Neurofeedback-Enhanced Cognitive Training for Anxiety via Mobile Apps

Thesis Abstract

This study investigates the efficacy of integrating neurofeedback with cognitive training delivered through mobile applications to alleviate anxiety symptoms and enhance cognitive-emotional regulation among adults in community settings. The problem addressed is the overreliance on pharmacological treatments and traditional psychotherapy alone, coupled with limited accessibility to evidence-based anxiety interventions for time-constrained individuals. The aim is to evaluate whether a scalable, ICT-driven neurofeedback-cognitive training platform can produce superior reductions in state and trait anxiety and improvements in attentional control, emotion regulation, and cognitive flexibility compared with standard cognitive training delivered via mobile apps. Specific objectives include (1) to examine changes in anxiety levels using validated scales (State-Trait Anxiety Inventory and Generalized Anxiety Disorder-7) over an eight-week intervention; (2) to assess improvements in executive function and attentional control using computerized tasks (Stroop, Flanker, and n-back) and self-report measures; (3) to analyze neurophysiological correlates of anxiety reduction through real-time electroencephalography (EEG) neurofeedback metrics such as frontal alpha asymmetry and theta/beta ratio; (4) to explore engagement, adherence, and user experience via mixed-methods data; and (5) to test a theoretical model combining neurofeedback mechanisms (operant conditioning and neurocognitive optimization) with cognitive-behavioral frameworks (theory of planned behavior and cognitive appraisal theory) to predict anxiety outcomes. A quasi-experimental design with a mixed-methods approach will be employed. The population comprises adults aged 18–45 with elevated anxiety symptoms recruited from community centers and online platforms (N ? 180). Participants will be randomized into two parallel groups an experimental group receiving neurofeedback-enhanced cognitive training via a mobile app (n ? 90) and an active control group receiving standard cognitive training via a comparable mobile app without neurofeedback (n ? 90). Data collection will occur at baseline, mid-intervention (Week 4), post-intervention (Week 8), and a 3-month follow-up. Instruments include standardized self-report scales (STAI, GAD-7, DERS for emotion regulation), cognitive task batteries (Stroop, Flanker, n-back), EEG neurofeedback metrics captured during training sessions, and a user experience questionnaire. Validity and reliability will be ensured through instrument calibration, pilot testing (n = 20), and inter-rater reliability checks for qualitative components. Quantitative data will be analyzed using mixed-model repeated-measures ANOVA to detect time-by-group interactions on anxiety and cognitive outcomes, supplemented by hierarchical linear modeling to account for nested training sessions. Regression analyses will assess the predictive power of neurofeedback metrics and emotion regulation scores on anxiety reduction. Mediation analyses will test whether improvements in attentional control and emotion regulation mediate the effect of the intervention on anxiety outcomes. EEG data will be processed with spectral power analyses and event-related potentials where applicable, and multilevel modeling will examine within-subject neurofeedback learning trajectories. Qualitative data from semi-structured interviews and user diaries will be analyzed using thematic analysis to elucidate user engagement, perceived mechanisms, and contextual facilitators or barriers. Integration of quantitative and qualitative findings will follow a convergent design approach for triangulation. Expected findings include greater reductions in state and trait anxiety in the neurofeedback-enhanced group compared with controls, with larger gains in attentional control and emotion regulation and corresponding changes in frontal alpha asymmetry and theta/beta ratios during training. It is anticipated that adherence rates will be higher in the neurofeedback condition due to enhanced biofeedback-driven motivation, leading to more robust transfer to daily functioning. The study contributes to knowledge by empirically validating a scalable ICT-driven intervention that combines neurophysiological feedback with cognitive training for anxiety, offering a theoretically grounded mechanism linking biofeedback, cognitive control, and emotional regulation. It integrates neurofeedback theory with cognitive-behavioral constructs to advance understanding of how real-time neural modulation can enhance cognitive strategies to mitigate anxiety. The main conclusion is that neurofeedback-enhanced cognitive training delivered via mobile apps can produce superior and sustained anxiety reductions and cognitive gains relative to standard app-based cognitive training, with practical implications for scalable, low-cost mental health interventions. Recommendations include refining adaptive difficulty algorithms, incorporating personalization features based on EEG profiles, exploring longer follow-up periods, and assessing applicability in diverse clinical populations, including adolescents and individuals with comorbid mood disorders.