Memory for Scene Details in Eye-Movement Behavior, with and Without Awareness.

Conflicting findings have led to uncertainty about the sensitivity of eye movements to unaware memory when materials are manipulated scenes. Here, we examine several factors (e.g., task instructions, manipulation type) that may determine whether and when unaware memory-based viewing effects are evident and detected in the data. Forty-eight young adults completed two encoding blocks, an indirect test block (i.e., eye-movements used alone to examine memory), and two direct test blocks. Test scenes were new, repeated, or manipulated, with an object removed from, or added to, the picture. Three explicit memory measures (i.e., scene recognition, change identification, change-location specification) were obtained in direct test 1 for indirect test scenes, whereas explicit responses and eye-movement data were obtained concurrently in direct test 2. Manipulated scene trials were categorized as unaware if all three responses were incorrect. Significant unaware viewing effects were evident when memory was tested indirectly and a new object was present in the scene, but not otherwise. We propose that elevated new-object viewing, without change awareness during the indirect test, promotes a reconsolidation process resulting in the development of a misleading familiarity signal. A lesser viewed, and therefore less familiar, object is then selected incorrectly as the new addition when memory is tested directly. In contrast, when changes are (spontaneously) noticed during indirect testing, a separate memory representation is encoded, permitting successful source discrimination later. Our findings reconcile incompatible results that have important implications for theories of hippocampal function and provide new insights into scene processing more generally.