The Modulating Role of Immediate Testing and Type of Retrieval

Data on the effect of sleep on false memories are mixed. Here we address two methodological issues which could explain these inconsistencies: the use of free recall vs. recognition tasks and the absence of immediate testing. In a mixed design, 8 word lists from the Deese-Roediger-McDermott task were administered to a Recall (n = 20) and a Recognition Group (n = 20) at 9 AM ('Wake' condition) or 9 PM ('Sleep' condition). Immediate Testing (IT) followed the presentation of 4 out of 8 lists. Delayed Testing was performed 12 h later. A week later, participants underwent the other condition with 8 different lists. In both groups: IT performance was similar between conditions; veridical recalls were facilitated by Sleep and by IT irrespectively of condition. False memories were enhanced in Sleep vs. Wake for non-tested lists only in the Recall Group. IT affected delayed false recognition regardless of the condition, whereas its effects on delayed recall differed between conditions: in Sleep, the false recall rate was similar between tested and non-tested lists, whereas in Wake, it was higher for tested than non-tested lists. Our results confirm that sleep enhances false recalls but not false recognitions when memory is tested directly after the retention interval. Moreover, immediate testing enhances veridical memory independently of condition, whereas its effect on false recall depends on the behavioural state in which retention occurs. Therefore, immediate testing could be usefully introduced in standard sleep-memory studies, while its use should be carefully evaluated in studies on sleep and false memories