My poster will present recent findings using mathematical models of large-scale brain dynamics to interpret EEG signals, with applications to understanding neurological disease. One application will focus on developing methods to forecast seizures in advance and using mathematical models to reveal the mechanistic underpinnings of these forecasts. The second application examines how modellingcan quantify the mechanisms underlying neurodevelopmental adversity. Both applications rely on novel, long-term wearable EEG recordings, which offer a promising alternative to traditional clinical-grade EEG.