Okayama Researchers Use Facial Recognition to Study Effects of Parkinson’s

A team of Okayama University researchers have been exploring the use of facial recognition to gain a better understanding of Parkinson’s disease. In cases of Parkinson’s, the patient’s loss of fine motor skills often extends to the muscles in the face. The symptom is known as hypomimia, and it can give people with the condition an expressionless, mask-like appearance because their facial muscles cannot display many common emotions.

With that in mind, the researchers set out to determine whether or not facial recognition would produce any meaningful insights when applied to people with Parkinson’s. The study included 97 patients with Parkinson’s and 96 healthy control subjects, and used facial recognition software to classify each participant based on characteristics like age, emotion, and gender.  

In doing so, the researchers found that people with Parkinson’s tended to appear older and were slightly less expressive than those without the disease. On average, the software assumed that people with Parkinson’s were 2.4 years older than their true age. That number jumped to 3.4 years for men, though the gap was smaller for more elderly patients.

The researchers also found modest evidence of hypomimia. People with Parkinson’s were more likely to have an expressionless face (89 percent instead of 77 percent). Meanwhile, only five percent of those with Parkinson’s were classified as happy, compared to 19 percent of those in the control group. However, there were no significant differences with regards to other emotions, nor was there any correlation between Parkinson’s and the health of a subject’s skin.

Given those limitations, the researchers concluded that facial recognition is not yet ready to be deployed in a clinical Parkinson’s environment. They noted that racial and gender bias could have affected the results of the study, and could create ethical concerns that would need to be addressed in the future.

The Okayama study was conducted using publicly available facial recognition software.

–

March 30, 2021 – by Eric Weiss