Thomas J. Palmeri

Rule-plus-exception model of classification learning

The authors propose a rule-plus-exception model (RULEX) of classification learning. According to RULEX, people learn to classify objects by forming simple logical rules and remembering occasional exceptions to those rules. Because the learning process in RULEX is stochastic, the model predicts that individual Ss will vary greatly in the particular rules that are formed and the exceptions that are stored. Averaged classification data are presumed to represent mixtures of these highly idiosyncratic rules and exceptions. RULEX accounts for numerous fundamental classification phenomena, including prototype and specific exemplar effects, sensitivity to correlational information, difficulty of learning linearly separable versus nonlinearly separable categories, selective attention effects, and difficulty of learning concepts with rules of differing complexity. RULEX also predicts distributions of generalization patterns observed at the individual subject level.

Inhibitory control in mind and brain: An interactive race model of countermanding saccades

The stop-signal task has been used to study normal cognitive control and clinical dysfunction. Its utility is derived from a race model that accounts for performance and provides an estimate of the time it takes to stop a movement. This model posits a race between go and stop processes with stochastically independent finish times. However, neurophysiological studies demonstrate that the neural correlates of the go and stop processes produce movements through a network of interacting neurons. The juxtaposition of the computational model with the neural data exposes a paradox-how can a network of interacting units produce behavior that appears to be the outcome of an independent race? The authors report how a simple, competitive network can solve this paradox and provide an account of what is measured by stop-signal reaction time.

Visual object understanding

Visual object understanding includes processes at the nexus of visual perception and visual cognition. A traditional approach separates questions that are more associated with perception — how are objects represented by high-level vision — from questions that are more associated with cognition — how are objects identified, categorized and remembered. However, to understand the bridge between perception and cognition, it is fruitful to abandon any sharp distinction between perceptual and cognitive aspects of visual object understanding. We provide a selective review of research from both the Object Recognition and Perceptual Categorization literatures, highlighting relevant behavioural, neuropsychological, neurophysiological and theoretical research into the representations and processes that underlie visual object understanding in humans and primates.

Neurally constrained modeling of perceptual decision making.

Stochastic accumulator models account for response time in perceptual decision-making tasks by assuming that perceptual evidence accumulates to a threshold. The present investigation mapped the firing rate of frontal eye field (FEF) visual neurons onto perceptual evidence and the firing rate of FEF movement neurons onto evidence accumulation to test alternative models of how evidence is combined in the accumulation process. The models were evaluated on their ability to predict both response time distributions and movement neuron activity observed in monkeys performing a visual search task. Models that assume gating of perceptual evidence to the accumulating units provide the best account of both behavioral and neural data. These results identify discrete stages of processing with anatomically distinct neural populations and rule out several alternative architectures. The results also illustrate the use of neurophysiological data as a model selection tool and establish a novel framework to bridge computational and neural levels of explanation.

Comparing modes of rule-based classification learning: A replication and extension of Shepard, Hovland, and Jenkins (1961)

We partially replicate and extend Shepard, Hovland, and Jenkinss (1961) classic study of task difficulty for learning six fundamental types of rule-based categorization problems. Our main results mirrored those of Shepard et al., with the ordering of task difficulty being the same as in the original study. A much richer data set was collected, however, which enabled the generation of block-by-block learning curves suitable for quantitative fitting. Four current computational models of classification learning were fitted to the learning data: ALCOVE (Kruschke, 1992), the rational model (Anderson, 1991), the configural-cue model (Gluck & Bower, 1988b), and an extended version of the configural-cue model with dimensionalized, adaptive learning rate mechanisms. Although all of the models captured important qualitative aspects of the learning data, ALCOVE provided the best overall quantitative fit. The results suggest the need to incorporate some form of selective attention to dimensions in category-learning models based on stimulus generalization and cue conditioning.

