for challenging behaviors in children with ASD. METHODS: The Caregiver Outreach for Autism Coaching at Home trial is a single-site, multiarm, parallel-group randomized controlled trial to evaluate the efficacy of delivering caregiver-mediated behavioral treatment for children with ASD who exhibit challenging behaviors. We aim to recruit 90 caregiver-child dyads with children with ASD aged 2 years, 0 months to 7 years, 11 months from metropolitan, suburban, and rural areas across New Mexico and the neighboring El Paso, Texas region. Following a practical functional assessment to identify the potential social-environmental variables maintaining the child's challenging behavior, caregiver-child dyads will be randomized into one of three parallel treatment arms: (1) caregiver-mediated behavioral treatment delivered via telehealth, (2) caregiver-mediated behavioral treatment delivered in person, or (3) caregiver online psychoeducation. Interventions will be delivered by clinicians in weekly 1-hour sessions in the home setting over 12 weeks. RESULTS: The study was funded in May 2022, and recruitment began in January 2024. As of March 2026, 81 caregiver-child dyads have been enrolled, with equal allocation across the 3 study groups (27 per group). Recruitment is expected to conclude in October 2026, with data analysis planned for Winter 2026 through Spring 2027. CONCLUSIONS: This study will address a critical gap in the scientific literature on scalable, family-centered approaches to safely and effectively reduce challenging behavior in children with ASD and decrease caregiver stress. Findings may inform best practices and payer policy for telehealth-delivered behavioral interventions targeting child safety and caregiver well-being.
29 May 2026
Read appraisal →Neuroscience
The effects of acute psychosocial stress on binge urges, inhibitory control, negative urgency and eating pathology in Binge Eating Disorder
Stress is frequently reported as a trigger for binge episodes, yet it remains unclear how subjective and physiological stress responses interact with cognitive and affective mechanisms to increase binge eating risk. The present study investigated the effects of acute psychosocial stress on binge urges, focusing on the role of state- and trait-level vulnerabilities. Twenty-eight individuals with BED were randomized to either a Stress condition (n = 14) or No-Stress condition (n = 14). Following an acute psychosocial stressor, participants completed the Stop-Signal Task to behaviourally assess inhibitory control. Self-reported mood, binge urge ratings and saliva samples for stress biomarker detection were collected at three timepoints throughout the experimental session. Results demonstrated that acute psychosocial stress elicited marked increases in total mood disturbance, without corresponding physiological stress responses and significantly impaired inhibitory control. Notably, only subjective stress responses were associated with greater binge urges, whereas physiological stress responses were not. This dissociation highlights subjective stress reactivity as a clinically meaningful mechanism of risk in BED, underscoring the potential value of interventions that target stress perception and regulation to reduce binge eating vulnerability.
25 May 2026
Read appraisal →PloS one
A systematic review and narrative synthesis of the psychometric properties and biopsychosocial correlates of the English version of the Intuitive Eating Scale-2
Intuitive eating is an adaptive eating approach characterised by having unconditional permission to eat when you are hungry, eating for physical rather than emotional reasons, relying on internal hunger and satiety cues, and honouring your health and practising gentle nutrition. The Intuitive Eating Scale-2 (IES-2) is currently the most commonly-used intuitive eating measure but, to date, there has not been a systematic review of how it performs psychometrically outside of the student sample in which it was developed. This systematic review aimed to assess the psychometric properties of the IES-2, including its associations with other variables, across all studies that used it as a measure of intuitive eating and reported psychometric data. MEDLINE, PsycINFO, Scopus and Web of Science were searched in April 2022, May 2024 and June 2025, identifying 90 papers from diverse populations including college students, people from the general population/community, and people seeking treatment for eating disorders or weight management. Results were presented and synthesised narratively, and risk of bias was assessed using two domains from the COSMIN Risk of Bias checklist. Findings suggest that the IES-2 has good construct validity but underperforms psychometrically in other areas such as response distribution, subscale inter-correlations and dimensionality, particularly in relation to the 'Unconditional permission to eat' subscale. Alternative factor structures were reported on and a three-factor solution excluding the 'Unconditional permission to eat' subscale was found to have promising results. The review contributes a comprehensive account of the biopsychosocial correlates of the IES-2, as well as identifying that studies relating to weight or eating disorders often report mean scores that tend towards the extremes of the scale. Limitations include the exclusion of non-English translations of the IES-2, and future reviews would benefit from being conducted in other languages. PROSPERO registration: CRD42022299436. Funder: ESRC (ES/P000711/1).
