Common and distinct neural systems support the generation retrieval phase of autobiographical memory and personal...

Prior research has documented engagement of a common 'core' retrieval network during autobiographical memory retrieval and higher-order prospective tasks, such as personal problem solving. This neural overlap has overwhelmingly been documented in the context of the 'elaboration phase' of retrieval, when a single mental event is simulated in detail . However, recollective and prospective tasks are often associated with generic cues, which require the retrieval and consideration of multiple conceptually-related events. This initial 'generation phase' of retrieval has received comparably little attention in the literature, leaving open questions as to how and when autobiographical memory and prospective tasks overlap within the brain. Here, we compare and contrast neural activity between autobiographical memory retrieval and personal problem solving with a focus on the initial generation phase of retrieval. In the MRI scanner, young adults completed both an autobiographical memory and a personal problem solving task. Each task consisted of a generation phase, which required participants to generate multiple past personal events or problem solutions to a given cue and a subsequent elaboration phase, where a single memory or solution was simulated in detail. A multivariate Partial Least Squares analysis revealed patterns of neural overlap between memory and problem solving during the generation phase that were distinct from the elaboration phase. Among regions commonly recruited during the generation phase was the anterior hippocampus, a structure involved in initiating mental construction and integrating concepts. Subsequent analyses demonstrated that the anterior hippocampus interacted with distinct cortical regions as a function of task, in particular the ventromedial prefrontal cortex. Together, these data provide novel evidence that neural overlap between autobiographical memory and personal problem solving does not occur solely in the context of detailed simulation but, instead, is driven by common retrieval demands.