188. The intervention effect of artificial intelligence-assisted recognition technology on schizophrenia in college students

Abstract Background Schizophrenia is a serious mental disorder characterized by disordered thinking and impaired social functioning. College students are at a critical stage of psychological development and social adaptation, facing multiple stressors such as academic pressure and social challenges, which may increase the risk of schizophrenia. Traditional treatment methods rely on clinical assessment by professionals, which have problems such as strong subjectivity and poor timeliness. In recent years, Artificial Intelligence (AI) -assisted recognition technology has demonstrated its potential in the early identification and intervention of mental disorders by analyzing multimodal data such as language and behavior. The research aims to explore the intervention effect of AI-assisted recognition technology on schizophrenia in college students, evaluate its practical role in improving symptoms and mental health, and provide technical support for mental health services in colleges and universities. Methods: The study recruited 180 college student patients who met the diagnostic criteria for schizophrenia, aged 18 to 25. Participants were randomly divided into the AI intervention group (n = 90) and the conventional intervention group (n = 90). The AI intervention group received a 16-week AI-assisted intervention, including emotional and cognitive state monitoring based on natural language processing, etc. The conventional intervention group received standard drug treatment. The assessment tools include the Positive and Negative Syndrome Scale (PANSS) and the Social Functioning Scale. The SFS and the General Health Questionnaire-12 (GHQ-12) were evaluated before the intervention (T0), in the middle of the intervention (week 8, T1), and after the intervention (week 16, T2). Data analysis was conducted using repeated measures analysis of variance, with the significance level set at p<.05, and Cohen's d was used to calculate the effect size. Results The total score of PANSS in the AI intervention group decreased from 82.5 ± 8.7 at T0 to 58.3 ± 6.4 at T2 (p<.001, d = 1.25), and both the negative symptom sub-item and the general psychopathological symptom sub-item showed significant improvement (p<.001). The total score of SFS in the AI intervention group showed a significant improvement in social function, increasing from 42.3 ± 5.6 in the T0 stage to 56.8 ± 4.9 in the T2 stage (p<.001, d = 0.92). Meanwhile, the total score of GHQ-12 also decreased significantly (p<.05), indicating an improvement in the overall mental health status. Although the conventional intervention group showed slight improvements in various indicators, the changes did not reach statistical significance (p>.05). Repeated measures analysis of variance further indicated that there was a significant interaction between the intervention time and the group in the total PANSS score (F = 24.18, p<.001) and the total SFS score (F = 19.75, p<.001). Discussion After 16 weeks of intervention with artificial intelligence-assisted recognition technology, college students with schizophrenia showed significant improvements in both symptom severity and social function. The total score of PANSS in the intervention group decreased significantly, the SFS score increased significantly, and the improvement effect was better than that of the conventional intervention group. The results show that AI-assisted recognition technology can effectively assist in the intervention and management of schizophrenia, providing quantifiable support for early identification. Future research can further explore the integration strategies of multimodal data to improve intervention programs and promote the construction of individualized and intelligent mental health service systems.