Abstract
Context shift decrement refers to the finding that memory for faces is undermined due to the alteration of scene contexts between encoding and retrieval. This effect has been interpreted as indicating automatic binding between a face and the context against which it was presented during encoding despite no intention to associate them. Here we examined the extent to which such an effect can be modulated by orienting attention. During encoding, participants oriented their attention either to faces and rated their personality traits (i.e., extrovert vs. introvert), or to scenes and rated their semantic attributes (i.e., indoors vs. outdoors). They performed the same task for both singly displayed face and scene images, which served as control, and face-scene compound images. During recognition, participants were shown face and scene images alone and judged whether they encountered the specific images during encoding. Brain images were collected using an fMRI scanner during both encoding and recognition. For faces, behavioral results showed no indication of automatic binding between the face and scene context in that activation of the hippocampus (HPC) failed to differ as a function of encoding success activity (ESA) (i.e., hit-miss) between the context-shift and control conditions. Furthermore, activation of the parahippocampal cortex (PHC) failed to show signs of reactivating encoded scenes in recognition, perhaps due to the lack of automatic binding during encoding in the first place. Finally, the context shift decrement of memory failed to correlate with measures of brain activation in regions presumably underpinning automatic binding. Analogously, for scenes, we only found the left HPC exhibited greater activation with respect to retrieval success activity (RSC) for those that were attended when viewing the face-scene compound images during encoding. Taken together, these findings suggest orienting attention to faces and scenes may have diminished the likelihood of automatic binding between them.