Contour interpolation automatically binds targets with distractors to impair multiple object tracking (Keane, Mettler, Tsoi, & Kellman, 2011). Is interpolation special in this regard, or can other features produce the same effect? To address this question, we examined the influence of eight features on tracking: color, contrast polarity, orientation, size, shape, depth, interpolation and a combination (shape, color, size). In each case, subjects tracked 4 of 8 objects that began as undifferentiated shapes, changed features as motion began (to enable grouping), and returned to their undifferentiated states before halting. The features were always irrelevant to the task instructions. We found that inter-target grouping improved performance for all feature types, except orientation and interpolation (Experiment 1 and Experiment 2). Most importantly, target-distractor grouping impaired performance for color, size, shape, combination, and interpolation. The impairments were at times large (>15% decrement in ac curacy) and occurred relative to a homogeneous condition in which all objects had the same features at each moment of a trial (Experiment 2) and relative to a “diversity” condition in which targets and distractors had different features at each moment (Experiment 3). We conclude that feature-based grouping occurs for a variety of features besides interpolation, even when irrelevant to task instructions and contrary to the task demands, suggesting that interpolation is not unique in promoting automatic grouping in tracking tasks. Our results also imply that various kinds of features are encoded automatically and in parallel during tracking.

Multiple object tracking (MOT) is an attentional task wherein observers attempt to track multiple targets among moving distractors. Contour interpolation is a perceptual process that fills-in nonvisible edges on the basis of how surrounding edges (inducers) are spatiotemporally related. In five experiments, we explored the automaticity of interpolation through its influences on tracking. We found that (1) when the edges of targets and distractors jointly formed dynamic illusory or occluded contours, tracking accuracy worsened; (2) when interpolation bound all four targets together, performance improved; (3) when interpolation strength was weakened (by altering the size or relative orientation of inducing edges), tracking effects disappeared; and (4) real and interpolated contours influenced tracking comparably, except that real contours could more effectively shift attention toward distractors. These results suggest that interpolation's characteristics-and, in particular, its automaticity-can be revealed through its attentional influences or "signatures" within tracking. Our results also imply that relatively detailed object representations are formed in parallel, and that such representations can affect tracking when they become relevant to scene segmentation.

Although much recent work in perceptual learning (PL) has focused on basic sensory discriminations, recent analyses suggest that PL in a variety of tasks depends on processes that discover and select information relevant to classifications being learned (Petrov, Dosher, & Lu, 2005; Kellman & Garrigan, 2009). In complex, real-world tasks, discovery involves finding structural invariants amidst task-irrelevant variation (Gibson, 1969), allowing learners to correctly classify new stimuli. The applicability of PL methods to such tasks offers important opportunities to improve learning. It also raises questions about how learning might be optimized in complex tasks and whether variables that influence other forms of learning also apply to PL. We investigated whether an adaptive, response-time based, category sequencing algorithm implementing laws of spacing derived from memory research would also enhance perceptual category learning and transfer to novel cases. Participants learned to classify images of 12 different butterfly genera under conditions of: 1) random presentation, 2) adaptive category sequencing, and 3) adaptive category sequencing with ‘mini-blocks’ (grouping 3 successive category exemplars). We found significant effects on efficiency of learning for adaptive category sequencing, reliably better than for random presentation and mini-blocking (Experiment 1). Effects persisted across a 1-week delay and were enhanced for novel items. Experiment 2 showed even greater effects of adaptive learning for perceptual categories containing lower variability. These results suggest that adaptive category sequencing increases the efficiency of PL and enhances generalization of PL to novel stimuli, key components of high-level PL and fundamental requirements of learning in many domains.

Understanding and optimizing spacing of learning events is a central topic in basic research in learning and memory and has widespread and substantial implications for learning and instruction in real-world settings. Spacing memory retrievals across time improves memory relative to massed practice – the well-known spacing effect. Most spacing research has utilized fixed (predetermined) spacing schedules. Some findings indicate advantages of expanding spacing intervals over equal spacing (e.g., Landauer & Bjork, 1978); however, evidence is mixed (e.g., Karpicke & Roediger, 2007). One potential account of differing findings is that spacing per se is not the primary determinant; rather learning may depend on interactions of spacing with an underlying variable of learning strength that varies for learners and items. If so, learning may be better optimized by adaptive schedules that change spacing in relation to a learner’s ongoing performance. In two studies, we investigated an adaptive spacing algorithm, Adaptive Response-Time-based Sequencing (ARTS; Mettler, Massey & Kellman, 2011) that uses response time along with accuracy in interactive learning to generate spacing. In Experiment 1, we compared adaptive scheduling with fixed schedules having either expanding or equal spacing. In Experiment 2, we compared adaptive scheduling to two fixed “yoked” schedules that were copied from adaptive participants; these equated average spacing and trial characteristics across conditions. In both experiments, adaptive scheduling outperformed fixed conditions at immediate and delayed tests of retention. No evidence was found for differences between expanding and equal spacing. The advantage of adaptive spacing in yoked conditions was primarily due to adaptation to individual items and learners. Adaptive spacing based on ongoing assessments of learning strength for individual items and learners yields greater learning gains than fixed schedules, a finding that helps to understand the spacing effect theoretically and has direct applications for enhancing learning in many domains.

Latent fingerprint examination is a complex task that, despite advances in image processing, still fundamentally depends on the visual judgments of highly trained human examiners. Fingerprints collected from crime scenes typically contain less information than fingerprints collected under controlled conditions. Specifically, they are often noisy and distorted and may contain only a portion of the total fingerprint area. Expertise in fingerprint comparison, like other forms of perceptual expertise, such as face recognition or aircraft identification, depends on perceptual learning processes that lead to the discovery of features and relations that matter in comparing prints. Relatively little is known about the perceptual processes involved in making comparisons, and even less is known about what characteristics of fingerprint pairs make particular comparisons easy or difficult. We measured expert examiner performance and judgments of difficulty and confidence on a new fingerprint database. We developed a number of quantitative measures of image characteristics and used multiple regression techniques to discover objective predictors of error as well as perceived difficulty and confidence. A number of useful predictors emerged, and these included variables related to image quality metrics, such as intensity and contrast information, as well as measures of information quantity, such as the total fingerprint area. Also included were configural features that fingerprint experts have noted, such as the presence and clarity of global features and fingerprint ridges. Within the constraints of the overall low error rates of experts, a regression model incorporating the derived predictors demonstrated reasonable success in predicting objective difficulty for print pairs, as shown both in goodness of fit measures to the original data set and in a cross validation test. The results indicate the plausibility of using objective image metrics to predict expert performance and subjective assessment of difficulty in fingerprint comparisons.