Shape of U: The relationship between object-location memory and expectedness
Based on a neuroscientific model of memory (SLIMM), we predicted that people’s memory for object locations would be a U-shaped function of the expectancy of those locations. Using immersive virtual reality, we manipulated expectancy by placing twenty familiar objects in locations within a virtual kitchen that were congruent, unrelated or incongruent with people’s schema (prior knowledge) for a typical kitchen. Using Bayes Factors across three experiments, we confirmed this (pre-registered) prediction, with better memory for highly expected or highly unexpected locations relative to neutral locations. This U-shape was found in location recall and, importantly, in three-alternative forced choice recognition using object-location images, for which locations in the foil images were approximately equally expected. The latter shows that (part of) the U-shape was not simply participants guessing expected locations when unsure. A second prediction of SLIMM is that the two ends of the U-shape would be associated with different expressions of memory: namely, recollection of objects at unexpected locations, but familiarity for objects at expected locations. Bayes Factors provided evidence against this second prediction, with recollection associated with both ends of the U-shape, and familiarity showing no effect of expectancy. These findings have implications for SLIMM and more general theories of the role of schema and surprise in episodic memory.