Cognivia is an evidence-based artificial intelligence therapist that integrates automatic cognitive distortion identification and rational response generation and is proposed the first hierarchical quality evaluation framework for assessing LLM-generated rational responses, developed through collaboration between AI researchers and behavioral science experts.
Abstract
Cognitive distortion amplifies negative emotions and contributes to mental health disorders. Cognitive Behavioral Therapy (CBT) is an effective way to address cognitive distortions, but its large-scale application is limited by the shortage of professional therapists. Although large language models (LLMs) have recently been explored for mental health applications, existing methods still suffer from limited domain specificity, overly flattering responses, and the absence of well-defined annotations for cognitive distortions. This paper proposes Cognivia, an evidence-based artificial intelligence therapist that integrates automatic cognitive distortion identification and rational response generation. Our framework is built on authoritative CBT texts widely regarded as core paradigms and standard references. It is further augmented with mental health question-answer (Q and A) data, and employs multi-stage prompting and structured generation strategies under the supervision of behavioral science experts. Then we fine-tune a lightweight LLM on this augmented CBT dataset to obtain Cognivia. In addition, we propose the first hierarchical quality evaluation framework for assessing LLM-generated rational responses, developed through collaboration between AI researchers and behavioral science experts. Cognivia is evaluated using lexical metrics, LLM-based Judges with two complementary criteria, and human evaluation by 10 behavioral science experts. It consistently outperforms the baseline methods in cognitive distortion recognition and rational response generation, demonstrating its effectiveness. Our code is available at https://github.com/SNOWTEAM2023/Cognivia.
Cognitive Behavioral Therapy (CBT) provides a structured framework for understanding a user's mental state by examining the interaction between cognitive and behavioral factors. However, out-of-the-box LLMs respond fluently and empathetically, yet collapse into validation&reflection, regardless of what the user actually needs. They know theoretical CBT (scoring up to 96% accuracy on licensing exam questions) but fail to apply it effectively. We explore this gap with a knowledge-guided framework that treats CBT dialogue as controlled affective reasoning: user narratives are decomposed into Beck's Cognitive Conceptualization structure, grounded in clinical SNOMED CT concepts validated via Natural Language Inference, and a Multiple Chain-of-Thought (MCoT) strategy selection between Validation&Reflection, Socratic Questioning, or Alternative Perspectives. To measure whether such guidance actually changes behavior, we introduce the Protocol Leverage Force (F), a behavior-level metric that captures how far an intervention shifts a model away from its default response. Across three open-weight LLMs and 14 RealCBT-derived case studies, evaluated with human experts, valence-arousal trajectories, and linguistic entrainment, F shows that simply introducing protocol definitions via single chain-of-thought prompting fails to change LLM behavior, while MCoT on these definitions guides strategy selection better. Still, the effect stays within 1% (approx. 1.2-1.3%), and all models remain biased toward Validation&Reflection. These results show CBT knowledge alone does not ensure effective application, giving the affective-computing community instrumentation to measure where LLMs fall short.
Vaishnavi Sinha, Pooja Guttal, Pranay Deep Reddy Katike et al.· 0 citations
A multimodal emotion-aware architecture, which pays attention to memory-enhanced personalization and emotion-specific reinforcement learning, is introduced and hybrid human-AI approaches, which focus on safety and empathetic conversation to improve current mental health systems are recommended.
This article synthesizes contemporary research on the multifaceted impacts of piano interaction on human cognition, mental health, and neural processes, and explores its translation into intelligent human-computer systems. Evidence from clinical psychology demonstrates that structured piano training can significantly alleviate anxiety and depression in the elderly, enhance executive functions and working memory in aging populations, and serve as an effective component within multi-element interventions for severe mental illness, as exemplified by the GET UP PIANO trial. Neuroscientific investigations reveal that these benefits are supported by a specialized auditory-motor network, encompassing premotor and parietal cortices, which exhibits significant plasticity in experts and is engaged during both music perception and mental imagery. Unique patterns of musical processing in special populations, such as preserved fronto-temporal connectivity for song in autism spectrum disorder, provide a neurobiological rationale for music-based therapies. Critically, these findings are now informing the development of intelligent technologies. We explore how principles of cross-modal correspondence and personalization are being leveraged to create adaptive systems for health-tech, including closed-loop neurofeedback for neuromodulation, smart keyboards for motor rehabilitation using biofeedback, and AI-driven analysis of musical improvisation for mental health assessment. Finally, the review addresses key challenges, including the need for methodological standardization, mechanistic elucidation, and ethical HCI design focused on data security and long-term engagement. The convergence of neuroscience, clinical practice, and technology positions the piano as a powerful and evolving interface for enhancing human health and cognitive resilience.
