Complex decision rules in categorization: Contrasting novice and experienced performance.

The ability of novice and experienced Ss to learn complex decision rules was tested with 3 categorization tasks. Each task involved 2 categories with exemplars that were normally distributed on 2 stimulus dimensions. Three separate sets of stimuli were used, and in each task the decision rule that maximized categorization accuracy was a highly nonlinear function of the stimulus dimension values. In the 3 tasks, all experienced Ss used highly nonlinear decision rules. Quadratic rules were supported over bilinear rules, and in many cases Ss used nearly optimal decision rules. These findings did not depend on whether the stimulus components were integral or separable. Novice Ss also did not use simple linear rules. A model that assumed Ss tried a succession of different linear rules was also rejected. Instead, novices appeared to use quadratic rules, although less consistently than experienced Ss. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Learning to choose: Cognitive aging and strategy selection learning in decision making.

Decision makers often have to learn from experience. In these situations, people must use the available feedback to select the appropriate decision strategy. How does the ability to select decision strategies on the basis of experience change with age? We examined younger and older adults' strategy selection learning in a probabilistic inference task using a computational model of strategy selection learning. Older adults showed poorer decision performance compared with younger adults. In particular, older adults performed poorly in an environment favoring the use of a more cognitively demanding strategy. The results suggest that the impact of cognitive aging on strategy selection learning depends on the structure of the decision environment. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Adaptive strategy selection in decision making.

The role of effort and accuracy in the adaptive use of decision processes is examined. A computer simulation using the concept of elementary information processes identified heuristic choice strategies that approximate the accuracy of normative procedures while saving substantial effort. However, no single heuristic did well across all task and context conditions. Of particular interest was the finding that under time constraints, several heuristics were more accurate than a truncated normative procedure. Using a process-tracing technique that monitors information acquisition behaviors, two experiments tested how closely the efficient processing patterns for a given decision problem identified by the simulation correspond to the actual processing behavior exhibited by subjects. People appear highly adaptive in responding to changes in the structure of the available alternatives and to the presence of time pressure. In general, actual behavior corresponded to the general patterns of efficient processing identified by the simulation. Finally, learning of effort and accuracy trade-offs are discussed. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

A layered network model of associative learning: Learning to learn and configuration.

Networks composed of layers of adaptive units provide a rigorous explanation for associative learning phenomena that otherwise have been relatively intractable, particularly learning to learn, spontaneous configuration, and negative patterning (the exclusive-OR problem). Layered network models can also reconcile these phenomena with better-understood phenomena, for example, stimulus summation, blocking, and conditioned inhibition. This article presents simulations based on a network of three adaptive units, each of which operates according to an associative competition rule, also known as the delta rule. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Group negotiation: Effects of decision rule, agenda, and aspiration.

In this study we characterized small group negotiation as a mixed motive task that involved both cooperation and competition. We examined the impact of two group decision-making processes (decision rule and agenda) and one cognitive-motivational frame (aspiration level) on the quality of negotiated outcomes in small groups. Negotiation groups that used a unanimous decision rule were more likely to integrate their interests to achieve higher group outcomes than were groups that used a majority rule. Negotiation groups that followed an explicit agenda and used a majority decision rule distributed resources more unequally, and were more likely to form coalitions against a remaining party than were groups with no agenda/majority rule, explicit agenda/unanimity rule, and no agenda/unanimity rule. There was no support for the hypotheses that group members who held high aspirations and followed a majority decision rule would distribute resources more unequally than would groups with high aspirations/unanimity rule, low aspirations/majority rule, and low aspirations/unanimity rule; that adherence to explicit agendas would lead to lower group profits; and that the absence of high aspirations would lead to lower group profit. We discuss the results in terms of a mixed motive analysis of group decision making. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Application of a computational decision model to examine acute drug effects on human risk taking.

In 3 previous experiments, high doses of alcohol, marijuana, and alprazolam acutely increased risky decision making by adult humans in a 2-choice (risky vs. nonrisky) laboratory task. In this study, a computational modeling analysis known as the expectancy valence model (J. R. Busemeyer & J. C. Stout, 2002) was applied to individual-participant data from these studies, for the highest administered dose of all 3 drugs and corresponding placebo doses, to determine changes in decision-making processes that may be uniquely engendered by each drug. The model includes 3 parameters: responsiveness to rewards and losses (valence or motivation); the rate of updating expectancies about the value of risky alternatives (learning/memory); and the consistency with which trial-by-trial choices match expected outcomes (sensitivity). Parameter estimates revealed 3 key outcomes: Alcohol increased responsiveness to risky rewards and decreased responsiveness to risky losses (motivation) but did not alter expectancy updating (learning/memory); both marijuana and alprazolam produced increases in risk taking that were related to learning/memory but not motivation; and alcohol and marijuana (but not alprazolam) produced more random response patterns that were less consistently related to expected outcomes on the 2 choices. No significant main effects of gender or dose by gender interactions were obtained, but 2 dose by gender interactions approached significance. These outcomes underscore the utility of using a computational modeling approach to deconstruct decision-making processes and thus better understand drug effects on risky decision making in humans. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Neural dynamics of decision making under risk: Affective balance and cognitive-emotional interactions.

