Approved Research

Characterizing Alzheimer's Disease and Dementia in South Korean Cohorts by Distillating Knowledge from Large-scale Data

Lay summary

In this project, we will develop Artificial Intelligence (AI) frameworks to understand Alzheimer's Disease (AD). Teaching an AI machine to investigate AD requires large amount of data. Unfortunately, existing AD datasets in South Korea are in small scale, which are not sufficient to develop AI methods. In this regards, we will use the AD data from UK Biobank to first teach an AI machine, and then apply the trained AI machine to investigate AD cohorts in South Korea. For this, we propose the following two aims:

Aim 1: We will develop an AI model to characterize AD using multi-modal data. We will use  both brain images (e.g., MRI) and gene data from the UK Biobank to understand AD.

Aim2: We will develop an AI model to analyze data from multiple sites. As the data collected from different sites may have different site-specific characteristics,  we will develop an AI model that neglects such effect and extract only the AD-related patterns from the data.

The frameworks developed in this project has potential to be adopted the research of other countries with insufficient data and help understanding of AD progression that stem from regional differences.