KMAN Publication Inc. AI and Tech in Behavioral and Social Sciences 3 Serial Number 11 2025 08 12 The Impact of Artificial Intelligence (AI) on Monitoring Athletes' Mental States: A Machine Learning Approach The Impact of Artificial Intelligence (AI) on Monitoring Athletes' Mental States: A Machine Learning Approach 1 12 10.61838/kman.aitech.3.3.8 EN Journal Article <p>This study aimed to develop and validate an artificial intelligence (AI) framework for monitoring athletes' mental states, addressing critical gaps in traditional assessment methods through advanced machine learning techniques. A mixed-methods longitudinal design was employed with 128 elite athletes, integrating multimodal data streams including physiological (Empatica E4, Polar H10), psychological (APSQ, CAT-MH), and behavioral (facial/voice analysis) measures. A hybrid ensemble model combining Temporal Convolutional Networks with multimodal fusion and explainable AI components was developed and validated through controlled stress induction protocols and ecological momentary assessments over 12 weeks. The AI model demonstrated superior performance (84.7% accuracy) in classifying mental states compared to unimodal approaches, identifying distinct stress phenotypes with differential intervention needs. Real-time feedback reduced acute stress duration by 42.3%, while subgroup analyses revealed gender- and sport-specific stress signatures. The system detected subclinical stress 6.7 minutes before athlete self-report, with strong validation against clinician ratings (κ=0.78) and biochemical markers (cortisol r=0.69). This research establishes that carefully designed AI systems can overcome limitations of conventional athlete mental health monitoring, providing sensitive, actionable insights while maintaining ethical standards. The framework offers sports medicine professionals a transformative tool for early intervention and personalized psychological support, representing a significant advancement in sports science with immediate practical applications. Future work should explore longitudinal implementation across diverse athletic populations.</p> https://journals.kmanpub.com/index.php/aitechbesosci/article/download/4325/7671