Leon Aksman

FastEBM improves disease progression modeling by using a diffusion-based Markov framework, achieving better accuracy and speed than existing methods in both synthetic and real-world data.

This study analyzes fMRI-complexity trajectories across the lifespan using multiscale sample entropy and finds a nonlinear peak in brain complexity around age 23, with gender differences and associations with executive function.

The paper identifies a tau-first subtype of Alzheimer's disease through analysis of post mortem and in vivo data, showing higher soluble TREM2 levels in tau-first cases.

The paper builds genetically adjusted hippocampal volume nomograms using polygenic scores and evaluates their clinical impact in detecting dementia patients.

pySuStaIn provides a Python implementation of the SuStaIn algorithm to infer disease subtypes and stages from cross-sectional data.

This paper introduces a Bayesian multi-task learning model for longitudinal imaging biomarkers, showing improved trajectory inference and disease effect detection in Alzheimer's disease data.