Deep Learning for Predicting Cerebral Metabolism Changes Along the Alzheimer’s Disease Continuum
Fernando García-Gutiérrez, Laura Hernández-Lorenzo, María Nieves Cabrera-Martín, Jordi A. Matias-Guiu, José L. Ayala
Proceedings of the 15th International Multi-Conference on Complexity, Informatics and Cybernetics: IMCIC 2024, pp. 115-122 (2024); https://doi.org/10.54808/IMCIC2024.01.115
|
The 15th International Multi-Conference on Complexity, Informatics and Cybernetics: IMCIC 2024
Virtual Conference March 26 - 29, 2024 Proceedings of IMCIC 2024 ISSN: 2771-5914 (Print) ISBN (Volume): 978-1-950492-78-7 (Print) |
|
Abstract
In recent years, there has been a significant increase in the application of Artificial Intelligence (AI) techniques in Alzheimer’s disease (AD). However, current research primarily focuses on differentiating clinical phenotypes based on cross-sectional designs. In this study, we hypothesize that modeling additional aspects of the disease, such as variations in brain metabolism measured by [18F]- fluorodeoxyglucose positron emission tomography (FDGPET), is possible, and can provide valuable insights into AD progression. For this purpose, we first identified the brain regions with the most pronounced brain hypometabolism in AD. Subsequently, Deep Learning (DL) models, based on feed-forward networks (FFNs) and convolutional neural networks (CNNs), were used to model variations in brain metabolism. Our findings demonstrated the feasibility of predicting trends in brain metabolism along the AD continuum. Overall, this study introduces a novel dimension to predictive modeling in AD, highlighting the relevance of predicting variations in brain metabolism.
|
||