CUHK to develop machine learning technologies for advancing digital biomarkers for Alzheimer’s disease

A multidisciplinary team of experts in artificial intelligence (AI), engineering and various medical specialties from the Chinese University of Hong Kong (CUHK) aims to develop AI technology to discover new digital biomarkers for early diagnosis of Alzheimer’s disease (AD). The project received a research grant of HKD 5.6 million from the Diagnostics Accelerator programme of Alzheimer’s Drug Discovery Foundation and a matching fund from SenseTime Technologies.

The team will propose interpretable machine learning algorithms to quantify the correlation of multimodal digital biomarkers from personal and mobile smart devices to assist early detection and accurate diagnosis of AD, as well as subsequent intervention. “Machine learning technique allows us to conduct objective longitudinal monitoring and comparison of indicators, enabling the detection of minimal abnormalities in the nervous system, thus achieving early identification of and intervention in AD,” commented Professor Bolei Zhou from the Department of Information Engineering, CUHK.

“However, the prevalence of sensors will lead to various issues such as privacy concerns and the ‘black box’ nature of AI algorithms. These must be solved in order to interpret the digital markers and their links with disease pathophysiology,” he added. “All of the collected data will be kept on the personal device in order to preserve user privacy.”

The new technologies that are to stem from the project will aim to detect any immediate risks that require prompt attention or action and will identify individuals at higher risk of developing AD and dementia. Early identification of people at risk of developing AD and timely intervention to slow the onset and progression of AD are crucial, because no AD-modifying treatment is available at present.

The proposed digital biomarkers will assess activities of daily living, behavioural and psychological symptoms of dementia, social interactions, motor function, and level of cognition. These accurate, easy-to-perform tests may enhance the effectiveness of screening and identification of AD and address the problem of denial of the illness.

Other planned outcomes include introducing appropriate advice based on the individual’s specific needs, modifying guidance based on response to prior suggestion or intervention, and providing feedback to patients and caregivers in order to improve self-care of the disease and stress adaptation.

“This project will bring together top experts from engineering and medical fields to address one of the biggest challenges faced by our ageing society. The sensing and AI technologies to be developed will empower doctors, caregivers and patients themselves to work together for early diagnosis of and intervention in AD. We believe this project will produce technologies to enable the paradigm of ‘smart health’, which will transform today’s reactive hospital-centered healthcare practice to an approach of proactive, individualized care and well-being,” stated principal investigator, Professor Guoliang Xing from the Department of Information Engineering, CUHK.