Exemplar similarity and the development of automaticity

Effects of exemplar similarity on the development of automaticity were investigated with a task in which participants judged the numerosity of random patterns of between 6 and 11 dots. After several days of training, response times were the same at all levels of numerosity, signaling the development of automaticity. In Experiment 1, response times to new patterns were a function of their similarity to old patterns. In Experiment 2, responses to patterns with high within-category similarity became automatized more quickly than responses to patterns with low within-category similarity. In Experiment 3, responses to patterns with high between-category similarity became automatized more slowly than responses to patterns with low between-category similarity. A new theory, the exemplar-based random walk (EBRW) model, was used to explain the results. Combining elements of G. D. Logans (1988) instance theory of automaticity and R. M. Nosofskys (1986) generalized context model of categorization, the theory embeds a dynamic similarity-based memory retrieval mechanism within a competitive random walk decision process.

The perceptual reality of synesthetic colors

Synesthesia is a remarkable, rare condition where an individual has multimodal perceptual experiences from a unimodal sensory event. We have studied such an individual, an adult male for whom achromatic words and alphanumeric characters are seen in vivid, reliable colors. We used a variety of perceptual tasks to document the perceptual reality of synesthetic colors and to begin to localize the stage of visual processing where this anomalous binding of externally specified form and internally generated color may take place. Synesthetic colors were elicited by forms defined solely by binocular cues or solely by motion cues, which implies a central locus of visual processing for synesthetic binding of form and color. Also included among our measurements was a difficult visual search task on which non-synesthetic subjects required an effortful search through the visual display. Our subject, in contrast to non-synesthetic subjects, accomplished the task with relative ease because the target of the search had a different synesthetic color from the distractors. Thus, synesthetic experiences appear to originate from a binding of color and form that takes place within central stages of visual processing.

Conditions for Facelike Expertise With Objects Becoming a Ziggerin Expert—but Which Type?

Compared with other objects, faces are processed more holistically and with a larger reliance on configural information. Such hallmarks efface processing can also be found for nonface objects as people develop expertise with them. Is this specifically a result of expertise individuating objects, or would any type of prolonged intensive experience with objects be sufficient? Two groups of participants were trained with artificial objects (Ziggerins). One group learned to rapidly individuate Ziggerins (i.e., subordinate-level training). The other group learned rapid, sequential categorizations at the basic level. Individuation experts showed a selective improvement at the subordinate level and an increase in holistic processing. Categorization experts improved only at the basic level, showing no changes in holistic processing. Attentive exposure to objects in a difficult training regimen is not sufficient to produce facelike expertise. Rather, qualitatively different types of expertise with objects of a given geometry can arise depending on the type of training.

Are there representational shifts during category learning

Early theories of categorization assumed that either rules, or prototypes, or exemplars were exclusively used to mentally represent categories of objects. More recently, hybrid theories of categorization have been proposed that variously combine these different forms of category representation. Our research addressed the question of whether there are representational shifts during category learning. We report a series of experiments that tracked how individual subjects generalized their acquired category knowledge to classifying new critical transfer items as a function of learning. Individual differences were observed in the generalization patterns exhibited by subjects, and those generalizations changed systematically with experience. Early in learning, subjects generalized on the basis of single diagnostic dimensions, consistent with the use of simple categorization rules. Later in learning, subjects generalized in a manner consistent with the use of similarity-based exemplar retrieval, attending to multiple stimulus dimensions. Theoretical modeling was used to formally corroborate these empirical observations by comparing fits of rule, prototype, and exemplar models to the observed categorization data. Although we provide strong evidence for shifts in the kind of information used to classify objects as a function of categorization experience, interpreting these results in terms of shifts in representational systems underlying perceptual categorization is a far thornier issue. We provide a discussion of the challenges of making claims about category representation, making reference to a wide body of literature suggesting different kinds of representational systems in perceptual categorization and related domains of human cognition.

Inverted faces are (eventually) processed holistically

Face inversion effects are used as evidence that faces are processed differently from objects. Nevertheless, there is debate about whether processing differences between upright and inverted faces are qualitative or quantitative. We present two experiments comparing holistic processing of upright and inverted faces within the composite task, which requires participants to match one half of a test face while ignoring irrelevant variation in the other half of the test face. Inversion reduced overall performance but led to the same qualitative pattern of results as observed for upright faces (Experiment 1). However, longer presentation times were required to observe holistic effects for inverted compared to upright faces (Experiment 2). These results suggest that both upright and inverted faces are processed holistically, but inversion reduces overall processing efficiency.