Two experiments tested the hypothesis that it is easier to bind a stimulus to context when the stimulus already has a stable (i.e., pre-existing) memory representation by comparing episodic memory of faces of celebrities vs. unknown individuals. Each face was superimposed on a picture of a well-known location (e.g., Eiffel Tower) during encoding and at a later unexpected recognition test but the background could change from encoding to test. Although recognition was to be based on the face, irrespective of background, performance was better when encoding context was reinstated. Further, a given background could be shown with many faces ("high fan") or only a few ("low fan") and this variable modulated the value added of context reinstatement. Importantly, manipulations of context only mattered for famous faces. As predicted, these effects were observed in recollection ("Remember") responses not in familiarity (“Know”) responses. Experiment 2 used the same design except that half of the subjects were administered midazolam, a drug that produces temporary anterograde amnesia, prior to encoding faces and backgrounds. Subjects injected with saline (control condition) showed the same pattern as Experiment 1; however subjects injected with midazolam showed a large decrease in the use of the "Remember" responses for famous faces and neither context reinstatement nor background fan affected performance. These results support the view that it is easier to bind stimuli to context when stimuli have a pre-existing, stable memory representation (e.g., faces of people whose identity we know) than when stimuli do not have pre-existing, stable memory representations.

No abstract

No abstract