Modeling Response Times for Two-Choice Decisions

The diffusion model for two-choice real-time decisions is applied to four psychophysical tasks. The model reveals how stimulus information guides decisions and shows how the information is processed through time to yield sometimes correct and sometimes incorrect decisions. Rapid two-choice decisions yield multiple empirical measures: response times for correct and error responses, the probabilities of correct and error responses, and a variety of interactions between accuracy and response time that depend on instructions and task difficulty. The diffusion model can explain all these aspects of the data for the four experiments we present. The model correctly accounts for error response times, something previous models have failed to do. Variability within the decision process explains how errors are made, and variability across trials correctly predicts when errors are faster than correct responses and when they are slower.     

Modeling the Covariance Structure of Complex Datasets Using Cognitive Models: An Application to Individual Differences and the Heritability of Cognitive Ability

Understanding individual differences in cognitive performance is an important part of understanding how variations in underlying cognitive processes can result in variations in task performance. However, the exploration of individual differences in the components of the decision process—such as cognitive processing speed, response caution, and motor execution speed—in previous research has been limited. Here, we assess the heritability of the components of the decision process, with heritability having been a common aspect of individual differences research within other areas of cognition. Importantly, a limitation of previous work on cognitive heritability is the underlying assumption that variability in response times solely reflects variability in the speed of cognitive processing. This assumption has been problematic in other domains, due to the confounding effects of caution and motor execution speed on observed response times. We extend a cognitive model of decision‐making to account for relatedness structure in a twin study paradigm. This approach can separately quantify different contributions to the heritability of response time. Using data from the Human Connectome Project, we find strong evidence for the heritability of response caution, and more ambiguous evidence for the heritability of cognitive processing speed and motor execution speed. Our study suggests that the assumption made in previous studies—that the heritability of cognitive ability is based on cognitive processing speed—may be incorrect. More generally, our methodology provides a useful avenue for future research in complex data that aims to analyze cognitive traits across different sources of related data, whether the relation is between people, tasks, experimental phases, or methods of measurement.     

Emotion dysregulation and threat-related attention bias variability

Although theory suggests that a bias for attending to threat information (ABT) may be a biobehavioral process underlying the transdiagnostic vulnerability factor of emotion dysregulation, there is a paucity of empirical evidence showing direct associations between emotion dysregulation and ABT. The purpose of the present study was to examine the relation between ABT and emotion dysregulation. Participants (N?=?200) completed a battery of self-report questionnaires and a modified dot-probe task with both neutral and threat stimuli and four stimulus presentation durations. Task response times were used to examine traditionally calculated ABT scores, as well as attention bias variability (ABV). As predicted, those with greater emotion dysregulation exhibited greater ABV. Importantly, emotion dysregulation was not associated with response time variability on trials for which only neutral stimuli were presented, thus increasing confidence that emotion dysregulation-related ABV is specific to the presence of threat stimuli and not merely a function of general variability in response times. Results suggest that those with greater emotion dysregulation exhibit attentional dyscontrol in the presence of perceived threat that is characterized by dynamic shifts between vigilance and avoidance.     

The role of age and perceptual symbols in language comprehension

Older and younger participants read sentences about objects and were then shown a picture of an object that either matched or mismatched the implied shape of the object in the sentence. Participants' response times were recorded when they judged whether the object had been mentioned in the sentence. Responses were faster in the shape-matching condition for all participants, but the mismatch effect was stronger for older than for younger adults, even when the larger variability of the older group's response times was controlled for. These results suggest that older adults may construct stronger situation models than younger adults.     

The impact of inhibition capacities and age on number–space associations

Numerical and spatial representations are tightly linked, i.e., when doing a binary classification judgment on Arabic digits, participants are faster to respond with their left/right hand to small/large numbers, respectively (Spatial-Numerical Association of Response Codes, SNARC effect, Dehaene et al. in J Exp Psychol Gen 122:371–396, 1993). To understand the underlying mechanisms of the well-established SNARC effect, it seems essential to explore the considerable inter-individual variability characterizing it. The present study assesses the respective roles of inhibition, age, working memory (WM) and response speed. Whereas these non-numerical factors have been proposed as potentially important factors to explain individual differences in SNARC effects, none (except response speed) has so far been explored directly. Confirming our hypotheses, the results show that the SNARC effect was stronger in participants that had weaker inhibition abilities (as assessed by the Stroop task), were relatively older and had longer response times. Interestingly, whereas a significant part of the age influence was mediated by cognitive inhibition, age also directly impacted the SNARC effect. Similarly, cognitive inhibition abilities explained inter-individual variability in number–space associations over and above the factors age, WM capacity and response speed. Taken together our results provide new insights into the nature of number–space associations by describing how these are influenced by the non-numerical factors age and inhibition.     

