Wise, T., Liu, Y., Chowdhury, F., & Dolan, R. J. (2021). Model-based aversive learning in humans is supported by preferential task state reactivation. Science Advances, 7(31): eabf9616. doi:10.1126/sciadv.abf9616.