Alexander A. Petrov

Task precision at transfer determines specificity of perceptual learning.

Perceptual learning, the improvement in performance with practice, reflects plasticity in the adult visual system. We challenge a standard claim that specificity of perceptual learning depends on task difficulty during training, instead showing that specificity, or conversely transfer, is primarily controlled by the precision demands (i.e., orientation difference) of the transfer task. Thus, for an orientation discrimination task, transfer of performance improvement is observed in low-precision transfer tasks, while specificity of performance improvement is observed in high-precision transfer tasks, regardless of the precision of initial training. The nature of specificity places important constraints on mechanisms of transfer in visual learning. These results contribute to understanding generalization of practiced improvements that may be key to the development of expertise and for applications in remediation.

The Dynamics of Scaling: A Memory-Based Anchor Model of Category Rating and Absolute Identification

A memory-based scaling model--ANCHOR--is proposed and tested. The perceived magnitude of the target stimulus is compared with a set of anchors in memory. Anchor selection is probabilistic and sensitive to similarity, base-level strength, and recency. The winning anchor provides a reference point near the target and thereby converts the global scaling problem into a local comparison. An explicit correction strategy determines the final response. Two incremental learning mechanisms update the locations and base-level activations of the anchors. This gives rise to sequential, context, transfer, practice, and other dynamic effects. The scale unfolds as an adaptive map. A hierarchy of models is tested on a battery of quantitative measures from 2 experiments in absolute identification and category rating.

A novel method for analyzing sequential eye movements reveals strategic influence on Raven's Advanced Progressive Matrices.

Eye movements are an important data source in vision science. However, the vast majority of eye movement studies ignore sequential information in the data and utilize only first-order statistics. Here, we present a novel application of a temporal-difference learning algorithm to construct a scanpath successor representation (SR; P. Dayan, 1993) that captures statistical regularities in temporally extended eye movement sequences. We demonstrate the effectiveness of the scanpath SR on eye movement data from participants solving items from Ravens Advanced Progressive Matrices Test. Analysis of the SRs revealed individual differences in scanning patterns captured by two principal components that predicted individual Raven scores much better than existing methods. These scanpath SR components were highly interpretable and provided new insight into the role of strategic processing on the Raven test. The success of the scanpath SR in terms of prediction and interpretability suggests that this method could prove useful in a much broader context.

Mapping and correcting the influence of gaze position on pupil size measurements

Pupil size is correlated with a wide variety of important cognitive variables and is increasingly being used by cognitive scientists. Pupil data can be recorded inexpensively and non-invasively by many commonly used video-based eye-tracking cameras. Despite the relative ease of data collection and increasing prevalence of pupil data in the cognitive literature, researchers often underestimate the methodological challenges associated with controlling for confounds that can result in misinterpretation of their data. One serious confound that is often not properly controlled is pupil foreshortening error (PFE)—the foreshortening of the pupil image as the eye rotates away from the camera. Here we systematically map PFE using an artificial eye model and then apply a geometric model correction. Three artificial eyes with different fixed pupil sizes were used to systematically measure changes in pupil size as a function of gaze position with a desktop EyeLink 1000 tracker. A grid-based map of pupil measurements was recorded with each artificial eye across three experimental layouts of the eye-tracking camera and display. Large, systematic deviations in pupil size were observed across all nine maps. The measured PFE was corrected by a geometric model that expressed the foreshortening of the pupil area as a function of the cosine of the angle between the eye-to-camera axis and the eye-to-stimulus axis. The model reduced the root mean squared error of pupil measurements by 82.5 % when the model parameters were pre-set to the physical layout dimensions, and by 97.5 % when they were optimized to fit the empirical error surface.