22 May 2026
Read appraisal →JMIR formative research
Impact of Restriction-Resumption Protocols on Mood and Anxiety in Healthy Adults: Randomized Controlled Trial
BACKGROUND: Simple behavioral and cognitive actions can reduce symptoms of anxiety and depression. However, there is limited research investigating whether restricting those same actions increases symptoms and whether resuming those actions reduces symptoms. OBJECTIVE: The primary aim was to examine the impact of restricting, then resuming (the "restriction-resumption protocol"), 5 groups of actions called the Things You Do (TYD), on symptoms of depression and anxiety. Participant satisfaction with life and perceptions of changes in behavior and mental health were also evaluated. METHODS: In total, 70 adults were randomly allocated to an intervention group (IG) or a control group (CG). IG participants completed a 3-phase protocol over 8 weeks: a 2-week baseline period (phase A), 2 weeks of restricted TYD actions (phase B), and a 4-week resumption phase (phase C). CG participants were instructed to maintain their usual activities and routines. Primary outcomes included symptoms of depression (Patient Health Questionnaire-9), anxiety (Generalized Anxiety Disorder-7), and frequency of TYD actions. Secondary outcomes included satisfaction with life and perceived changes in mental well-being. Outcomes were measured weekly throughout the 8-week trial and again after the trial at week 9. RESULTS: Scores on outcome measures did not differ between groups during phase A and after the trial. However, by the end of phase B, IG participants showed significantly increased symptoms of depression and anxiety compared to CG participants (Ps<.001). Large within-group effect sizes were observed on the Patient Health Questionnaire-9 and Generalized Anxiety Disorder-7 from baseline to the end of phase B (Cohen d>1.8) and from phase B to after the trial (Cohen d>1.9). Changes in satisfaction with life and evaluations of mental health corresponded with symptom changes. CONCLUSIONS: These findings demonstrate a relationship between the frequency of performing specific actions and symptoms of depression and anxiety. The restriction-resumption protocol has the potential to increase our understanding of causal mechanisms underlying common mental disorders and may lead to the development of new models of prevention and treatment. TRIAL REGISTRATION: Australian and New Zealand Clinical Trials Registry ACTRN12624001491550; https://www.anzctr.org.au/Trial/Registration/TrialReview.aspx?id=388828&isReview=true.
21 May 2026
Read appraisal →PloS one
Optimising supervised machine learning algorithms predicting cigarette cravings and lapses for a smoking cessation just-in-time adaptive intervention (JITAI).
This study aimed to optimise the balance between participant burden and algorithm performance for predicting high-risk moments in a smoking cessation just-in-time adaptive intervention (JITAI) by systematically varying ecological momentary assessment (EMA) prompt frequency, predictor count, and training data source. Thirty-seven participants completed 16 EMAs per day for the first 10 days of their smoking cessation attempt, reporting mood, context, behaviour, cravings, and smoking lapses. Random forest algorithms predicting lapses and cravings were evaluated in terms of F1-score and ROC-AUC via mixed effects models accounting for clustering within individuals. Performance across out-of-sample individuals ranged from excellent to poor but was, on average, modest. Lapse prediction outperformed craving prediction, particularly for ROC-AUC (Median F1-score: Lapses 0.436 [IQR 0.180-0.625], Cravings 0.400 [IQR 0.048-0.649]; Median ROC-AUC: Lapses 0.659 [IQR 0.514-0.809], Cravings 0.628 [IQR: 0.510-0.729]). A substantial proportion of configurations fell below commonly used minimum performance thresholds, particularly for F1-score. Reducing EMA frequency had outcome- and metric-dependent effects. Lapse F1-scores improved with fewer prompts (16 EMAs: 0.254 [IQR 0.081-0.500], 3 EMAs: 0.588 [IQR 0.353-0.667]), while ROC-AUC showed a slight, inconsistent decline (16 EMAs: 0.661 [IQR 0.520-876], 4 EMAs: 0.613 [IQR 0.494-0.786], 3 EMAs: 0.704 [IQR 0.567-0.809]). For cravings, both metrics declined with fewer prompts (F1-score: 16 EMAs: 0.470 [IQR 0.141-0.745]; 3 EMAs: 0.333 [IQR 0.000-0.600]; ROC-AUC: 16 EMAs 0.700 [IQR 0.582-0.811], 3 EMAs 0.544 [IQR 0.421-0.676]). Feature reduction had negligible impact on lapse prediction (F1-score: all features 0.435, selected features 0.441; ROC-AUC: all 0.660, selected 0.657), but slightly reduced craving performance (F1-score: all 0.410 [IQR 0.117-0.646], selected 0.400 [IQR 0.000-0.650]; ROC-AUC: all 0.632, selected 0.622). Including participant-specific data improved lapse F1-scores (None 0.286 [IQR 0.000-0.571], 30 pc 0.542 [IQR: 0.329-0.667]), but not ROC-AUC (None 0.655 [IQR: 0.512-0.786], 30 pc 0.694 [IQR 0.513-0.852]); and impaired craving ROC-AUC (None 0.650 [IQR: 0.544-0.734], 30 pc 0.614 [IQR 0.493-0.730]; F1-score: None 0.424 [IQR 0.143-0.649], 30 pc 0.400 [IQR 0.000-0.703]). Overall, EMA-based machine learning detected lapse risk but showed modest overall performance and substantial inter-individual variability. Algorithms using higher EMA density, larger predictor sets, and participant-specific training data did not consistently outperform over more parsimonious approaches. However, machine learning prediction alone is unlikely to be sufficient for real-world JITAI implementation and may be best combined with complementary rules-based approaches.
