Cassandra W. Bacher¹, Karsten Babin¹, Marieke Mur¹, Melvyn A. Goodale¹, Jody C. Culham¹; ¹Unversity of Western Ontario, London, Ontario, Canada

Actively manipulating novel objects rather than passively observing them can facilitate object recognition, but the mechanisms behind this effect still remain largely unknown. One potential explanation is that active learning facilitates an iterative process of generating and testing hypotheses about the effects of actions on the objects. We investigated how the gaze strategies utilized during active learning differ from those utilized during passive learning. Of particular interest was the question of whether participants would display any preference for the “leading edge” (the edge of a rotating object where previously unseen features are becoming visible) or the “trailing edge” (the edge of a rotating object where previously visible features are disappearing from view) of an object and whether such a preference would differ between learning conditions. Participants (n = 26) learned novel virtual objects either actively or passively while eye tracking data was collected, then completed an old/new discrimination task. Results indicated that during active learning, participants spent significantly more time looking at the leading than trailing edge (53%:47%). In contrast, participants spent equal amounts of time looking at the leading and trailing edges during passive learning (50%:50%). These results suggest that people may indeed be more likely to generate and test hypotheses about objects that are being learned actively rather than passively. However, no effect of active versus passive learning on recognition accuracy or speed was observed, likely because the objects we used had less distinctive parts than those used in prior studies did (Harman et al., 1999; Curr Biol). Thus, further investigation is needed to determine what role such predictive behavior plays in the “active learning effect” observed in past studies.

Acknowledgements: Natural Sciences and Engineering Research Council of Canada (04271-2022-RGPIN)