Several studies have reported that spontaneous eye movements occur with visual imagery and that they closely reflect content and spatial relations from an original picture or scene (e.g., Brandt & Stark, 1997; Spivey & Geng, 2001; Johansson, Holsanova, & Holmqvist, 2006). Nevertheless, the exact purpose of these eye movements to “nothing” is elusive and has currently been the target of a hot topic of debate (cf., Ferreira et al., 2008; Richardson et al., 2009). Do they have an active and functional role when visuospatial memories are retrieved or are they merely an epiphenomenon? In a recent study we reported that when eye movements were prohibited for participants who orally described pictures from memory, their recollections became altered and impaired (Johansson, Holsanova, Dewhurst, & Holmqvist, 2011). The current study was designed as a follow-up, with the purpose to uncover exactly how imposing different eye movements on participants affect memory retrieval of visuospatial memories.
Eye movements were recorded – using a SMI RED 500-system – from 16 participants during an experiment where sets of objects were visually encoded and subsequently retrieved from memory. In the encoding phase, the participants encoded 24 objects in different locations on a computer screen. In the retrieval phase, they listened to pre-recorded spoken statements that either dealt with a property of an object – intra-object – or with the spatial arrangement between two objects – inter-object. The participants were instructed to orally decide whether those statements were true (by saying ‘yes’) or false (by saying ‘no’). The retrieval phase was divided into blocks of four different conditions: (1) free viewing on a blank screen; (2) gazing at a fixation cross; (3) looking at an area which was matched with the original location of the object(s) to be recalled; (4) looking at an area which did not match the original location of the object(s) to be recalled. Over the entire experiment each participant responded to 192 statements.
The data was analyzed within-subjects over the four conditions in respect to reaction time (RT) for correct responses. The eye movement data was used to verify if the participants were able to comply in the conditions when they were restricted to look at the fixation cross or inside the matched/non-matched area. If not, those trials were excluded.
Results revealed a significant main effect for RT in regard to inter-object statements but not for intra-object statements. Post-hoc comparisons revealed that looking at the fixation cross and looking at the area which did not match the original location of the objects to be retrieved yielded significantly longer RT when compared to free viewing and looking at the area which matched the original location of the objects to be retrieved.
Consequently, these results demonstrate that eye movements to “nothing” do indeed have an active and supportive role when visuospatial information is retrieved from memory and show that those eye movements primarily influence processes that integrate spatial properties between objects.