16 May 2026
Read appraisal →Journal of medical Internet research
Stimulating Future-Oriented Thinking and Goal-Achievement Through the Future Self Using Virtual Reality and a Smartphone App: Randomized Controlled Trial
BACKGROUND: Research suggests that identification with one's future self-encompassing vividness, connectedness, and valence-plays a key role in motivating future-oriented choices and goal pursuit. Interventions aiming to strengthen future self-identification have been shown to reduce maladaptive behaviors and promote well-being, but traditional approaches often rely heavily on imagination. Emerging technologies offer novel opportunities to make the future self more vivid and tangible, potentially reducing cognitive burden, and enhancing intervention effectiveness. OBJECTIVE: This randomized controlled trial evaluated the effectiveness of a digital intervention on the following primary outcomes: future self-identification, future orientation, consideration of future consequences, self-defeating behavior, goal commitment, and goal achievement. METHODS: In this parallel, 3-arm randomized controlled trial, 321 first-year students from a large public university in the Netherlands were randomized (1:1:1) with blocks of 9 to a smartphone app-based future-self intervention, an immersive virtual reality (VR) version of the same intervention, or an active goal-setting control condition. Participants assigned to the smartphone or VR condition engaged with the intervention over a 3-week period, during which they interacted with their 10-year-old self, their future self. Participants in the control condition received no further support. Each of the 3 conditions consisted of 107 participants, who were all included in the analyses following the intention-to-treat principle. Due to the nature of the intervention, blinding was not possible. RESULTS: Compared to the goal-setting control condition, both intervention conditions yielded significant short-term improvements on all 3 aspects of future self-identification (smartphone: dvividness=0.49; dvalence=0.44; dconnectedness=0.43; VR: dvividness=0.35; dvalence=0.44; dconnectedness=0.43), and a small decline in vividness (smartphone: d=-0.36; VR: d=-0.23) and connectedness (smartphone: d=-0.36; VR: d=-0.32) at 6-month follow-up. Additionally, the intervention buffered declines in future orientation during the study period (smartphone: d=.16; VR: d=0.18), and VR delivery led to significantly higher weekly goal achievement (d=0.88). However, effects on future orientation were not sustained over time (smartphone: d=-0.18; VR: d=-0.25), resulting in similar levels across conditions at 6-month follow-up. Furthermore, no significant effects emerged for other primary or secondary outcomes, such as self-defeating behavior (smartphone: doverall=-0.10; VR: doverall=-0.03), impulsivity (smartphone: doverall=-0.07; VR: doverall=-0.09), or academic performance (η2partial=0.00). No adverse events were reported in any of the conditions. CONCLUSIONS: This randomized controlled trial innovatively compares smartphone and immersive VR future-self interventions. Findings suggest that digital interventions leveraging visual and interactive representations of the future self can strengthen future self-identification and future orientation, and support (short-term) goal pursuit. Together, they highlight the potential of scalable digital future self approaches for educational and preventive health contexts.