A real-time neural network model, called affective balance theory, is developed to explain many properties of decision making under risk that heretofore have been analyzed using formal algebraic models, notably prospect theory. The model describes cognitive-emotional interactions that are designed to ensure adaptive responses to environmental demands but whose emergent properties nonetheless can lead to paradoxical and even irrational decisions in risky environments. Emotional processing in the model is carried out by an opponent processing network called a gated dipole. Learning enables cognitive representations to generate affective reactions of the dipole. Habituating chemical transmitters within a gated dipole determine an affective adaptation level, or context, against which later events are evaluated. Neutral events can become affectively charged either through direct activations or antagonistic rebounds within a previously habituated dipole. The theory describes the affective consequences of strategies in which an individual compares pairs of events or statements that are not necessarily explicitly grouped within the stimuli. The same preference orders may sometimes, but not always, emerge from different sequences of pair-wise alternatives. The role of short-term memory updating and expectancy-modulated matching processes in regulating affective reactions is described. The formal axioms of prospect theory are dynamically explicated through this analysis. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

"Two-stage dynamic signal detection: A theory of choice, decision time, and confidence": Erratum.

Reports an error in "Two-stage dynamic signal detection: A theory of choice, decision time, and confidence" by Timothy J. Pleskac and Jerome R. Busemeyer (Psychological Review, 2010[Jul], Vol 117[3], 864-901). The name of the philosopher Charles Peirce was misspelled throughout as Charles Pierce. (The following abstract of the original article appeared in record 2010-14834-006.) The 3 most often-used performance measures in the cognitive and decision sciences are choice, response or decision time, and confidence. We develop a random walk/diffusion theory—2-stage dynamic signal detection (2DSD) theory—that accounts for all 3 measures using a common underlying process. The model uses a drift diffusion process to account for choice and decision time. To estimate confidence, we assume that evidence continues to accumulate after the choice. Judges then interrupt the process to categorize the accumulated evidence into a confidence rating. The model explains all known interrelationships between the 3 indices of performance. Furthermore, the model also accounts for the distributions of each variable in both a perceptual and general knowledge task. The dynamic nature of the model also reveals the moderating effects of time pressure on the accuracy of choice and confidence. Finally, the model specifies the optimal solution for giving the fastest choice and confidence rating for a given level of choice and confidence accuracy. Judges are found to act in a manner consistent with the optimal solution when making confidence judgments. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Generalization hypothesis of abstract-concept learning: Learning strategies and related issues in Macaca mulatta, Cebus apella, and Columba livia.

The generalization hypothesis of abstract-concept learning was tested with a meta-analysis of rhesus monkeys (Macaca mulatta), capuchin monkeys (Cebus apella), and pigeons (Columba livia) learning a same/different (S/D) task with expanding training sets. The generalization hypothesis states that as the number of training items increases, generalization from the training pairs will increase and could explain the subjects' accurate novel-stimulus transfer. By contrast, concept learning is learning the relationship between each pair of items; with more training items subjects learn more exemplars of the rule and transfer better. Having to learn the stimulus pairs (the generalization hypothesis) would require more training as the set size increases, whereas learning the concept might require less training because subjects would be learning an abstract rule. The results strongly support concept or rule learning despite severely relaxing the generalization-hypothesis parameters. Thus, generalization was not a factor in the transfer from these experiments, adding to the evidence that these subjects were learning the S/D abstract concept. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Perspectives of probabilistic inferences: Reinforcement learning and an adaptive network compared.

