Statistical prediction of the future impairs episodic encoding of the present

Significance We use memory to recall the past and make predictions about the future. However, these functions can conflict: Recalling a past experience involves returning to a specific moment in time, but this exact experience will never occur again because the world and observer have changed. Rather, it is often adaptive to focus on aspects of the world that repeat across experiences and are likely to generalize. Encoding individual experiences (episodic memory) and extracting regularities across experiences (statistical learning) require fundamentally different kinds of learning, yet both have been linked to the same brain region, the hippocampus. Here we reveal the consequence of this codependence: Predicting what comes next from statistical learning prevents encoding of the current experience into episodic memory.