Figure 2.

Illustration of how the event-based model (EBM) aims to extract temporal information from a cross-sectional data set. Gaussian distributions of fractional anisotropy (FA) biomarker readings within a tract affected early in the course of the disease, such as the corticospinal tract, would be expected to demonstrate substantial separation between ALS and Control imaging data (A). However, FA from another area affected at later stages demonstrates much less separation between distributions (B). By exploiting and characterizing these differences across all biomarkers, the EBM attempts to order the change from “normal” to “diseased” across the entire disease course.