# Facial Emotion Analysis During Rehabilitation Treatment of a Patient With Acute Generalized Peritonitis and Delirium

## TL;DR

This study used facial emotion analysis to assess emotional changes in a delirious patient during rehabilitation, showing increased happiness after treatment.

## Contribution

The study demonstrates the feasibility of using facial emotion analysis as an objective tool for assessing emotions in delirious patients.

## Key findings

- Happy emotion increased significantly after rehabilitation training.
- Surprised emotion decreased significantly following training.
- Happiness showed an increasing trend during delirium after training.

## Abstract

Rehabilitation treatment is known to improve not only physical function but also emotional well-being. However, objectively assessing emotions in critically ill patients with impaired consciousness by using conventional subjective measures is difficult. In this study, we focused on a specific method for analyzing emotions from facial expressions and aimed to perform facial emotion analysis in a single patient with delirium. The patient was a man in his 70s with acute generalized peritonitis who developed postoperative delirium. Rehabilitation treatment was initiated on the first day after surgery, and facial emotion analysis was performed before and after rehabilitation treatment on day 14. Five emotional factors (Neutral, Happy, Sad, Angry, and Surprised) were measured, and 38 measurements were taken during 19 sessions. Visual analog scale (VAS) and salivary α-amylase activity (sAA) levels were also measured 22 times in 11 sessions from day 30 until discharge. Throughout the imaging period, the Happy emotion increased significantly after training when compared with the period before training, whereas the Surprised emotion decreased significantly. Furthermore, during the delirium period, happiness showed an increasing trend following training, whereas other emotional factors tended to decline. The VAS score decreased after training, whereas sAA levels did not show a consistent change. These observations suggest that facial emotion analysis may be feasible as an objective adjunct, but its clinical utility and generalizability require confirmation in larger studies.