MAPPING AFFECTIVE PROFILES IN DEPRESSION, BURNOUT, NORMAL SADNESS, AND EUTHYMIC STATE: A SELF-REPORT SCREENING TOOL...

Area of Science:

• Psychiatry and Mental Health
• Computational Psychology
• Health Informatics

Background:

• Modern society faces challenges impacting mental health, increasing affective disorders.
• A need exists for sensitive tools to differentiate preclinical and clinical affective states.
• Affective spectrum disorders require accurate and accessible screening methods.

Purpose of the Study:

• To develop a sensitive self-questionnaire for differentiating affective states.
• To utilize machine learning (ML) for accurate classification of mental health conditions.
• To identify key affective symptom domains for improved diagnostic accuracy.

Main Methods:

• A two-stage online survey was developed and administered to university staff and students.
• Machine learning methods were applied for data analysis and respondent classification.
• A final 34-question survey was refined based on statistical analysis and classification tasks.

Main Results:

• Significant differences were found between euthymia, normal sadness, emotional burnout, and depression groups (p < .001).
• The ML-based tool achieved 100% accuracy in classifying affective states in the final iteration.
• The final distribution showed euthymia (38.8%), normal sadness (27.3%), emotional burnout (25.2%), and depression (8.7%).

Conclusions:

• The developed self-report tool accurately classifies the affective spectrum using ML.
• Integration of clinical assessments with ML algorithms enhances diagnostic accuracy and reduces costs.
• Further clinical research is needed to refine the tool's sensitive symptom domains.