Reproduce “OPTIMUS: Predicting Multivariate Outcomes in Alzheimer's Disease Using Multi-modal Data amidst Missing Values” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Christelle Schneuwly Diaz, Duy-Thanh Vu, Julien Bodelet, Duy-Cat Can, Guillaume Blanc, Haiting Jiang, Lin Yao, Guiseppe Pantaleo and 2 more
The OPTIMUS framework predicts multivariate cognitive outcomes in Alzheimer's disease using diverse multimodal data, effectively managing missing values and providing biological explanations. It demonstrates that integrating structural MRI, blood transcriptomics, APOE genotype, and CSF biomarkers with a TabNet Regressor yields superior predictive accuracy, while explainable AI identifies specific brain regions and genes associated with distinct cognitive deficits.
Key insights
- Multimodal data integration consistently outperforms unimodal approaches in predicting multivariate AD cognitive outcomes, underscoring the complementary value of diverse biological information.
- The TabNet Regressor, a deep learning model, demonstrates superior performance for complex many-to-many prediction tasks on tabular multimodal data, even against strong tree-based models.
- Explainable AI can effectively identify domain-specific neuroimaging, genetic, and transcriptomic biomarkers and map their contributions to distinct cognitive functions, providing nuanced insights into AD pathology.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.