15 May 2026
Read appraisal →PloS one
Narrative-affect discrepancy as a regulated degree of freedom in 351,734 relationship narratives
In naturalistic emotional narratives, the intensity of expressed affect does not scale proportionally with narrative structure. Using 351,734 English-language relationship narratives from online support communities (the ANEST Narrative-Affect Dataset, ANAD v1.1.0), we constructed a two-dimensional expressive space defined by narrative complexity (N) and linguistically inferred affective intensity (A), with their signed discrepancy ([Formula: see text]) treated as a derived coordinate. Rather than converging toward low discrepancy, human narratives occupied a broad but structured space consistent with trade-offs between relational exposure and cognitive effort. We identified four empirically separable regimes of expressive organization: coupled expression (non-extreme discrepancy; the complement of the extreme regimes), strategic understatement (high A, low N, D < 0), strategic overstatement (high N, low A, D > 0), and collapse (high A with limited narrative scaffolding). A data-anchored cost model (NCS) formalized these regimes as arising from the interaction of exposure risk and cognitive effort, not from discrepancy minimization per se. Coupled expression dominated the corpus (91.3%), while the remaining regimes formed smaller but non-negligible subpopulations (Understatement: n = 20,223; Collapse: n = 8,040; Overstatement: n = 2,223), indicating that extreme discrepancy configurations occur systematically rather than as isolated outliers. As a comparative probe, we projected an RLHF-aligned large language model into the same space using matched prompts and identical feature extraction. Because D is a deterministic function of N and A, expressive extent was quantified as convex hull area in the clipped [Formula: see text] plane. Under this procedure, the model occupied a markedly smaller region (approximately [Formula: see text] smaller hull area; bootstrap 95% CI [1.68, 1.70]; permutation p < 0.0001) and concentrated near low-discrepancy configurations, with sparse occupancy of extreme under- and overstatement regimes. Together, these findings suggest that narrative-affect discrepancy is a measurable and regulated dimension of emotional expression and provide a reproducible geometric basis for comparing expressive degrees of freedom across populations and systems.
14 May 2026
Read appraisal →Sleep
Testing affect regulation theories of dreaming
STUDY OBJECTIVES: Dreaming is theorized to regulate waking affect via fear extinction learning, similar to exposure therapy. We tested whether fear in dreams was associated with lower negative affect (NA) and higher positive affect (PA) on subsequent mornings and whether these effects were moderated by emotion regulation strategies or mixed dream emotion states. Additionally, we examined whether individuals with higher average dream fear showed more adaptive emotion regulation. METHODS: We employed a natural language processing tool and Bayesian multilevel modeling using online daily survey data in a community sample (n = 536; 85.6% women; Mage = 39.3 ± 17.3; days = 4715). RESULTS: Frightening dreams were associated with more negative affect the following morning (+7% NA, pd = 100%), which was amplified for individuals with more adaptive emotion regulation (+3% NA, pd = 98.69). Dreams concurrently high in fear and joy were related to 20% higher odds of zero NA the next morning (pd = 97.82%). Higher average dream joy was related to 9% higher morning PA (pd = 98.79%). Individuals with higher average dream fear showed greater adaptive emotion regulation (β = 0.18, pd = 99.93%). CONCLUSIONS: Findings supported affect continuity from dream to wake, yet affect regulation at the individual level. Frightening dreams were related to worse morning affect, with a more pronounced effect among high-regulation individuals. High-regulation individuals tended to experience more fearful dreams. Mixed dream emotions were related to mornings without NA, and dream joy was related to higher morning PA. Our findings suggest adaptive, though dynamic, effects of dream emotions on psychological functioning.
14 May 2026
Read appraisal →American journal of speech-language pathology
Effect of the Fake Phone Call App on Patient Adherence Behavior and Motivation in Conversation Training Therapy: A Randomized Crossover Trial
PURPOSE: The Fake Phone Call app was previously developed and feasibility-tested for conversational target voice practice in resonant voice therapy. The purpose of the present study was to examine its effect on adherence and motivation for conversational practice in conversation training therapy (CTT) as well as its usability and utility in CTT. METHOD: Eleven patients referred for voice therapy served as their own control in a randomized crossover trial comparing 1 week of app-supported homework practice with 1 week of standard-of-care homework completion in CTT. Dependent variables included adherence and motivation variables. In addition, usability and utility as well as outcomes were examined at study completion. RESULTS: App support significantly increased practice frequency, practice fidelity, and goal commitment for practice. Usability and utility were good. Exploratory investigation revealed that patients randomized to receive the app condition first and the standard-of-care condition second had greater gains on all study variables and significantly greater self-reported generalization at treatment completion. CONCLUSION: Fake Phone Call support significantly improves adherence and motivation for CTT, with additional benefits when provided earlier in therapy.
