If neuroscientists who are not memory researchers were asked, “What does the hippocampus do?” they might answer that the hippocampus (HC) is important for memory, and perhaps they might specify that HC is critical for episodic memory (remembering personally experienced past events) but not for implicit (nonconscious) memory or for working memory (briefly maintaining information online). Although this view from the 1990s is still largely accepted, the way many memory researchers conceptualize HC has gradually changed during the last four decades. First, there is now abundant evidence that HC is involved not only in episodic memory but also in implicit memory and working memory (1). As a result, many researchers today think that HC is not a key region of a memory system but a component that mediates a specific cognitive operation, which can be recruited by tasks traditionally associated with different memory systems if the tasks require that specific operation (1, 2). Although the nature of this operation is a topic of active research, a popular hypothesis is that the main function of HC is to associate different kinds of information (3). Of course, no brain region can support a cognitive task by itself; a region must team up with other brain areas, and HC often collaborates with the functionally and structurally connected angular gyrus (AG) (4). A second important development in our conceptualization of HC function was brought by evidence that HC—in collaboration with AG and a subset of default network regions known as the core network —mediates not only remembering past events but also imagining future events (5). This second expansion of HC functions is still within the memory domain, because imagining future events consists mostly in recombining fragments from memories of the past (6). In contrast, the third extension of HC’s “job description” has been … [↵][1]1To whom correspondence may be addressed. Email: cabeza{at}duke.edu. [1]: #xref-corresp-1-1
