Probing spatiotemporal context reveals differences between true and false memories

Our memories do not faithfully represent the past, but rather are reconstructive, rendering them susceptible to the intrusion of false memories. However, the nature of the false memories is debated: Do false memories reflect the retrieval of a genuine, erroneous memory trace, or are they the product of semantic associations? This question reflects the fact that although some findings suggest that false memories can be phenomenologically compelling and rich in details, others suggest that there are subtle cues that distinguish true and false memories. Here, we take a novel approach to this question by assessing memory for the spatiotemporal context of false memories — the when and where of those memories. We demonstrate that context memory can provide a probabilistic cue to whether a memory is true or false in the Deese-Roediger-McDermott paradigm. Across participants, the reliability of reported context memory (i.e., the consistency with which people report remembering when or where a false memory occurred) revealed differences between true and false memories. Additionally, confidence in memory for context (both within and across participants) differed between true and false memories — even when controlling for confidence in having encountered the false memory. However, these effects were malleable: Under certain circumstances, false memories appeared to have genuine context traces. Together, these data provide new insights into the circumstances under which true and false memories differ in subtle but meaningful ways, speaking to theories of how false memories may arise from the reconstructive nature of memory.