11 May 2026
Read appraisal →Behavioral and brain functions : BBF
Neuropeptide S modulation of learning and memory: a systematic review
BACKGROUND: In the last two decades, Neuropeptide S (NPS) has been identified as a key bioactive peptide in the mammalian brain, influencing fear, anxiety, wakefulness, reward, and learning. While some reviews have addressed its role in reward-seeking and anxiety, few have addressed its particular role in learning and memory. The neuropeptide S receptor 1 is highly expressed in key areas for learning processing, such as the hippocampus, cortex, thalamus, and amygdala. This review aims to examine evidence from human and animal studies that focused on the NPS system's role in modulating learning and memory. A special focus is given to experiments addressing the impact of NPS on associative learning leading to addiction and in fear conditioning, pointing to its potential therapeutic value in associated pathologies. MAIN BODY: An advanced search was conducted using the databases PubMed, Google Scholar, Web of Science, and Scopus, focusing on memory and Neuropeptide S. The reviewed data suggest that NPS modulation occurs at all memory phases, including acquisition, consolidation, and retrieval, and in extinction learning, whether motivated by appetitive or aversive stimuli. The summarized evidence shows that the NPS system interferes with working and short-term memory, mitigates learning impairments, enhances spatial and object memory consolidation, supports fear extinction learning and inhibitory avoidance consolidation, and reinstates drug-seeking behaviors. The NPS system closely interacts with key neuromodulators, including orexinergic, dopaminergic, and noradrenergic systems, in influencing memory. CONCLUSION: The Neuropeptide S system emerges as a critical modulator of memory processes. The NPS signaling may preferentially influence learning that involves emotionally or motivationally relevant stimuli. This highlights the NPS system's potential as a target for therapeutic interventions for particular memory impairments.
11 May 2026
Read appraisal →European journal of pediatrics
The effect of motivational interviewing on internet addiction and digital game addiction in adolescents: a randomized controlled trial
UNLABELLED: The purpose of this study was to measure the effect of motivational interviewing on both reducing internet addiction and digital game addiction in adolescents. A parallel-group randomised controlled trial was adopted. The study population consisted of ninth-grade (14-15 years of age) high school students in a city in Turkiye. The study was completed by 88 participants (experimental: 44; control: 44). The data were collected using a Personal Information Form, the Young Internet Addiction Test, and the Digital Game Addiction Scale. The experimental group received a preparatory session and five weekly motivational interviewing sessions. Instruments were administered to both groups before (pre-test) and after the intervention (post-test), and at follow-up tests 3 and 6 months after the final session. The data were analysed using the two-way mixed design and the Bonferroni Comparison Test. The mean scores of internet addiction and digital game addiction decreased significantly after the motivational interviewing in the experimental group compared to the control group (p < 0.001) in both the post-test and follow-up tests. CONCLUSION: The present study concluded that motivational interviewing may be associated with reductions in mitigating symptoms of internet addiction and digital game addiction behaviours among adolescents. Motivational interviewing could be implemented to reduce internet addiction and digital game addiction behaviours. Trial registration: The study was registered on a clinical trial database (NCT06721702). The study started on December 11, 2023 (actual date on which the first participant was enrolled). WHAT IS KNOWN: • Internet addiction and digital game addiction are two increasingly important problems among adolescents. • Digital games and online activities negatively affect adolescents' physical, social, and psychological health. WHAT IS NEW: • Motivational interviewing was an effective technique to reduce online gaming and internet addiction. • A motivational interviewing program comprising at least six sessions could be implemented to promote behavioural change in adolescents.
10 May 2026
Read appraisal →JMIR mHealth and uHealth
Alcohol Use Trajectories During the First 72 Weeks of WHOOP Wearable Platform Membership: Observational Cohort Study
Among 30,000 new wearable device users (11.6 million person-days), self-reported alcohol consumption significantly declined over 72 weeks, with the daily probability of drinking decreasing by 5.8 percentage points (P<.001) and reductions across age and biological sex.
2 May 2026
Read appraisal →Behavioural brain research
Influences of high-definition transcranial direct current stimulation (HD-tDCS) and cognitive reappraisal on emotions and aggression following social exclusion
Social exclusion can trigger negative emotions and aggressive behavior, while emotion regulation strategies and neuromodulation techniques may mitigate these effects. This study employed a 2 × 2 factorial design to investigate the impacts of high-definition transcranial direct current stimulation (HD-tDCS) and cognitive reappraisal on emotions and aggression following social exclusion. Thirty-two university students were randomly assigned to four groups: active HD-tDCS + cognitive reappraisal, sham HD-tDCS + cognitive reappraisal, active HD-tDCS + no reappraisal, and sham HD-tDCS + no reappraisal. Results showed that active HD-tDCS significantly reduced aggressive behavior in both the Competitive Reaction Time Task and Job-relevant Evaluations Task, likely by enhancing prefrontal regulatory control. Cognitive reappraisal significantly decreased negative emotions but did not directly reduce aggressive behavior, suggesting a dissociation between emotional and behavioral regulation pathways. No significant interaction was found between HD-tDCS and cognitive reappraisal, possibly due to their independent neural mechanisms. These findings highlight the potential of HD-tDCS and cognitive reappraisal as interventions for social exclusion-induced aggression and emotional dysregulation.