Dissociable perceptual-learning mechanisms revealed by diffusion-model analysis

Performance on perceptual tasks improves with practice. Most theories address only accuracy data and tacitly assume that perceptual learning is a monolithic phenomenon. The present study pioneers the use of response time distributions in perceptual learning research. The 27 observers practiced a visual motion-direction discrimination task with filtered-noise textures for four sessions with feedback. Session 5 tested whether the learning effects transferred to the orthogonal direction. The diffusion model (Ratcliff, Psychological Review, 85, 59–108, 1978) achieved good fits to the individual response time distributions from each session and identified two distinct learning mechanisms with markedly different specificities. A stimulus-specific increase in the drift-rate parameter indicated improved sensory input to the decision process, and a stimulus-general decrease in nondecision time variability suggested improved timing of the decision process onset relative to stimulus onset (which was preceded by a beep). A traditional d’ analysis would miss the latter effect, but the diffusion-model analysis identified it in the response time data.

The Leabra architecture: Specialization without modularity

The posterior cortex, hippocampus, and prefrontal cortex in the Leabra architecture are specialized in terms of various neural parameters, and thus are predilections for learning and processing, but domain-general in terms of cognitive functions such as face recognition. Also, these areas are not encapsulated and violate Fodorian criteria for modularity. Andersons terminology obscures these important points, but we applaud his overall message.

Symmetry-based methodology for decision-rule identification in same--different experiments.

The standard practice of reducing every same-different data set to two numbers (hits and false alarms) is wasteful, because the response pattern to all four stimulus pairs carries information about the decision rule adopted by the observer. We describe eight rules organized in three families: differencing, covert classification, and likelihood ratio. We prove that each family produces a characteristic pattern of (in)equalities among the response probabilities. We propose two simple qualitative tests. Is the performance on stimulus pairs AA and BB statistically indistinguishable? If not, differencing and likelihood-ratio strategies can be rejected. Is the performance on pairs AB and BA indistinguishable? If not, covert classification can be rejected. We present algorithms for fitting two covert-classification models and illustrate the new methodology in a perceptual learning experiment on visual motion-direction discrimination. The standard assumption of symmetric decision criteria was violated.

Relational priming plays a supporting but not leading role in adult analogy-making

Leech et al.s analysis adds to an emerging consensus of the role of priming in analogy-making. However, their model cannot scale up to adult-level performance because not all relations can be cast as functions. One-size-fits-all accounts cannot capture the richness of analogy. Proportional analogies and transitive inferences can be made by nonstructural mechanisms. Therefore, these tasks do not generalize to tasks that require structure mapping.

The visual identification of relational categories

An experiment was performed to investigate the ability of human observers to identify configural relations among three dots. Four stimulus categories were defined on the basis of whether or not the dots were arranged collinearly and whether or not the central dot was equally spaced relative to the two flanking dots. Observers were initially trained with feedback to identify these categories at a single orientation with a fixed uniform background, and then they were tested with variable orientations and backgrounds without feedback. The results revealed almost perfect generalization. We also simulated the same task using a recent feature hierarchy model (J. Mutch & D. G. Lowe, 2008) that is among the most successful for object recognition. This model performed well for fixed orientations and backgrounds, but it could not accurately identify these categories with variable orientations and backgrounds even when given training with those conditions. These findings suggest that feature hierarchy models represent the spatial relations within an image quite differently than human observers.

Category Rating Is Based on Prototypes and Not Instances: Evidence from Feedback-Dependent Context Effects

Context effects in category rating on a 7-point scale are shown to reverse direction depending on feedback. Context (skewed stimulus frequencies) was manipulated between and feedback within subjects in two experiments. The diverging predictions of prototype- and exemplar-based scaling theories were tested using two representative models: ANCHOR and INST. To gain coverage on one side of the continuum, a prototype-based category must lose on the opposite side. ANCHOR can exhibit both assimilative and compensatory context effects depending on feedback. INST always exhibits assimilative effects. The human data show a significant context-by-feedback interaction. The main context effect is assimilative in one data set and compensatory in the other. This pattern is consistent with ANCHOR but rules out INST, which fails to account for the compensatory effect and the interaction. This suggests that human category rating is based on unitary representations.