An Intelligent Mental Health Identification Method for College Students: A Mixed-Method Study

Area of Science:

• Psychology
• Artificial Intelligence
• Machine Learning

Background:

• Traditional large-scale mental health screening in students is labor-intensive and time-consuming.
• Combining behavioral and facial expression analysis shows promise for mental health assessment.
• Need for efficient and accurate methods to evaluate college students' psychological well-being.

Purpose of the Study:

• To develop an efficient and accurate intelligent method for assessing college students' mental health.
• To integrate artificial intelligence (AI) with traditional psychological assessments.
• To assist in the early diagnosis and treatment of mental health issues in students.

Main Methods:

• A mixed-method approach combining the Depression Anxiety and Stress Scale-21 (DASS-21) questionnaire with facial emotion recognition.
• Facial emotion recognition model developed using transfer learning on neural networks, pre-trained on FER2013 and CFEE datasets.
• Online questionnaires administered to 400 college students, with 374 responses analyzed (350 usable results).

Main Results:

• Facial emotion recognition model's classification results align with mental health survey findings, demonstrating improved efficiency.
• The proposed AI model achieved higher accuracy than traditional single-questionnaire methods.
• Lower absolute errors in detecting depression (0.8%), anxiety (8.1%), and stress (3.5%, 1.8%) compared to existing methods.

Conclusions:

• The combined intelligent and scale-based method for mental health assessment exhibits high recognition accuracy.
• This approach effectively supports efficient, large-scale screening of psychological problems in student populations.
• AI integration offers a promising solution for enhancing the speed and precision of mental health evaluations.