1 May 2026
Read appraisal →JMIR formative research
Multimodal Sentiment and Emotion Analysis Framework for Personalized Health Coaching Messages: Proof-of-Concept Study
BACKGROUND: Text generation approaches in health care communication have evolved along 2 major paths. The first path involves generative adversarial networks, progressing from basic architectures to specialized variants like Text-to-Text Generative Adversarial Network (TT-GAN) and Time and Frequency Domain-Based Generative Adversarial Network (TF-GAN), which address challenges in discrete text generation through techniques such as Gumbel-Softmax and reinforcement learning. The second path emerges from transformer-based architectures, particularly Generative Pretrained Transformer-2 (GPT-2), which uses extensive pretraining and self-attention mechanisms to generate contextually appropriate text. GPT-2's transformer architecture enhances persuasive health communication by generating personalized messages using various strategies like task support, dialogue support, and social support for effective health interventions. OBJECTIVE: This study aimed to use GPT-2 as a generative method to construct persuasive text in a dataset and compare the performance of sentiment analysis and emotion detection analysis. METHODS: We combined sentiment analysis tools (VADER [Valence Aware Dictionary and Sentiment Reasoner] and TextBlob) with emotion detection methods (Text2Emotion and NRCLex [National Research Council Lexicon]) to analyze health coaching messages across different persuasive types: reminder, reward, suggestion, and praise. RESULTS: TextBlob and VADER achieved accuracies of 57% and 69%, respectively, while RoBERTa (robustly optimized BERT approach)-sentiment outperformed them with an accuracy of 88%. Emotion detection showed a high prevalence of "joy" and "happy" labels (93.69% positive skew). While transformers excel in accuracy, lexicon-based models like VADER offer a better performance-efficiency balance for real-time health communication systems. For emotion detection, all categories showed perfect accuracy (1.0), while trust showed mixed results, with precision, recall, and F1-score values ranging from 0.81 to 0.96. The emotion detection analysis revealed varying success rates across different emotions, with some categories, such as anger and neutral, showing reasonable performance and others, such as trust, showing mixed performance. CONCLUSIONS: This research contributes to understanding the emotional dynamics of persuasive health communication and highlights both the capabilities and limitations of current natural language processing tools in analyzing health-related persuasive messaging. This proof-of-concept study using synthetically generated data establishes a methodological framework for multimodal sentiment and emotion analysis. The findings require validation with real-world health coaching messages before clinical deployment.
24 Apr 2026
Read appraisal →JMIR formative research
Designing Psychologically Grounded Artificial Intelligence for Supporting Bystander-Based Cyberaggression Intervention: Mixed Methods Exploratory Study
BACKGROUND: Cyberaggression poses a growing threat to mental health, contributing to increased distress, reduced self-esteem, and other adverse psychosocial outcomes. Although bystander intervention can mitigate the escalation and impact of cyberaggression, individuals often lack the confidence, strategies, or language to respond effectively in these high-stakes online interactions. Advances in generative artificial intelligence (AI) present a novel opportunity to facilitate digital behavior change by assisting bystanders with contextually appropriate, theory-informed intervention messages that promote safer online environments and support mental well-being. OBJECTIVE: This mixed methods design study aimed to explore the feasibility of using generative AI to support bystander intervention in cyberaggression on social media. Specifically, we examined whether AI can generate effective responses aligned with established intervention strategies and how these responses are perceived in terms of their potential to de-escalate online harm and foster behavior change. METHODS: We collected 1000 real-world cyberaggression examples from public social media datasets and generated bystander intervention responses using 3 distinct prompt strategies: a generic policy reminder, a baseline GPT prompt, and a theory-driven GPT prompt (AllyGPT). To evaluate the responses, we conducted computational linguistic analyses to assess their psycholinguistic features and carried out a mixed methods evaluation. Three trained coders rated each message on favorability, conversational impact, and potential to change behavior and later participated in semistructured interviews to reflect on their evaluation process and perceptions of intervention effectiveness. RESULTS: Linguistic analyses revealed that baseline GPT responses exhibited more emotionally positive and authentic language compared to AllyGPT responses, which showed a more analytical and assertive tone. Policy reminder messages were linguistically rigid and lacked emotional nuance. Human evaluation results showed that AllyGPT responses received the highest effectiveness ratings for low-incivil cyberaggression cases in 2 dimensions (favorability and changing behavior), and baseline GPT works better for mid and high levels for all effectiveness dimensions. For medium- and high-incivility aggressions, baseline GPT responses received the highest ratings across all 3 dimensions of effectiveness (favorability, discussion-shifting potential, and likelihood of changing bullying behavior), followed by AllyGPT, with policy reminders rated lowest. Qualitative feedback further emphasized that baseline GPT responses were perceived as natural and inclusive, while AllyGPT responses, although grounded in psychological theory, were sometimes viewed as overly direct. Policy reminders were considered clear but lacked persuasive impact. CONCLUSIONS: Our work showed that designing effective AI-generated bystander interventions requires a deep sensitivity to platform culture, social context, and user expectations. By combining psychological theory with adaptive, conversational design and ongoing feedback loops, future systems can better support bystanders, delivering interventions that are not only contextually appropriate but also socially resonant and behaviorally impactful. As such, this work serves as a foundation for scalable, human-centered AI systems that promote safer online spaces and users' mental well-being.
