Document Type : Original Article

Authors

• Atefe Aghaei ¹
• Mohsen Ebrahimi Moghaddam ²

¹ Computer Science and Engineering Department, Shahid Beheshti University

² Faculty of Computer Science and Engineering Shahid Beheshti University, Tehran, Iran

Abstract

Abstract— Alzheimer’s disease (AD) is a progressive neurodegenerative disorder characterized by cognitive decline and the accumulation of amyloid-beta plaques. Early detection is crucial for timely intervention, and the Aβ42/Aβ40 ratio is a key biomarker for identifying amyloid deposition. In this study, we propose a method to predict the Aβ42/Aβ40 ratio using the extracted features from MRI images using 3D Convolutional Neural Network (3D CNN). Moreover, Random Forest Regression is employed to obtain the relationship between MRI features and the Aβ42/Aβ40 ratio. Our results demonstrate a strong correlation (r = 0.72) between the predicted and actual Aβ42/Aβ40 ratios, effectively predicting amyloid accumulation. This result also makes the proposed feature extraction model more reliable. By leveraging MRI and molecular biomarkers such as the Aβ42/Aβ40 ratio, the proposed method provides valuable insights into disease progression and early diagnosis. By leveraging MRI and molecular biomarkers such as the Aβ42/Aβ40 ratio, the proposed method provides valuable insights into disease progression and early diagnosis.

Keywords

• Keywords—3DCNN
• Alzheimer’s Disease
• Aβ42/Aβ40 ratio
• MRI
• Random Forest Regression

Main Subjects

• Artificial Intelligence, Robotics, Cognitive Computing