Using finite mixture of GLMs to explore variability in children's flexibility in a task-switching paradigm

The present study illustrates the usefulness of finite mixture of generalized linear models (GLMs) to examine variability in cognitive strategies during childhood. More precisely, it addresses this variability in set-shifting situations where task-goal updating is endogenously driven. In a task-switching paradigm 5–6-year-olds had to switch between color- and shape-matching rules as a function of a predetermined, predictable task sequence. A finite mixture of GLMs was fitted to explore individual differences in performance. The statistical model revealed five response profiles, defined by accuracy and response times. These response profiles likely correspond to different cognitive strategies with varying efficiency and differential relations to working memory capacity (assessed by backward digit span). These results illustrate the heuristic value of statistical modeling to reveal the behavioral and cognitive variability in the temporal dynamics of children's cognitive functioning.     

Improved Test–retest Reliability Criteria as Applied to Probability Assessments

Test–retest reliability is a common indicator of response consistency. It is argued that using regression coefficients for detecting systematic response error is less appropriate than testing for shifts in the mean (median) and variance. This procedure is exemplified using probability response data. For this data, shifts in centrality were found to be about 2.5 times more likely than shifts in variability. The shifts in centrality did not favor any particular direction; however, variability tended to decrease over time in early sessions.     

An electrophysiological model of human visual reaction time

Human visual reaction times were fractionated into component latencies measuring visual reception time, opto-motor integration time, central motor outflow time, and peripheral motor time on the basis of evoked cortical activity recorded from the intact scalp and the occurrence of the response electromyogram. Normative data are presented for a right-foot dorsiflexion task studied on 18 male subjects, together with an analysis of inter- and intra-subject variability in response timing. Faster reactors were found to display briefer opto-motor integration times and motor times, while an individual's faster responses were characterized by shorter motor outflow times and motor time. These results are interpreted in terms of varying physiological mechanisms.     

Work flow policy and within-worker and between-workers variability in performance

Work flow policies are shown to induce a change in average between-workers variability (worker heterogeneity) and within-worker variability in performance times. In a laboratory experiment, the authors measured the levels of worker heterogeneity and within-worker variability under an individual performance condition, a work sharing condition, and a fixed assignment condition. The work sharing policy increased the levels of worker heterogeneity and worker variability, whereas the fixed assignment policy decreased them. These effects, along with work flow policy main effects on mean performance times and variability are examined. This article represents an initial step in understanding effects that may be important in the selection of an operating policy, the ignorance of which may lead to costly misestimates of performance.     

Expectancy in humans in multisecond peak-interval timing with gaps

In two experiments, the peak-interval procedure was used with humans to test effects related to gaps in multisecond timing. In Experiment 1, peak times of response distributions were shorter when the gap occurred later during the encoding of the criterion time to be reproduced, suggesting that gap expectancy shortened perceived durations. Peak times were also positively related to objective target durations. Spreads of response distributions were generally related to estimated durations. In Experiment 2, peak times were shortest when gaps were expected but did not occur, confirming that the shortening effect of gap expectancy is independent of the gap occurrence. High positive start-stop correlations and moderate positive peak-time-spread correlations showed strong memory variability, whereas low and negative start-spread correlations suggest small response-threshold variability. Correlations seemed not to be influenced by expectancy. Overall, the peak-interval procedure with gaps provided relevant information on similarities and differences in timing in humans and other animals.     

How social network heterogeneity facilitates lexical access and lexical prediction

People learn language from their social environment. As individuals differ in their social networks, they might be exposed to input with different lexical distributions, and these might influence their linguistic representations and lexical choices. In this article we test the relation between linguistic performance and 3 social network properties that should influence input variability, namely, network size, network heterogeneity, and network density. In particular, we examine how these social network properties influence lexical prediction, lexical access, and lexical use. To do so, in Study 1, participants predicted how people of different ages would name pictures, and in Study 2 participants named the pictures themselves. In both studies, we examined how participants’ social network properties related to their performance. In Study 3, we ran simulations on norms we collected to see how age variability in one’s network influences the distribution of different names in the input. In all studies, network age heterogeneity influenced performance leading to better prediction, faster response times for difficult-to-name items, and less entropy in input distribution. These results suggest that individual differences in social network properties can influence linguistic behavior. Specifically, they show that having a more heterogeneous network is associated with better performance. These results also show that the same factors influence lexical prediction and lexical production, suggesting the two might be related.     

Subjective reports of stimulus,response, and decision times in speeded tasks: How accurate are decision time reports?

Four experiments examined how accurately participants can report the times of their own decisions. Within an auditory reaction time (RT) task, participants reported the time at which (a) the tone was presented, (b) they decided on the response, or (c) the response key was pressed. Decision time reports were checked for plausibility against the actual RTs, and we compared the effects of experimental manipulations on these two measures to see whether the reported decision times showed appropriate effects. In addition, we estimated the amount of error associated with individual decision time reports by checking how often participants’ decision time reports were implausibly early (i.e., before stimulus onset) or late (i.e., after response), and by using several quantitative models. Overall, the results suggest that decision time reports are not very accurate but they may be usable for some purposes.     

