No allocentric distractor-location learning following self-movement even with contextual anchors and explicit knowledge
In the authors' words
Adaptive behavior requires learning where relevant information can likely be found, but also where distraction is likely to occur. For such learning to be useful in everyday life, it must generalize across changing viewpoints as observers move through the environment. Recent work suggests a striking asymmetry in this regard: whereas target-location learning can be grounded in world-centered coordinates, distractor-location learning appears tied to the observer's current viewpoint. One possibility is that allocentric distractor-location learning requires stronger contextual support than allocentric target-location learning. Alternatively, distractor suppression may be fundamentally constrained by viewer-centered spatial coding. To distinguish between these possibilities, participants performed an additional-singleton search task displayed on a tabletop and continuously changed their standing position around the table, while the high-probability distractor location remained fixed in the external environment. Unlike previous studies using this paradigm, contextual anchors were embedded directly within the search display, allowing distractor locations to be encoded relative to a stable task-relevant allocentric structure. In a second experiment, participants were additionally explicitly informed about the distractor-location regularity. Despite task-relevant contextual structure and explicit knowledge, distractor-location learning failed to emerge in world-centered coordinates. Together with previous findings showing robust egocentric distractor-location learning using the same set-up, these findings strengthen the view that learned distractor suppression relies primarily on viewer-centered spatial coding, potentially reflecting a process linked to viewpoint-specific action selection.
Appeared: Wednesday, November 18. bioRxiv. Preprint, not yet peer-reviewed.