Contributions of configuration, individual location, and reference frame in contextual guided visual search

Abstract

The thesis investigated the Contextual Cueing Effect, characterized by enhanced efficiency in visual search through the acquisition of spatial regularities in repeated visual stimuli. To gain deeper insights into the mechanisms behind such effects, this dissertation analyzes various specific factors that could potentially act as contextual cues, directing visual search in repeated displays. The investigation covered four distinct factors: global configuration, individual location, as well as egocentric and allocentric reference frames. All experimental procedures were conducted with the participation of healthy, young individuals, ensuring a robust foundation for exploring the contextual cueing effect. Study I scrutinized the influence of global configuration and individual spatial locations on contextual cueing. During the learning session, participants performed 20 searches of 12 displays to comprehensively familiarize themselves with these displays. Subsequently, the transfer session encompassed four conditions: fully repeated configurations (same as the displays in the learning session), recombined configurations (preserving distractor locations but not configuration by combining distractor locations from two learned displays), rotated configurations (preserving configuration but not distractor locations by rotating displays), and new configurations (randomly selected distractor locations). The reaction times for each condition were collected and analyzed. Building upon study I, study II extended the investigation to the oculomotor dimension with a parallel experimental design. The primary objective of study II was to discern whether learned spatial configuration or individual locations prompted scanpaths similar to those observed in the original displays. The combined findings from studies I and II revealed that participants exhibited accelerated target detection in both fully and partially repeated displays, indicating contribution from both global configuration and individual location to contextual cueing. Moreover, study II highlighted that the number of fixations and scanpath pattern ratio were higher in the new displays compared to the other three repeated displays. Notably, significantly high similarities were observed in the fixation density map between recombined and fully repeated displays, as well as between rotated and fully repeated displays when the rotated display was back-rotated. Overall, this demonstrated a remarkably flexible use of the oculomotor system for search in partially repeated displays. Study III explored the influence of egocentric and allocentric reference frames on contextual cueing. Participants learned various combinations of distractor-target configurations and external frame orientations during a training phase. In the subsequent testing phase, either the frame orientation or the configuration underwent rotation, disrupting either the allocentric or egocentric predictions for the target location. Despite these manipulations, contextual cueing persisted. However, when both reference frames were simultaneously invalidated for predicting target location, contextual cueing was eliminated. This highlights the role of both egocentric and allocentric reference frames in providing effective search guidance in repeated contexts, as long as they contain valid information about the search goal. Collectively, the three studies presented in this thesis offer valuable insights into a nuanced understanding of contextual cueing, shedding light on the roles of global and local factors, as well as egocentric and allocentric reference frames in facilitating efficient visual search. The findings carry implications for theories of spatial attention and memory, enriching our comprehension of the cognitive mechanisms underlying the contextual cueing effect.