Artificial Intelligence-Derived Emotionality State Score Correlates with Layperson Assessment and Objective...

Area of Science:

• Medical technology
• Artificial intelligence in healthcare
• Facial paralysis research

Background:

• Growing interest in AI for tracking facial paralysis (FP) patient outcomes.
• Limited validation studies correlating AI output with established outcome measurements.
• Need for objective and automated assessment tools in facial reanimation.

Purpose of the Study:

• To correlate an AI-derived emotionality state score (ESS) with layperson assessment (Terzis score).
• To correlate ESS with objective oral commissure excursion measurements.
• To evaluate AI's utility in assessing facial reanimation outcomes.

Main Methods:

• Retrospective cohort study of 63 patients with FP undergoing facial reanimation.
• Analysis of voluntary smile videos using FaceReader software to derive ESS.
• Comparison of AI ESS with Terzis scores from lay observers and Emotrics software measurements.

Main Results:

• Significant postoperative improvement in ESS and reduction in negative emotion scores.
• Strong correlation between AI ESS and Terzis score (r=0.78, p<0.001).
• Moderate correlation between AI ESS and oral commissure excursion (r=0.59, p<0.001).
• AI demonstrated better performance in patients with complete paralysis.

Conclusions:

• AI-driven emotionality recognition shows promise for automating layperson assessments.
• AI contributes to objective evaluation of facial reanimation outcomes.
• Future research should focus on AI training and multicenter validation studies for FP patients.