Individual Differences in the Development of Depth Cue Combination
56.444, Tuesday, May 14, 2:45 - 6:45 pm, Orchid Ballroom
Bauke van der Velde1,2, Tessa Dekker2, Georgina Aisbitt3, Marko Nardini2; 1Cognitive Science Center Amsterdam, University of Amsterdam, The Netherlands, 2Dept of Visual Neuroscience, UCL Institute of Ophthalmology, UK, 3Dept of Experimental Psychology, University of Oxford, UK
Adults integrate multiple depth cues to reduce their discrimination thresholds. In children, however, integration of stereo and texture cues to slant remains immature until 12 years (Nardini et al, PNAS 2010). This suggests that the neural mechanisms underlying integration of depth cues (Ban et al, Nature Neuroscience 2012) are still developing in mid-childhood. Here we investigated whether reliable individual differences in depth-cue integration are present in the "transitional" period of 8 to 12 years. Such differences would allow us to investigate the changes in neural processing of visual information underlying the acquisition of mature cue integration. We tested 25 observers aged 8-12 years in two sessions. Children judged which surface, a 45° standard or a variable comparison, was the most slanted, based on texture, disparity or both. 75% discrimination thresholds were estimated using a Bayesian adaptive staircase based on 80 trials in each condition. Thresholds for all three conditions were strongly correlated across sessions (all ρ > 0.6, p <0.001). To understand the structure of variability across observers, we entered all six thresholds (3 conditions x 2 sessions) into a principal component analysis (PCA). The first component (60% of variance) described variation in global ability, while the second (19% of variance) described variation in cue integration ability, i.e. in the difference between single-cue and combined-cue conditions. The correlations indicate that individual children at this age provide repeatable measures, while the PCA indicates that an important source of variability is in the degree to which they integrate depth cues. These results provide a basis for identifying subgroups of good and poor cue-integrators. In an ongoing follow-up study we are using fMRI and MVPA to relate these individual differences in behaviour to differences in the cortical representation of single vs. combined depth cues.