We’ve made progress in understanding how memory systems support real-world event comprehension. Yet we still lack computational models that generate precise predictions about how episodic memory (EM) will be used when processing naturalistic, high-dimensional stimuli. (2/n)