The goal of this observational study is to test an artificial intelligence (AI) tool that can help screen for mental health risks . The main questions it aims to answer are: Can an AI model that analyzes a person's voice, facial expressions, and language accurately identify students who may be at high risk for mental health conditions, such as depression or OCD? How accurate is the AI model when compared to results from standard mental health questionnaires? Participants will be asked to: Complete a standard mental health questionnaire. Provide consent for their data to be used in the research. Participate in a recorded session to collect video and audio data for the AI model to analyze.
This large-scale, multi-center observational study aims to develop and validate a novel artificial intelligence (AI) model for the early and objective screening of mental health risks, such as depression and OCD, in university students. The model will be trained and internally validated on multimodal data (including vocal, facial, and linguistic features) from a large student cohort. A subsequent neuroscience sub-study will explore the neurobiological correlates of the AI-identified risk levels using electroencephalography (EEG) and functional magnetic resonance imaging (fMRI) to establish biological validity. The primary outcome is to assess the final model's diagnostic accuracy, quantified by its sensitivity, specificity, and AUC, with the ultimate goal of providing a scalable and efficient early warning tool to facilitate timely clinical intervention for university populations.
Study Type
OBSERVATIONAL
Enrollment
17,386
An AI model provides an objective and rapid assessment of potential mental health risks in students by holistically analyzing their facial expressions, vocal characteristics, and linguistic content from data.
Peking Union Medical College
Beijing, Beijing Municipality, China
Sensitivity
Time frame: through study completion, an average of 1 year
AUROC
Area Under the Receiver Operating Characteristic Curve
Time frame: through study completion, an average of 1 year
Specificity of the AI Model for Mental Health Screening
The ability of the AI model to correctly identify students without significant psychological distress. It will be calculated as the percentage of participants correctly classified as 'low-risk' by the AI model compared to a 'gold standard' classification
Time frame: through study completion, an average of 1 year
Positive and Negative Predictive Values
Time frame: through study completion, an average of 1 year
Correlation Between AI-Identified Risk Scores and Neurobiological Markers
To assess the biological validity of the AI model, the model's output will be correlated with specific neurobiological markers obtained from a sub-study. The correlation will be assessed using a Pearson correlation coefficient.
Time frame: through study completion, an average of 1 year
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.