If the pupils move anxiously, it indicates a 'dementia risk group'

Korea Institute of Oriental Medicine
Development of Eye Movement Measurement-Dementia Risk Diagnosis Model

Domestic researchers have developed a technology that identifies dementia risk groups by measuring eye movements.

The Korea Institute of Oriental Medicine has developed a technology that analyzes eye movement data to identify individuals at risk of dementia. The photo shows an actual experiment scene. Photo by Korea Institute of Oriental Medicine

The Korea Institute of Oriental Medicine announced on the 22nd that Dr. Kim Jae-wook's research team from the Digital Clinical Research Department, in collaboration with the Gwangju Dementia Cohort research team, produced research results that more accurately identify dementia risk groups using eye movement data.

Early identification of patients with cognitive impairment who are at high risk of dementia is crucial. While it is difficult to cure once the condition has progressed, in the mild cognitive impairment stage, disease progression can be prevented through appropriate physical exercise, cognitive training, dietary therapy, and cardiovascular function management.

The research team conducted experiments on a total of 594 elderly individuals (428 normal control group, 166 mild cognitive impairment patient group). They collected and analyzed eye movement data of subjects performing simple cognitive tasks on a computer for 5 minutes and developed classification models using machine learning techniques.

Subsequently, the performance of three classification models applying different data combinations was evaluated. The results showed AUROC (Area Under the Receiver Operating Characteristic) scores of 0.752, 0.767, and 0.840, respectively. The AUROC score is an indicator used to evaluate the performance of classification models, with scores closer to 1 indicating better classification performance. Generally, an AUROC score above 0.8 is considered a good performing classification model, and these results demonstrate that using simple eye movement data along with existing information can create more effective classification models.

Dr. Kim stated, “Dementia, which arises due to rapid aging and other factors, is a serious social issue,” and added, “If this research is applied to digital health devices such as VR, it is expected that dementia risk groups can be identified early at dementia safety centers, public health centers, and primary medical institutions, allowing them to participate in various programs for dementia prevention.”

The research results were published on the 15th in the international academic journal Frontiers in Neuroscience (IF 5.152).

This content was produced with the assistance of AI translation services.