Don’t predict if you cannot interpret: investigating the clinical viability of facial movements for machine-learning assisted diagnostics of bipolar disorder

Key Points

• The study aims to evaluate the interpretability of machine learning models in diagnosing bipolar disorder using facial movements.
• Exploratory analysis of facial movement data related to bipolar disorder
• Identification of AU4 as a discriminative feature
• Discussion of clinical implications for diagnosis
• Establishment of AU4 as a significant indicator for bipolar disorder
• Highlighting the need for interpretable machine learning models in clinical settings
• Emphasis on the value of specific facial movements for diagnostics