Greebles in 3D: Depth sensitivity is altered by classification training involving novel objects

Abstract

Previous work has indicated that depth-position judgment can be influenced by the object’s identity, raising an intriguing question as the role of object expertise in defining stereoscopic depth retrieval. In this study, we investigated whether training that attaches meaning (i.e., classification labels) to otherwise novel objects (i.e., Greebles), changes observers’ sensitivity for judging their depth position. Observers (N = 20) completed an experiment comprising a pre-test session, 3-5 training sessions (depending on performance attainment), and a post-test session. In the pre-test and post-test stages, participants were presented with disparity-defined Greebles and asked to complete a noise-based depth position (near versus far) judgment task. In the training stage, participants were trained to associate disparity-defined Greebles with given names and gender. Results showed that depth discrimination thresholds were significantly lower (i.e., better) following Greeble-classification training, than before training, despite the fact that object identity is irrelevant to judging depth-position. Performance did not change between “pre” and “post” tests for two additional groups: (1) a group (N = 13) who completed tests and re-tests without classification-training; and (2) another group (N = 13) who completed tests and re-tests, but trained on an orientation task (using identical stimuli), rather than on the Greeble-classification task. Our data suggest that depth sensitivity can be actively altered through acquiring object meaning for otherwise nonsensical objects, highlighting a robust interplay between object recognition mechanisms and stereoscopic outcomes.