The assumption that people possess a strategy repertoire for inferences has been raised repeatedly. The strategy selection learning theory specifies how people select strategies from this repertoire. The theory assumes that individuals select strategies proportional to their subjective expectations of how well the strategies solve particular problems; such expectations are assumed to be updated by reinforcement learning. The theory is compared with an adaptive network model that assumes people make inferences by integrating information according to a connectionist network. The network's weights are modified by error correction learning. The theories were tested against each other in 2 experimental studies. Study 1 showed that people substantially improved their inferences through feedback, which was appropriately predicted by the strategy selection learning theory. Study 2 examined a dynamic environment in which the strategies' performances changed. In this situation a quick adaptation to the new situation was not observed; rather, individuals got stuck on the strategy they had successfully applied previously. This "inertia effect" was most strongly predicted by the strategy selection learning theory. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Dynamical evolutionary psychology: Individual decision rules and emergent social norms.

A new theory integrating evolutionary and dynamical approaches is proposed. Following evolutionary models, psychological mechanisms are conceived as conditional decision rules designed to address fundamental problems confronted by human ancestors, with qualitatively different decision rules serving different problem domains and individual differences in decision rules as a function of adaptive and random variation. Following dynamical models, decision mechanisms within individuals are assumed to unfold in dynamic interplay with decision mechanisms of others in social networks. Decision mechanisms in different domains have different dynamic outcomes and lead to different sociospatial geometries. Three series of simulations examining trade-offs in cooperation and mating decisions illustrate how individual decision mechanisms and group dynamics mutually constrain one another, and offer insights about gene-culture interactions. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Signal-detection analysis of group decision making.

How effectively can groups of people make yes-or-no decisions? To answer this question, we used signal-detection theory to model the behavior of groups of human participants in a visual detection task. The detection model specifies how performance depends on the group's size, the competence of the members, the correlation among members' judgments, the constraints on member interaction, and the group's decision rule. The model also allows specification of performance efficiency, which is a measure of how closely a group's performance matches the statistically optimal group. The performance of our groups was consistent with the theoretical predictions, but efficiency decreased as group size increased. This result was attributable to a decrease in the effort that members gave to their individual tasks rather than an inefficiency in combining the information in the members' judgments. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Effect of generic advance instructions on learning a classification task.

Investigated whether giving students generic advance instructions about how to learn a classification task would be effective in facilitating learning of a specific classification task. The effect of instructions was compared with the effects of adjunct postquestions, simply reading the materials (control group), and both instructions and adjunct postquestions. The experimental task involved learning to classify instances of several different categories. Ss were 72 US Navy enlisted personnel. It was hypothesized that advance instructions to Ss regarding how and what to learn would facilitate learning of classification material. Results show that the instruction group and adjunct questions group did equally well on classifying new and old instances of US Navy call signs. All groups did better than control groups. Results suggest that students can learn a general strategy for processing classification material and that providing students with generic information about how and what to learn is an effective instructional strategy. A text of the instructions used to teach Ss how to learn a classification task is appended. (30 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Effect of temporal position, proportional variance, and proportional duration on decision weights in temporal pattern discrimination

Two experiments investigated how listeners allocate their attention to different segments of a temporal pattern. The experiments allowed a direct test of the predictions of the Proportion of Total Duration (PTD) rule and the Component Relative Entropy (CoRE) model. Listeners had to decide whether two sequences of nine tones had the same or different temporal patterns (tone duration = 25 ms, tone frequency = 1000 Hz). A sequence's temporal pattern was determined by the time intervals between each tone's offset and the next tone's onset. On same trials, the time intervals at corresponding temporal positions in the two sequences were identical. On different trials, the corresponding time intervals were randomly varied. Listener attention to different temporal positions within a sequence was assessed by calculating the decision weights at each position. The results supported the CoRE model and were inconsistent with the PTD rule. Manipulating the mean of the time intervals within the sequence had no consistent effect on the pattern of weights (or on overall performance), indicating that listener attention was not affected by either the proportion of total duration or the perceptual salience of a longer or shorter time interval. However, manipulating the variance of the time intervals had a significant effect: the highest weight was given to the highest variance segment. This weighting strategy leads to better performance because higher variance segments are more diagnostic of whether the sequences are the same or different.     

Base-rate effects in category learning: A comparison of parallel network and memory storage-retrieval models.

Exemplar-memory and adaptive network models were compared in application to category learning data, with special attention to base rate effects on learning and transfer performance. Subjects classified symptom charts of hypothetical patients into disease categories, with informative feedback on learning trials and with the feedback either given or withheld on test trials that followed each fourth of the learning series. The network model proved notably accurate and uniformly superior to the exemplar model in accounting for the detailed course of learning; both the parallel, interactive aspect of the network model and its particular learning algorithm contribute to this superiority. During learning, subjects' performance reflected both category base rates and feature (symptom) probabilities in a nearly optimal manner, a result predicted by both models, though more accurately by the network model. However, under some test conditions, the data showed substantial base-rate neglect, in agreement with M. A. Gluck and G. H. Bower (see record 1989-00340-001). (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Learning processes affecting human decision making: An assessment of reinforcer-selective Pavlovian-to-instrumental transfer following reinforcer devaluation.

