A technology has been developed that can detect the pre-dementia stage at an early stage using big data, and it is expected to be of great help in preventing dementia in the future.


The Korea Electrotechnology Research Institute (KERI) announced on the 7th that the research team led by Dr. Park Young-jin of the Auditory Cognitive Brain Function Research Team of the Electromedical Device Research Group is receiving a lot of attention for their challenge in developing technology that can detect high-risk groups for 'mild cognitive impairment', a common pre-dementia stage, at an early stage through analysis of everyday conversations in the elderly.

With the rapid aging of the population, the prevalence of dementia continues to increase. According to the 'Central Dementia Center Annual Report' recently published by the Central Dementia Center, as of 2023, 1 in 10 elderly people aged 65 or older in Korea will suffer from dementia.

According to the Ministry of Health and Welfare's 4th Comprehensive Dementia Management Plan, this number is expected to continue to increase, reaching 1.36 million in 2030 and 3.02 million in 2050.

Dementia is emerging as a global issue that must be addressed, with the number expected to reach 75 million worldwide by 2030.

Alzheimer's disease, a degenerative brain disease that typically causes dementia, progresses over a long period of time, but in many cases, symptoms are not discovered in the early stages and the disease enters the dementia stage, or patients visit local government dementia centers or hospitals in person to receive a belated diagnosis.

Furthermore, although it would be good for seniors living alone to visit the center every year to receive care, many of them do not visit the center due to physical difficulties and discomfort, and thus fall into a blind spot of care.

To address this situation, KERI is carrying out the 'Development of AI-based degenerative brain function decline assessment technology through the establishment of big data on everyday speech of the elderly' project as the general organization with the support of the National Research Council of Science and Technology's Creative Convergence Research Project.

KIST, ETRI, Seoul National University Hospital, and Ewha Womans University are participating as joint research institutes and contracted research institutes.

Utterance is the expression of language through speech, such as reading a book or answering a question. The business unit is developing a 'senior-friendly speech data collection device', and is collecting big data (speech, auditory cognitive brain waves, hearing) from it, and is using artificial intelligence (AI) to select and monitor elderly people at high risk of mild cognitive impairment. The reason why screening and management of high-risk groups are necessary at the mild cognitive impairment stage is that while the incidence of dementia in normal people over 65 years of age is 1-2% per year, that of patients with mild cognitive impairment is 10-15%. Experts report that if long-term follow-up is conducted for up to 6 years, 80% of patients with mild cognitive impairment will develop dementia.

The goal of the technology is convenience and accuracy. All you have to do is wear a light hearing aid-like device and install a neurocognitive function test app. The business group plans to analyze the speech paradigms that are mainly used in everyday life environments through this app and screen high-risk groups for degenerative brain function decline with an accuracy (sensitivity) of over 80% using only information on the average 20 conversational turns.

In particular, the speech of the elderly is more difficult to recognize due to pronunciation disorders, and there are more difficulties such as using dialects or not being able to hear questions properly due to hearing loss. The business unit is continuously solving these problems through AI and auditory recognition decoding technology.

KERI is currently conducting a verification test on approximately 100 elderly people in the community, including the Sangnok-gu Senior Welfare Center in Ansan-si, Gyeonggi-do, using the developed research results, and has so far successfully screened 6 patients with mild cognitive impairment and 7 suspected patients. In addition, by August of this year, we plan to further enhance healthcare support and technology for the elderly living in Ansan City by conducting a verification test on an additional 150 elderly people at welfare centers. In the future, we aim to expand the target range to 1,000 people by discovering local governments that wish to conduct a verification test.

Furthermore, the technology can be applied to manage dementia crises in advance and delay symptoms by linking it to personalized cognitive function improvement based on collected data, so it is expected to have a huge ripple effect.

KERI Dr. Park Young-jin said, “If we can advance the treatment period by just one year through early detection of dementia, we can reduce the medical expenses of tens of millions of won per person, and we can also expect an astronomical reduction in national and social costs. Above all, there is the advantage of being able to spend time with our families for a longer time.” He added, “The project team is conducting research to improve hearing cognitive function based on hearing augmentation devices and increase the accuracy of neurocognitive function testing devices using generative AI.” He continued, “We will produce good results so that it will be possible to screen high-risk groups for mild cognitive impairment by conveniently participating in a short test at home.”

Meanwhile, KERI is a government-funded research institute under the National Research Council of Science and Technology of the Ministry of Science and ICT. Currently, the institute is continuously developing easy and convenient 'high-quality digital healthcare devices for the elderly' to support the happy life of the elderly in the super-aged society and to solve national and social problems such as hearing loss and dementia, centered on the Electromedical Device Research Group.