Neurofeedback-Driven Mobile App for Anxiety Regulation in Real-World Settings

Thesis Abstract

Neurofeedback-based interventions have shown promise for anxiety modulation in clinical and laboratory settings, yet there remains a gap in translating these gains to real-world environments where contextual stressors and digital engagement influence outcomes. This study proposes a mobile neurofeedback app designed to regulate anxiety by guiding users to modulate neural correlates of arousal (somatic and autonomic markers) through real-time feedback, coupled with behaviorally adaptive interventions. The aim is to evaluate whether a smartphone-delivered neurofeedback paradigm can reduce trait and state anxiety and improve-day-to-day functioning in naturalistic contexts. The objectives are (1) to assess the efficacy of the app in lowering self-reported anxiety levels across a 8-week intervention period; (2) to examine changes in physiological indicators of arousal (heart rate variability, galvanic skin response) and EEG-based markers (relative alpha/theta power) during real-world tasks; (3) to investigate user engagement, acceptability, and adherence as mediators of treatment effects; (4) to explore the moderating role of baseline cognitive-behavioral factors (trait anxiety, rumination, executive function) on outcomes; and (5) to synthesize qualitative experiences to refine the intervention framework. The study adopts a mixed-methods, randomized controlled trial design guided by the Tripartite Model of anxiety and the neuropsychophysiological framework of affect regulation. The population comprises adults aged 18–45 residing in urban and suburban settings, recruited across two metropolitan regions. A total sample of 180 participants will be randomly assigned to three arms (a) active neurofeedback via the mobile app, (b) sham neurofeedback controlling for expectancy, and (c) wait-list control. Data collection instruments include validated scales (State-Trait Anxiety Inventory, Generalized Anxiety Disorder-7, and Rumination Response Scale), ecological momentary assessment (EMA) for context-specific anxiety episodes, wearable sensors for heart rate variability and galvanic skin response, and mobile EEG headsets capturing frontal alpha and theta activity. Instruments will be piloted for reliability (Cronbach’s alpha ? .80 for scales) and concurrent validity against clinical assessments. Data analysis will integrate longitudinal multilevel modeling to examine trajectories of anxiety over time and mixed-effects models to assess within-subject changes across daily contexts. In addition, regression analyses will test mediation effects of physiological and neural markers on anxiety outcomes, while moderation analyses will explore baseline cognitive-behavioral factors. A Bayesian approach will be employed for model comparison and to quantify evidence for app efficacy. Thematic analysis of semi-structured post-intervention interviews will extract user experiences, perceived usefulness, and barriers to adherence, with triangulation against quantitative findings. The study anticipates that participants in the active neurofeedback arm will exhibit statistically significant reductions in state and trait anxiety (p < .05) compared with controls, accompanied by increases in heart rate variability and favorable shifts in EEG indices (increased relative alpha power, reduced theta activity during regulation attempts). It is expected that higher engagement and adherence will predict larger effect sizes, and that baseline rumination and executive function will moderate outcomes, with greater benefits observed among individuals with lower baseline rumination and robust executive control. The anticipated contribution to knowledge includes empirical evidence on the feasibility, efficacy, and mechanisms of a mobile neurofeedback intervention for anxiety in everyday life, integration of multimodal physiological and neural measures in a scalable consumer technology, and a refined theoretical model linking real-world neurofeedback to affect regulation. The study will inform clinical and digital health practice by outlining procedural guidelines for deployment, ethical considerations surrounding ambulatory neurofeedback, and recommendations for personalization to maximize engagement. The main conclusion is that a well-designed neurofeedback-enabled mobile app can produce meaningful anxiety reductions in real-world settings, mediated by physiological regulation and neural changes, with adherence and individual cognitive-behavioral profiles shaping the magnitude of benefit. Recommendations include iterative app enhancements to optimize real-time feedback fidelity, user-centered design to sustain engagement, and broader trials in diverse populations to establish generalizability.