16 Apr 2026
Read appraisal →Journal of affective disorders
Who engages? Machine learning insights into digital mindfulness-based intervention for generalized anxiety disorder.
BACKGROUND: Although mindfulness ecological momentary interventions (MEMI) appear effective in alleviating worry symptoms, treatment engagement remains suboptimal. Determining baseline variables of MEMI over self-monitoring placebo (SM) can inform tailored interventions for individuals with generalized anxiety disorder (GAD). METHOD: Machine meta-learning methods (ML) were applied to predict two-week engagement (log-transformed number of prompts completed) among individuals randomized to MEMI or SM (N = 110). Sixteen baseline variables comprised the predictor set: clinical, demographic, process, and executive functioning (EF) factors. Random forest using a five-fold nested cross-validation approach mitigated overfitting. X-learner meta-algorithms estimated conditional average treatment engagement (CATE). Shapley additive explanations evaluated relative importance. RESULTS: The 16-predictor model displayed strong predictive performance (R-squared [R2] = 82.7%; root mean squared error [RMSE] = 0.780; mean absolute error [MAE] = 0.512). The top-10 predictor model also yielded good predictive performance (R2 = 82.1%; RMSE = 0.547; MAE = 0.307). As predicted by the CATE analysiscate, participants had the highest treatment engagement when assigned to MEMI instead of SM (d = 1.447, p < .001). The following baseline variables predicted more engagement with MEMI over SM: lower GAD severity, inhibition response time (RT), and EF errors, higher attentional control, empathy, and verbal fluency (capitalization theory); lower mindfulness, and treatment expectancy, poorer working memory, and higher set-shifting RT (compensation model). LIMITATIONS: The small sample size, single engagement metric, and brief duration might constrain generalizability. DISCUSSION: Integrating robust ML approaches could optimally identify prescriptive predictors of engagement to brief digital mental health interventions to inform targeted treatments. TRIAL REGISTRATION: ClinicalTrials.gov ID (NCT04846777).
16 Apr 2026
Read appraisal →JMIR mHealth and uHealth
Promoting Self-Regulated Social Media Use on Smartphones With a Mobile Intervention App (Wellspent): Randomized Controlled Trial
BACKGROUND: Problematic social media use has been linked to reduced well-being and impulse control difficulties. While digital self-control apps show potential for reducing general app usage, they often lack customization, leading to limited effectiveness and increased user resistance. Their impact on problematic social media use remains uncertain. OBJECTIVE: This study evaluates the effectiveness of the Wellspent app, a customizable mobile intervention app designed to promote self-regulated social media use by targeting user-defined problematic app use and offering tailored behavioral nudges. METHODS: In a 3-week randomized controlled trial, 70 iPhone users (mean age 26.2, SD 5.6 years; 47/70, 67% female), regularly using at least 1 social media app, were randomly assigned to an intervention (n=35) or control group (n=35). The intervention group received personalized full-screen reminders with the option to quit or continue social media app use whenever an app session exceeded a self-defined time limit. Participants completed weekly online surveys measuring problematic social media use, problematic smartphone use, self-efficacy, and daily screen time on their most problematic app. Linear mixed models tested intervention effects. RESULTS: While no significant reduction in problematic social media use or increase in self-efficacy was observed, the intervention group showed a significant reduction in daily screen time on their most problematic app by approximately 29 minutes (estimate=-29.35, SE 6.84, 95% CI -42.79 to -15.99; P<.001), and a significant decrease in perceived problematic smartphone use (estimate=-0.46, SE 0.18, 95% CI -0.80 to -0.11; P=.01). CONCLUSIONS: The Wellspent app demonstrated short-term efficacy in reducing problematic smartphone use. By allowing users to tailor interventions to their personal goals, the app shows promise as a self-directed tool to support healthier digital habits. Further research should explore long-term effects and feature-specific impacts.