Constraints have different concurrent effects and aftereffects on variability

Three experiments showed that constraints imposed early in learning have different effects on variability when they are in effect and after they are removed. Task constraints, which determine how something can be done, limited the number of possible responses in a computer game. Variability constraints, which specify how differently something must be done, required that each response differ from some number of prior responses. Less restrictive constraints (Experiments 1 and 2) produced higher variability during the constraints. More restrictive constraints (Experiments 2 and 3) led to higher variability after the constraints were relaxed. The authors discuss how these differences reflect strategies acquired during the constraints (default rules) and modified in closely related ways (exception rules) afterward.     

The operant conditioning of response variability: Free-operant versus discrete-response procedures

The operant conditioning of response variability under free-operant and discrete-response procedures was investigated. Two pigeons received food only if their pattern of four pecks on two response keys differed from the patterns emitted on the two immediately preceding trials. Under the free-operant procedure, the keys remained illuminated and operative throughout each trial. There was little variability in the response patterns that resulted, and the pigeons received fewer than one third of the available reinforcers. Under the discrete-response procedure, a brief timeout period followed each response. Variability increased under this procedure, and the pigeons obtained three fourths of the available reinforcers. Previous successes and failures to produce response variability may have been due to the use or failure to use, respectively, a discrete-response procedure. Respondent effects inherent in the free-operant procedure may encourage the development of response stereotypy and, in turn, prevent the development of response variability.     

The influence of age on consistent and varied semantic-category search performance

Three experiments examined age-related consistent mapping (CM) and varied mapping (VM) practice effects. In separate experiments, young (age, 19-22), middle-aged (37-50), and older (64-88) adults' performance was examined using semantic-category and letter-based search paradigms. After extensive practice, major age differences occurred in CM search. Young and middle-aged subjects showed near-zero comparison slopes, large reductions in mean reaction times, and substantial reductions in response variability. Although older adults' reaction time decreased with CM practice, the reduction in reaction time, comparison slope, and response variability was small compared with the other groups. In VM search, older subjects were slower than the other age groups, but all groups exhibited similar linear set-size functions, search termination, and comparison-load effects. We concluded that age-associated declines in cognitive performance may be due in large part to the degree with which older adults can acquire or use automatic processes to perform tasks.     

An optimal period for setting sustained variability levels

In two experiments, we investigated how explicit reinforcement of highly variable behavior at different points in training affected performance after the requirement was eliminated. Two versions of a computer game, differing in the number of possible solution paths, were used. In each, an optimal period of training for producing sustained high variability was found. Exposure to a high lag requirement shortly after acquisition sustained variability. Rewarding variability at other times did not have a sustained effect. The implications for learning and problem solving are discussed.     

Operant variability: evidence,functions, and theory

Although responses are sometimes easy to predict, at other times responding seems highly variable, unpredictable, or even random. The inability to predict is generally attributed to ignorance of controlling variables, but this article is a review of research showing that the highest levels of behavioral variability may result from identifiable reinforcers contingent on such variability. That is, variability is an operant. Discriminative stimuli and reinforcers control it, resulting in low or high variability, depending on the contingencies. Schedule-of-reinforcement effects are orderly, and choosing to vary or repeat is lawfully governed by relative reinforcement frequencies. The operant nature of variability has important implications. For example, learning, exploring, creating, and problem solving may partly depend on it. Abnormal levels of variability, including those found in psychopathologies such as autism, depression, and attention deficit hyperactivity disorder, may be modified through reinforcement. Operant variability may also help to explain some of the unique attributes of voluntary action.     

Optimal Predictions in Everyday Cognition: The Wisdom of Individuals or Crowds?

Griffiths and Tenenbaum (2006) asked individuals to make predictions about the duration or extent of everyday events (e.g., cake baking times), and reported that predictions were optimal, employing Bayesian inference based on veridical prior distributions. Although the predictions conformed strikingly to statistics of the world, they reflect averages over many individuals. On the conjecture that the accuracy of the group response is chiefly a consequence of aggregating across individuals, we constructed simple, heuristic approximations to the Bayesian model premised on the hypothesis that individuals have access merely to a sample of k instances drawn from the relevant distribution. The accuracy of the group response reported by Griffiths and Tenenbaum could be accounted for by supposing that individuals each utilize only two instances. Moreover, the variability of the group data is more consistent with this small-sample hypothesis than with the hypothesis that people utilize veridical or nearly veridical representations of the underlying prior distributions. Our analyses lead to a qualitatively different view of how individuals reason from past experience than the view espoused by Griffiths and Tenenbaum.     

Accounting for Parcel-Allocation Variability in Practice: Combining Sources of Uncertainty and Choosing the Number of Allocations

Item parceling remains widely used under conditions that can lead to parcel-allocation variability in results. Hence, researchers may be interested in quantifying and accounting for parcel-allocation variability within sample. To do so in practice, three key issues need to be addressed. First, how can we combine sources of uncertainty arising from sampling variability and parcel-allocation variability when drawing inferences about parameters in structural equation models? Second, on what basis can we choose the number of repeated item-to-parcel allocations within sample? Third, how can we diagnose and report proportions of total variability per estimate arising due to parcel-allocation variability versus sampling variability? This article addresses these three methodological issues. Developments are illustrated using simulated and empirical examples, and software for implementing them is provided.