Dissociating Principal and Medial Axis Predictions: Effects of Training Enclosure Size

Document Type


Publication Date



Incidental-learning procedures are often used to determine the stimuli that control spatial reorientation in diverse species. A common procedure is to train subjects to approach a discrete landmark prior to testing in the absence of the landmark. Global geometric cues, such as the medial axis of space and the principal axis of space, have both been suggested to account for spatial reorientation in variety of animals (e.g., rats, pigeons and humans). Additionally, the control of spatial reorientation by the global geometric cues has been suggested to vary based upon the size of the enclosure—with smaller enclosures producing greater control by global geometry. In the present study, we trained human participants to approach rotationally-equivalent corners (i.e., north-east and south-west) of a rectangular enclosure in a desktop virtual environment. We manipulated the size of the training enclosure between groups. Following training, all participants were tested in an I-shaped enclosure which dissociated predictions made by medial-axis and principal-axes accounts. Results provide evidence of orientation by principal axis that is dependent on training enclosure size—supporting principal axis accounts of spatial reorientation across species.


Comparative Cognition Society’s International Conference on Comparative Cognition (CCS)


Melbourne, FL