Research and development of a machine learning model and a self-check application for simple detection of subtle cognitive decline in MCI

Dementia is a significant challenge in Japan’s aging society. Early detection of its precursor, mild cognitive impairment (MCI), is crucial as timely intervention can delay its progression. However, conventional assessment tools are often impractical for daily use due to their length and the need for skilled administrators.

The objective of this project is to develop an AI-powered healthcare application that enables self-monitoring of cognitive decline associated with MCI. The application, designed for use with smartphones, features a fully automated check-up in just 3–5 minutes, with users simply responding to voice prompts. This revolutionary approach eliminates the need for a tester and complex operations, overcoming the shortcomings of conventional tools. We aim for high accuracy, with both sensitivity and specificity exceeding 0.8, enabling early MCI detection in daily life. The app also generates new content for each test, making it suitable for continuous monitoring.

The app encourages timely lifestyle improvements and medical consultations by fostering early awareness of cognitive decline. This facilitates the early use of new Alzheimer’s drugs, contributing to the prevention of dementia and the extension of a healthy lifespan. By helping the elderly maintain their independence, it also reduces the burden on families and carers.

This project will also expand the applications of AI in dementia prevention, potentially creating a new digital health market. Ultimately, the application can serve as a health support tool for entire communities, contributing to the development of a sustainable aging society.