AI-POWERED TOOL DEVELOPED TO DECODE ALZHEIMER’S DISEASE
In the past few years, there has been an array of innovative and insightful research projects funded by the Ministry of Higher Education, Research and Innovation. ‘A customised machine learning and deep learning model for predicting the risk of Alzheimer’s disease at an early stage’ by principal investigator Dr Abraham Varghese, Senior Lecturer at the IT Department, University of Technology and Applied Sciences Muscat, is among the research projects funded by the Block Funding Programme of the Ministry of Higher Education, Research and Innovation.
In this research project, Dr
Abraham Varghese explained that Alzheimer’s disease (AD) has become one of the leading causes of death worldwide. The number of people suffering from AD is expected to rise from 55 million to 139 million by 2050. A sharp increase in Alzheimer’s disease cases necessitates the development of immediate early diagnostic tools, something that is also true in the Sultanate of Oman, where the number of patients with Alzheimer’s disease is on the rise.
The study therefore aimed to develop an Ai-driven diagnostic tool for Alzheimer’s Disease (AD) that leverages both psychological parameters and image features derived from clinical and MRI measurements, integrate Explainable AI (XAI) techniques into the diagnostic tool to enable practitioners and researchers to understand the reasoning behind the model’s decisions, and develop a user-friendly graphical user interface (GUI) for the diagnostic tool, which facilitates model predictions and explanations, thereby supporting informed clinical decisionmaking.
According to Dr Abraham, the study emphasised the development of Ai-based tools for the early diagnosis of AD, focusing on the stage of Mild Cognitive Impairment (MCI) to prevent the complete neurodegeneration of Alzheimer’s progression. To understand how AD affects the brain over time, the research team examined MRI scans along with psychological and demographic information from the Alzheimer’s Disease Neuroimaging Initiative (ADNI).
Dr Abraham added that after rigorous preprocessing and application of feature selection algorithms, a set of black box algorithms was applied, and Random Forest was selected as having the highest accuracy of 92 per cent. Incorporating Explainable AI techniques, the research team enhanced model transparency, enabling medical professionals to gain clear insights into the predictive outcomes.
Dr Abraham and his team also developed a web application that enables medical experts to use the model effectively. This tool facilitates the provision of personalised care and quality treatment by offering accurate, Ai-driven insights into each patient’s condition. Through this initiative, the team aimed to enhance clinical decisionmaking processes, bridging the gap between advanced AI technologies and their practical application in healthcare settings for Alzheimer’s disease.
This research project was published in the peer-reviewed journal Alzheimer’s & Dementia in 2023. The research project was conducted by Dr Abraham Varghese, Dr Vinu Sherimon, Dr Ben George, Aisha bint Khalid al Hinaiyah and Dr Prashanth Gouda.