The power law repealed: the case for an exponential law of practice

How changes in white matter might underlie improved reaction time due to practice

Why does training on a task reduce the reaction time for performing it? New research points to changes in white matter pathways as one likely mechanism. These pathways connect remote brain areas involved in performing the task. Genetic variations may be involved in individual differences in the extent of this improvement. If white matter change is involved in improved reaction time with training, it may point the way toward understanding where and how generalization occurs. We examine the hypothesis that brain pathways shared by different tasks may result in improved performance of cognitive tasks remote from the training.

Network Profiles of the Dorsal Anterior Cingulate and Dorsal Prefrontal Cortex in Schizophrenia During Hippocampal-Based Associative Memory

Schizophrenia is a disorder characterized by brain network dysfunction, particularly during behavioral tasks that depend on frontal and hippocampal mechanisms. Here, we investigated network profiles of the regions of the frontal cortex during memory encoding and retrieval, phases of processing essential to associative memory. Schizophrenia patients (n = 12) and healthy control (HC) subjects (n = 10) participated in an established object-location associative memory paradigm that drives frontal-hippocampal interactions. Network profiles were modeled of both the dorsal prefrontal (dPFC) and the dorsal anterior cingulate cortex (dACC) as seeds using psychophysiological interaction analyses, a robust framework for investigating seed-based connectivity in specific task contexts. The choice of seeds was motivated by previous evidence of involvement of these regions during associative memory. Differences between patients and controls were evaluated using second-level analyses of variance (ANOVA) with seed (dPFC vs. dACC), group (patients vs. controls), and memory process (encoding and retrieval) as factors. Patients showed a pattern of exaggerated modulation by each of the dACC and the dPFC during memory encoding and retrieval. Furthermore, group by memory process interactions were observed within regions of the hippocampus. In schizophrenia patients, relatively diminished modulation during encoding was associated with increased modulation during retrieval. These results suggest a pattern of complex dysfunctional network signatures of critical forebrain regions in schizophrenia. Evidence of dysfunctional frontal-medial temporal lobe network signatures in schizophrenia is consistent with the illness' characterization as a disconnection syndrome.

Neuropsychological Practice Effects in the Context of Cognitive Decline: Contributions from Learning and Task Novelty

Although cognitive decline is typically associated with decreasing practice effects (PEs) (presumably due to declining memory), some studies show increased PEs with declines in cognition. One explanation for these inconsistencies is that PEs reflect not only memory, but also rebounds from adapting to task novelty (i.e., novelty effect), leading to increased PEs. We examined a theoretical model of relationships among novelty effects, memory, cognitive decline, and within-session PEs. Sixty-six older adults ranging from normal to severely impaired completed measures of memory, novelty effects, and two trials each of Wechsler Adult Intelligence Scale, 4 th Edition Symbol Search and Coding. Interrelationships among variables were examined using regression analyses. PEs for Symbol Search and Coding (a) were related to different proposed PE components (i.e., memory and novelty effects), such that novelty effect predicted Symbol Search PE (R2 =.239, p<.001) and memory predicted Coding PE (R2 =.089, p=.015), and (b) showed different patterns across stages of cognitive decline, such that the greatest cognitive decline was associated with smallest Coding PE (R2 =.125, p=.004), whereas intermediate cognitive decline was associated with the greatest Symbol Search PE (R2 =.097, p=.040). The relationship between cognitive decline and PE for Symbol Search was partially mediated by novelty effect among older adults with abnormal cognitive decline (model R2 =.286, p<.001). These findings (a) suggest that PE is not a unitary construct, (b) offer an explanation for contradictory findings in the literature, and (c) highlight the need for a better understanding of component processes of PE across different neuropsychological measures.

Assistive peripheral phosphene arrays deliver advantages in obstacle avoidance in simulated end-stage retinitis pigmentosa: a virtual-reality study

OBJECTIVE

The prospective efficacy of peripheral retinal prostheses for guiding orientation and mobility in the absence of residual vision, as compared to an implant for the central visual field (VF), was evaluated using simulated prosthetic vision (SPV).

