🤖 AI Summary
This study systematically reviews AI applications in schizophrenia rehabilitation from 2012 to 2024, addressing five core clinical challenges: symptom monitoring, medication management, relapse prediction, functional training, and psychosocial support. Synthesizing evidence from 61 empirical studies, the review employs supervised learning algorithms—including SVM and random forests—combined with feature engineering and structured clinical data modeling to construct, for the first time, a comprehensive taxonomy of AI-driven rehabilitation interventions in schizophrenia. Results demonstrate robust performance of AI models in symptom detection and relapse forecasting, achieving mean AUC > 0.85. The study further proposes three translational development pathways: (1) multimodal data fusion, (2) integration of interpretable AI models, and (3) AI-orchestrated closed-loop rehabilitation task delivery. These pathways provide empirically grounded, actionable implementation guidelines to accelerate clinical translation and improve longitudinal recovery outcomes.
📝 Abstract
This systematic review assessed the current state and future prospects of artificial intelligence (AI) in schizophrenia rehabilitation management. We reviewed 61 studies on AI-related data types, feature engineering methods, algorithmic models, and evaluation metrics published from 2012-2024. The review categorizes AI applications into the following key application areas: symptom monitoring, medication management, risk management, functional training, and psychosocial support. Findings indicate that supervised machine learning techniques, particularly for symptom monitoring and relapse risk management, remain the predominant approaches, effectively leveraging structured data while incorporating interpretable algorithms. This study underscores the potential of AI in transforming long-term management strategies for schizophrenia, offering valuable insights into improving the quality of life of patients. Future research should focus on expanding data sources through multimodal data integration, exploring deep learning models, and integrating AI-driven interventions into training tasks to fully capitalize on AI's potential in schizophrenia rehabilitation.