In reinforcer-selective transfer, Pavlovian stimuli that are predictive of specific outcomes bias performance toward responses associated with those outcomes. Although this phenomenon has been extensively examined in rodents, recent assessments have extended to humans. Using a stock market paradigm adults were trained to associate particular symbols and responses with particular currencies. During the first test, individuals showed a preference for responding on actions associated with the same outcome as that predicted by the presented stimulus (i.e., a reinforcer-selective transfer effect). In the second test of the experiment, one of the currencies was devalued. We found it notable that this served to reduce responses to those stimuli associated with the devalued currency. This finding is in contrast to that typically observed in rodent studies, and suggests that participants in this task represented the sensory features that differentiate the reinforcers and their value during reinforcer-selective transfer. These results are discussed in terms of implications for understanding associative learning processes in humans and the ability of reward-paired cues to direct adaptive and maladaptive behavior. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

What is learned in sequential learning? An associative model of reward magnitude serial-pattern learning.

A computational model of sequence learning is described that is based on pairwise associations and generalization. Simulations by the model predicted that rats should learn a long monotonic pattern of food quantities better than a nonmonotonic pattern, as predicted by rule-learning theory, and that they should learn a short nonmonotonic pattern with highly discriminable elements better than 1 with less discriminable elements, as predicted by interitem association theory. In 2 other studies, the model also simulated behavioral "rule generalization", "extrapolation", and associative transfer data motivated by both rule-learning and associative perspectives. Although these simulations do not rule out the possibility that rats can use rule induction to learn serial patterns, they show that a simple associative model can account for the classical behavioral studies implicating rule learning in reward magnitude serial-pattern learning. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Relationship of actin, microtubules, and crosswall synthesis during septation in Aspergillus nidulans

A significant limitation of neural networks is that the representations they learn are usually incomprehensible to humans. We have developed an algorithm, called TREPAN, for extracting comprehensible, symbolic representations from trained neural networks. Given a trained network, TREPAN produces a decision tree that approximates the concept represented by the network. In this article, we discuss the application of TREPAN to a neural network trained on a noisy time series task: predicting the Dollar-Mark exchange rate. We present experiments that show that TREPAN is able to extract a decision tree from this network that equals the network in terms of predictive accuracy, yet provides a comprehensible concept representation. Moreover, our experiments indicate that decision trees induced directly from the training data using conventional algorithms do not match the accuracy nor the comprehensibility of the tree extracted by TREPAN.     

Feedback Specificity, Learning Opportunities, and Learning.

Although increasing feedback specificity is generally beneficial for immediate performance, it can undermine certain aspects of the learning needed for later, more independent performance. The results of the present transfer experiment demonstrate that the effects of increasing feedback specificity on learning depended on what was to be learned, and these effects were partially mediated through the opportunities to learn how to respond to different task conditions during practice. More specific feedback was beneficial for learning how to respond to good performance and detrimental for learning how to respond to poor performance. The former relationship was partially mediated by feedback specificity's effect on learning opportunities during practice. The results have implications for designing feedback interventions and training to maximize the learning of various aspects of a task. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Executive functions, information sampling, and decision making in narcolepsy with cataplexy.

Objective: Narcolepsy with cataplexy (NC) affects neurotransmitter systems regulating emotions and cognitive functions. This study aimed to assess executive functions, information sampling, reward processing, and decision making in NC. Method: Twenty-one NC patients and 58 healthy participants performed an extensive neuropsychological test battery. Results: NC patients scored as controls in executive function tasks assessing set shifting, reversal learning, working memory, and planning. Group differences appeared in a task measuring information sampling and reward sensitivity. NC patients gathered less information, tolerated a higher level of uncertainty, and were less influenced by reward contingencies than controls. NC patients also showed reduced learning in decision making and had significantly lower scores than controls in the fifth block of the IOWA gambling task. No correlations were found with measures of sleepiness. Conclusions: NC patients may achieve high performance in several neuropsychological domains, including executive functions. Specific differences between NC patients and controls highlight the importance of the hypocretin system in reward processing and decision making and are in line with previous neuroimaging and neurophysiological studies. (PsycINFO Database Record (c) 2011 APA, all rights reserved)