APPROACH

Sighted volunteers wearing a head-mounted display performed an obstacle circumvention task under SPV. Mobility and orientation performance with three layouts of prosthetic vision were compared: peripheral prosthetic vision of higher visual acuity (VA) but limited VF, of wider VF but limited VA, as well as centrally restricted prosthetic vision. Learning curves using these layouts were compared fitting an exponential model to the mobility and orientation measures.

MAIN RESULTS

Using peripheral layouts, performance was superior to the central layout. Walking speed with both higher-acuity and wider-angle layouts was 5.6% higher, and mobility errors reduced by 46.4% and 48.6%, respectively, as compared to the central layout. The wider-angle layout yielded the least number of collisions, 63% less than the higher-acuity and 73% less than the central layout. Using peripheral layouts, the number of visual-scanning related head movements was 54.3% (higher-acuity) and 60.7% (wider-angle) lower, as compared to the central layout, and the ratio of time standing versus time walking was 51.9% and 61.5% lower, respectively. Learning curves did not differ between layouts, except for time standing versus time walking, where both peripheral layouts achieved significantly lower asymptotic values compared to the central layout.

SIGNIFICANCE

Beyond complementing residual vision for an improved performance, peripheral prosthetic vision can effectively guide mobility in the later stages of retinitis pigmentosa (RP) without residual vision. Further, the temporal dynamics of learning peripheral and central prosthetic vision are similar. Therefore, development of a peripheral retinal prosthesis and early implantation to alleviate VF constriction in RP should be considered to extend the target group and the time of benefit for potential retinal prosthesis implantees.

Jump, Hop, or Skip: Modeling Practice Effects in Studies of Determinants of Cognitive Change in Older Adults

Improvements in cognitive test scores upon repeated assessment due to practice effects (PEs) are well documented, but there is no empirical evidence on whether alternative specifications of PEs result in different estimated associations between exposure and rate of cognitive change. If alternative PE specifications produce different estimates of association between an exposure and rate of cognitive change, this would be a challenge for nearly all longitudinal research on determinants of cognitive aging. Using data from 3 cohort studies-the Three-City Study-Dijon (Dijon, France, 1999-2010), the Normative Aging Study (Greater Boston, Massachusetts, 1993-2007), and the Washington Heights-Inwood Community Aging Project (New York, New York, 1999-2012)-for 2 exposures (diabetes and depression) and 3 cognitive outcomes, we compared results from longitudinal models using alternative PE specifications: no PEs; use of an indicator for the first cognitive visit; number of prior testing occasions; and square root of the number of prior testing occasions. Alternative specifications led to large differences in the estimated rates of cognitive change but minimal differences in estimated associations of exposure with cognitive level or change. Based on model fit, using an indicator for the first visit was often (but not always) the preferred model. PE specification can lead to substantial differences in estimated rates of cognitive change, but in these diverse examples and study samples it did not substantively affect estimated associations of risk factors with change.

On the time course of short-term forgetting: a human experimental model for the sense of balance

The primary aim of this study was to establish whether the decline of the memory of an angular displacement, detected by the semicircular canals, is best characterized by an exponential function or by a power function. In 27 subjects a conflict was created between the semicircular canals and the graviceptive systems. Subjects were seated, facing forwards, in the gondola of a large centrifuge. The centrifuge was accelerated from stationary to 2.5Gz. While the swing out of the gondola (66°) during acceleration constitutes a frontal plane angular-displacement stimulus to the semicircular canals, the graviceptive systems persistently signal that the subject is upright. During 6 min at 2.5Gz the perceived head and body position was recorded; in darkness the subject repeatedly adjusted the orientation of a luminous line so that it appeared to be horizontal. Acceleration of the centrifuge induced a sensation of tilt which declined with time in a characteristic way. A three-parameter exponential function (Y = Ae(-bt) + C) and a power function (Y = At(-b) + C) were fitted to the data points. The inter-individual variability was considerable. In the vast majority of cases, however, the exponential function provided a better fit (in terms of RMS error) than the power function. The mean exponential function was: y = 27.8e(-0.018t) + 0.5°, where t is time in seconds. Findings are discussed with connection to possible underlying neural mechanisms; in particular, the head-direction system and short-term potentiation and persistent action potential firing in the hippocampus are considered.

