Parkinson’s (PD) disease is a well-studied neurodegenerative disorder that affects between 7 and 10 million people worldwide. Despite the fact that PD has been a recurring topic in the medical literature for more than 200 years, its mechanisms are largely unclear, and existing therapies focus on the symptoms of improve patients.
The most common symptoms of PD include motor problems, including tremors, slowness, and muscle stiffness. These, along with many non-motor symptoms, cause many PD patients to develop facial disorders, such as facial skin problems and problems making facial expressions. Such problems should not be taken lightly, because a person’s face plays a vital role not only in self-esteem, but also when interacting with others. Unfortunately, facial changes in PD patients are difficult to track and measure with conventional methods.
But what if artificial intelligence (AI) could provide the tools we needed? In a recent study published in Brain development, a team of scientists from Okayama University, Japan, tested the feasibility of a modern face detection and analysis algorithm in a PD context. Led by Dr. Koh Tadokoro from the Department of Neurology, the team recruited 193 people, half of whom were PD patients at various stages of the disease and the other half were healthy control subjects. The subjects were taken to a counseling room in the University clinic, and the researchers took one photo of their face without giving any specific instructions.
Once the photos were taken, the team used commercially available AI software to locate the subjects ’faces and explore some of their features; that is, obvious age, strong feeling, and facial skin condition. Through statistical analyzes, they examined whether AI-based algorithms detected significant differences in these effects between the control group and the PD group.
Interestingly, the age gap, i.e. the age at which the software appears to reflect the actual age of the subject, was a factor that differed between the two groups. On average, the age gap for PD patients was higher than for healthy subjects, indicating that the former tended to look slightly older than the correct age in the eyes of the AI. Moreover, PD patients’ facial expressions were more likely to be classified as “meaningless” compared to healthy subjects, who were significantly more likely to carry a phrase classified as “nonsense.” On the other hand, in terms of skin condition, the researchers found no obvious differences between the two groups.
Overall, the strategy used in the study and the results collected could pave the way for AI-based software for future study of PD signals. . The approach taken by the team is relatively simple and cost-effective, as Dr Tadokoro explains: “Although most previous studies have evaluated the face-to-face changes caused PD during movement or actions, we found evidence of specific facial changes based on a single image using modern AI. “In addition, the team also noted some limitations of facial analysis software, such as worse performance in some people for darker skin or being less flawless when they estimated the age of Asians.Such problems can and should be addressed before the software is used in clinical research.
In short, it is possible that AI-based devices carry an undeveloped capability for rapid, quantitative analysis of patients with PD. Excited about the results, Dr Tadokoro says: “We hope our study will accelerate the use of AI technology for the detection and treatment of patients with Parkinson’s Disease and other neurodegenerative diseases. “We hope that future researchers will pay attention to the power of AI in the context of brain disorders to push the field forward to a deeper understanding.