9 Apr 2026
Read appraisal →Journal of medical Internet research
A Gamified Mobile Health Intervention to Promote Physical Activity, Executive Function, and Mental Health in College Students: Randomized Controlled Trial
BACKGROUND: College students commonly experience suboptimal health conditions, including insufficient physical activity (PA), excessive body weight, and declining physical fitness. Traditional interventions face low adherence, while gamified mobile health (mHealth) programs may improve engagement and outcomes. OBJECTIVE: This study aimed to evaluate the feasibility and effectiveness of a novel gamified, incentive-based mHealth intervention on primary outcomes (PA and adherence) and secondary outcomes (physical fitness, body composition, executive function [EF], and mental health). METHODS: A 2-arm parallel-group randomized controlled trial (RCT) was conducted in 2025 at Yantai University with 160 college students (18-25 years; BMI 18.5-30.0) who were randomized 1:1 (computer-generated, sex-stratified blocks of 4; concealed allocation) to the intervention group (IG) or control group (CG; n=80 each); major exclusions were contraindications to exercise, severe physical/mental illness, recent PA interventions, or psychotropic medication use. Both used the same fitness watch-app system and identical PA targets (≥150 min moderate-to-vigorous physical activity [MVPA] per week or ≥900 metabolic equivalent-minutes [MET-min] per week); IG additionally received team-based gamification (competition, points/leaderboards, feedback, and rewards), while CG received monitoring only. PA and adherence were monitored throughout the 8-week intervention; other outcomes were assessed at baseline and 8 weeks (fitness, body composition, EF, and mental health). Open-label with blinded outcome assessors/analysts; intention-to-treat (ITT) with multiple imputation. RESULTS: At 8 weeks, data were available for 154 participants (IG 78; CG 76); all 160 were analyzed per ITT. Compared to the CG, the IG demonstrated significantly higher mean levels in all primary PA outcomes over 8 weeks (daily steps: mean 10,356, SD 1245 versus 8242, SD 1087; Δ=2114; d=1.81, 95% CI 1.44-2.18; P<.001; daily MVPA: mean 71, SD 15 versus 43, SD 12 min; Δ=28 min; d=2.06, 95% CI 1.68-2.45; P<.001; and weekly MET-min: mean 1650, SD 310 versus 1340, SD 285; Δ=310; d=1.04, 95% CI 0.71-1.37; P<.001). Adherence was also higher in the IG (both P<.001). Key secondary outcomes, including skeletal muscle mass (Δ +0.54 kg), body fat percentage (Δ -0.46 percentage points), 800/1000 m run time (Δ -6.5 s), and depressive symptoms (Δ -2.0 points), significantly favored the IG (all P<.001). No serious adverse events occurred. CONCLUSIONS: This RCT demonstrated that integrating gamification into an mHealth platform significantly enhanced PA, intervention adherence, and selected health outcomes among college students. The key innovation lies in the trial design, which held the technology platform and activity goals constant across groups, thereby allowing a clearer estimate of the incremental contribution of gamification. Unlike many existing studies, this approach reduces confounding from co-occurring intervention components in between-group comparisons. It therefore provides more comparable and methodologically robust evidence on the specific efficacy of gamification in mHealth interventions. For real-world implementation, this model leverages the existing smartphone and wearable ecosystem, offering a low-cost, scalable strategy for university health-promotion programs.
8 Apr 2026
Read appraisal →Journal of behavioral addictions
When rejection backfires: Social exclusion and economic decision-making in gamblers
BACKGROUND AND AIMS: Economic decision-making deviates from rational choice models and is influenced by social factors, particularly fairness considerations and belonging needs. Social exclusion and altruistic punishment are intended to promote behavioral conformity, yet paradoxically may exacerbate rather than correct targeted behaviors among individuals with addictions already experiencing chronic rejection. Several preliminary studies show that social exclusion increases money desirability in general populations, however, the interaction between acute social exclusion and economic decision-making among individuals with gambling disorders (PGs)-marked by aberrant financial decision-making-remains unexplored. METHODS: In a field study, 203 gamblers were recruited outside gambling venues and randomized to experience either social inclusion or exclusion via Cyberball, a virtual ball-tossing paradigm. Participants then completed an ultimatum game, making accept/reject decisions on fair and unfair monetary offers, followed by the Problem Gambling Severity Index. RESULTS: Social exclusion increased unfair offer acceptance among PGs but had no effect on low- and moderate-risk gamblers. PGs accepted more unfair offers following exclusion than inclusion. Gambling severity had no effect on unfair offers acceptance. Neither social exclusion nor gambling severity affected fair offer acceptance. DISCUSSION AND CONCLUSIONS: Social exclusion triggered utilitarian decision-making only among PGs. This selective effect suggests rejection compromises PGs' economic decision-making processes, shifting priorities toward immediate monetary gains over equitable outcomes. These findings reveal how social punishment creates self-reinforcing cycles where exclusion triggers the utilitarian mindset characterizing problematic gambling. Treatment and societal approaches relying on social consequences may inadvertently strengthen harmful patterns, highlighting the need for inclusion-based interventions.
5 Apr 2026
Read appraisal →