Fig. 2
From: Soluble amyloid-beta isoforms predict downstream Alzheimer’s disease pathology
Machine learning framework for predicting tau pathology and neurodegeneration. A Cognitively unimpaired (CU) individual’s cerebrospinal fluid (CSF) levels of Aβ1–38, Aβ1–40 and Aβ1–42, demographics data and APOE ε4 status were used for feature generation. B All possible combinations of features were generated using the feature set. C The subsets were used for generating tuned machine learning models validated with nested cross-validation aiming to (D) identify tau pathology (T+) and neurodegeneration (N+) positivity