Quantitative Biology > Neurons and Cognition
[Submitted on 4 Oct 2023 (v1), last revised 10 Jan 2024 (this version, v2)]
Title:Developing a Novel Holistic, Personalized Dementia Risk Prediction Model via Integration of Machine Learning and Network Systems Biology Approaches
View PDF HTML (experimental)Abstract:The prevalence of dementia has increased over time as global life expectancy improves and populations age. An individual's risk of developing dementia is influenced by various genetic, lifestyle, and environmental factors, among others. Predicting dementia risk may enable individuals to employ mitigation strategies or lifestyle changes to delay dementia onset. Current computational approaches to dementia prediction only return risk upon narrow categories of variables and do not account for interactions between different risk variables. The proposed framework utilizes a novel holistic approach to dementia risk prediction and is the first to incorporate various sources of tabular environmental pollution and lifestyle factor data with network systems biology-based genetic data. LightGBM gradient boosting was employed to ensure validity of included factors. This approach successfully models interactions between variables through an original weighted integration method coined Sysable. Multiple machine learning models trained the algorithm to reduce reliance on a single model. The developed approach surpassed all existing dementia risk prediction approaches, with a sensitivity of 85%, specificity of 99%, geometric accuracy of 92%, and AUROC of 91.7%. A transfer learning model was implemented as well. De-biasing algorithms were run on the model via the AI Fairness 360 Library. Effects of demographic disparities on dementia prevalence were analyzed to potentially highlight areas in need and promote equitable and accessible care. The resulting model was additionally integrated into a user-friendly app providing holistic predictions and personalized risk mitigation strategies. The developed model successfully employs holistic computational dementia risk prediction for clinical use.
Submission history
From: Srilekha Mamidala [view email][v1] Wed, 4 Oct 2023 02:47:29 UTC (2,852 KB)
[v2] Wed, 10 Jan 2024 21:08:59 UTC (4,148 KB)
Current browse context:
q-bio.NC
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.