Modelling Alzheimer's Protein Dynamics: A Data-Driven Integration of Stochastic Methods, Machine Learning and Connectome Insights
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866909377720483840 |
|---|---|
| author | MacIver, Alec Shaheen, Hina |
| author_facet | MacIver, Alec Shaheen, Hina |
| contents | Alzheimer's disease (AD) is a complex neurodegenerative disorder characterized by the progressive accumulation of misfolded proteins, leading to cognitive decline. This study presents a novel stochastic modelling approach to simulate the propagation of these proteins within the brain. We employ a network diffusion model utilizing the Laplacian matrix derived from MRI data provided by the Human Connectome Project (https://braingraph.org/cms/). The deterministic model is extended by incorporating stochastic differential equations (SDEs) to account for inherent uncertainties in disease progression. Introducing stochastic components into the model allows for a more realistic simulation of the disease due to the multi-factorial nature of AD. By simulation, the model captures the variability in misfolded protein concentration across brain regions over time. Bayesian inference is a statistical method that uses prior beliefs and given data to model a posterior distribution for relevant parameter values. This allows us to better understand the impact of noise and external factors on AD progression. Deterministic results suggest that AD progresses at different speeds within each lobe of the brain, moreover, the frontal takes the longest to reach a perfect disease state. We find that in the presence of noise, the model never reaches a perfect disease state and the later years of AD are more unpredictable than earlier on in the disease. These results highlight the importance of integrating stochastic elements into deterministic models to achieve more realistic simulations, providing valuable insights for future studies on the dynamics of neurodegenerative diseases. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2411_02644 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Modelling Alzheimer's Protein Dynamics: A Data-Driven Integration of Stochastic Methods, Machine Learning and Connectome Insights MacIver, Alec Shaheen, Hina Neurons and Cognition Alzheimer's disease (AD) is a complex neurodegenerative disorder characterized by the progressive accumulation of misfolded proteins, leading to cognitive decline. This study presents a novel stochastic modelling approach to simulate the propagation of these proteins within the brain. We employ a network diffusion model utilizing the Laplacian matrix derived from MRI data provided by the Human Connectome Project (https://braingraph.org/cms/). The deterministic model is extended by incorporating stochastic differential equations (SDEs) to account for inherent uncertainties in disease progression. Introducing stochastic components into the model allows for a more realistic simulation of the disease due to the multi-factorial nature of AD. By simulation, the model captures the variability in misfolded protein concentration across brain regions over time. Bayesian inference is a statistical method that uses prior beliefs and given data to model a posterior distribution for relevant parameter values. This allows us to better understand the impact of noise and external factors on AD progression. Deterministic results suggest that AD progresses at different speeds within each lobe of the brain, moreover, the frontal takes the longest to reach a perfect disease state. We find that in the presence of noise, the model never reaches a perfect disease state and the later years of AD are more unpredictable than earlier on in the disease. These results highlight the importance of integrating stochastic elements into deterministic models to achieve more realistic simulations, providing valuable insights for future studies on the dynamics of neurodegenerative diseases. |
| title | Modelling Alzheimer's Protein Dynamics: A Data-Driven Integration of Stochastic Methods, Machine Learning and Connectome Insights |
| topic | Neurons and Cognition |
| url | https://arxiv.org/abs/2411.02644 |