Depression and anxiety disorders are among the most prevalent and debilitating mental health conditions worldwide, imposing substantial personal, social, and economic burdens. Although recent advances in Large Language Models (LLMs) have shown promise in supporting mental health assessment and intervention, existing approaches often lack contextual awareness, real-time adaptability, and privacy-preserving personalization. To address these limitations, we propose a novel, context-aware and privacy-preserving mental health evaluation architecture that synergistically integrates LLM-driven intelligence. The proposed system enables personalized, continuous, and stigma-free mental health support by combining structured multiple-choice questionnaires with advanced language models, including GPT-3.5-turbo and Groq, to analyze user inputs, identify behavioral patterns, and predict potential mental health conditions such as depression and anxiety. Furthermore, the platform provides individualized recommendations, including self-care strategies, lifestyle adjustments, mindfulness practices, and referrals to healthcare professionals when appropriate. Recognizing the critical importance of reliability in sensitive healthcare settings, we introduce an ensemble-based aggregation framework that explicitly incorporates classification confidence and uncertainty quantification across multiple LLMs. Experimental results demonstrate that the proposed approach outperforms existing LLM models. By prioritizing user anonymity and data privacy, the proposed system reduces psychological barriers to seeking mental health support and promotes early intervention.
Jashraj Jani, Sara Akif, Wassila Lalouani· International Conference on...· 0 citations
Background: Depression and anxiety disorders remain among the leading contributors to global disability and represent a major public health challenge. Although evidence-based psychotherapies are available, access to treatment remains limited due to structural, economic, geographical, and workforce-related barriers. Digital mental health interventions have emerged as scalable approaches to reducing this treatment gap, with artificial intelligence (AI)-guided cognitive behavioral therapy (CBT) representing a rapidly developing and clinically relevant extension of digital psychotherapy. Objective: This review aims to synthesize current evidence on digital and AI-guided CBT interventions for depression and anxiety, with a focus on clinical utility, scalability, mechanisms of change, safety considerations, and public health relevance. In addition, the review proposes a clinically oriented conceptual framework for understanding the role of AI-guided CBT within contemporary digital psychiatry. Methods: A focused narrative review was conducted using PubMed, Scopus, and Google Scholar databases, covering publications from 2010 to 2025. Relevant peer-reviewed studies, systematic reviews, meta-analyses, and conceptual papers addressing digital CBT, AI-assisted CBT, conversational agents, symptom monitoring, and digital mental health implementation were identified and analyzed qualitatively. Results: Existing evidence suggests that internet-delivered CBT, mobile applications, and AI-based conversational agents may reduce depressive and anxiety symptoms, particularly in individuals with mild to moderate conditions. However, the evidence base remains heterogeneous, with limitations including short follow-up periods, variability in intervention quality, reliance on self-reported outcomes, and insufficient data on long-term effectiveness, safety, and real-world implementation. Emerging concepts such as digital therapeutic alliance, continuous symptom monitoring, adaptive intervention delivery, and AI-driven personalization may represent key factors influencing engagement and clinical outcomes. Conclusions: AI-guided CBT represents a promising but still evolving component of modern mental health care. These technologies have the potential to improve accessibility, optimize resource allocation, and support stepped-care and hybrid models of treatment. Future research should prioritize rigorous clinical validation, long-term outcome evaluation, transparent safety protocols, ethical governance, and integration into real-world health systems. AI-guided CBT should not be understood as a replacement for clinicians, but as a complementary and scalable extension of evidence-based psychotherapy.
Aleksandra Stojanović, Miodrag Stanković, Aleksandra Ristic· Healthcare· 0 citations
The rapid expansion of artificial intelligence (AI) in healthcare has prompted growing interest in its application to mental health support. This review compares AI-based mental health tools to human psychotherapy from neuroscientific, computational, and clinical perspectives. The review outlines the structure and evidence-based human therapy, with a focus on cognitive behavioral therapy (CBT) and the therapeutic alliance; then, the mechanisms underlying AI mental health models, including large language models, natural language processing, and training techniques such as reinforcement learning from human feedback are explained. A comparison of the human brain and artificial neural networks, and the analysis of empathy plus emotional processing in both systems, is presented. Evidence suggests that while AI tools can temporarily reduce symptoms and improve accessibility to professional help for mild to moderate conditions, they are less effective in cases of severe or complex disorders. Their simulated empathy, which is strongly associated with the success of treatment, differs from natural human emotions and also contributes to the decrease in effectiveness. Key limitations, including privacy concerns, algorithmic bias, and inadequate crisis handling, are also discussed. The review concludes that a hybrid model integrating AI tools with human-delivered care is the most promising direction for the future of mental health treatment.
Nanxi Zhang· Theoretical and Natural Scie...· 0 citations