Mapping Functional Brain Connectivity Alterations in Alzheimer’s Disease with Machine Learning
||Mapping Functional Brain Connectivity Alterations in Alzheimer’s Disease with Machine Learning|
||NAISS Small Compute|
||Jiawei Sun <firstname.lastname@example.org>|
||2023-11-01 – 2024-11-01|
Alzheimer’s disease (AD) is characterized by the accumulation of amyloid-β plaques and tau neurofibrillary tangles. These pathologies spread in a distinct spatial pattern, with evidence suggesting a strong influence of brain connectivity on their progression. Early AD stages show amyloid-β and tau accumulation in functionally connected regions, particularly within the default mode network. However, their effects on functional connections are complex, leading to both strengthened and weakened connections, impacting cognitive functions.
Given the nonlinear nature of these effects, advanced artificial intelligence techniques are essential to decipher these intricate relationships. This project employs transformer neural networks to map nonlinear alterations in functional brain connectivity across aging and AD stages. By analyzing fMRI scans from diverse cohorts, including CamCAN, ADNI, H70, and subjects from the Karolinska University Hospital, we aim to identify the most affected brain connections. Furthermore, we will correlate these changes with amyloid and tau pathologies, explore the potential role of synaptic pruning in AD, and translate our findings into clinical tools for predicting cognitive decline.
In essence, this project bridges advanced machine learning techniques with clinical insights, offering a comprehensive approach to understanding AD's neural intricacies and providing tools for early detection and intervention.