S-shaped learning curves

In this article, learning curves for foreign vocabulary words are investigated, distinguishing between a subject-specific learning rate and a material-specific parameter that is related to the complexity of the items, such as the number of syllables. Two experiments are described, one with Turkish words and one with Italian words. In both, S-shaped learning curves were observed, which were most obvious if the subjects were not very familiar with the materials and if they were slow learners. With prolonged learning, the S shapes disappeared. Three different mathematical functions are proposed to explain these S-shaped curves. A further analysis clarifies why S-shaped learning curves may go unnoticed in many experiments.

The Impacts of Video Games on Cognition (and How the Government Can Guide the Industry)

Video game play has become a pervasive part of American culture. The dramatic increase in the popularity of video games has resulted in significant interest in the effects that video gaming may have on the brain and behavior. The scientific research to date indicates that some, but not all, commercial video games do indeed have the potential to cause large-scale changes in a wide variety of aspects of human behavior, including the focus of this review—cognitive abilities. More recent years have seen the rise of a separate form of video games, the so-called “brain games,” or games designed with the explicit goal of enhancing cognitive abilities. Although research on such brain games is still in its infancy, and the results have definitely not been uniformly positive, there is nonetheless reason for continued optimism that custom games can be developed that make a lasting and positive impact on human cognitive skills. Here, we discuss the current state of the scientific literature surrounding video games and human cognition with an emphasis on points critically related to public policy.

From means and variances to persons and patterns

A novel approach for conceptualizing and analyzing data from psychological studies is presented and discussed. This approach is centered on model building in an effort to explicate the structures and processes believed to generate a set of observations. These models therefore go beyond the variable-based, path models in use today which are limiting with regard to the types of inferences psychologists can draw from their research. In terms of analysis, the newer approach replaces traditional aggregate statistics such as means, variances, and covariances with methods of pattern detection and analysis. While these methods are person-centered and do not require parametric assumptions, they are both demanding and rigorous. They also provide psychologists with the information needed to draw the primary inference they often wish to make from their research; namely, the inference to best explanation.

Robot assistance of motor learning: A neuro-cognitive perspective

The last several years have seen a number of approaches to robot assistance of motor learning. Experimental studies have produced a range of findings from beneficial effects through null-effects to detrimental effects of robot assistance. In this review we seek an answer to the question under which conditions which outcomes should be expected. For this purpose we derive tentative predictions based on a classification of learning tasks in terms of the products of learning, the mechanisms involved, and the modulation of these mechanisms by robot assistance. Consistent with these predictions, the learning of dynamic features of trajectories is facilitated and the learning of kinematic and dynamic transformations is impeded by robotic guidance, whereas the learning of dynamic transformations can profit from robot assistance with error-amplifying forces. Deviating from the predictions, learning of spatial features of trajectories is impeded by haptic guidance, but can be facilitated by divergent force fields. The deviations point to the existence of additional effects of robot assistance beyond the modulation of learning mechanisms, e.g., the induction of a passive role of the motor system during practice with haptic guidance.

Conditional discrimination and response chains by worker bumblebees (Bombus impatiens Cresson, Hymenoptera: Apidae)

We trained worker bumblebees to discriminate arrays of artificial nectaries (one, two, and three microcentrifuge tubes inserted into artificial flowers) from which they could forage in association with their location in a three-compartmental maze. Additionally, we challenged bees to learn to accomplish three different tasks in a fixed sequence during foraging. To enter the main three-compartmented foraging arena, they had first to slide open doors in an entry box to be able to proceed to an artificial flower patch in the main arena where they had to lift covers to the artificial nectaries from which they then fed. Then, the bees had to return to the entrance way to their hive, but to actually enter, were challenged to rotate a vertically oriented disc to expose the entry hole. The bees were adept at associating the array of nectaries with their position in the compartmental maze (one nectary in compartment one, two in two, and three in three), taking about six trials to arrive at almost error-free foraging. Over all it took the bees three days of shaping to become more or less error free at the multi-step suite of sequential task performances. Thus, they had learned where they were in the chain sequence, which array and in which compartment was rewarding, how to get to the rewarding array in the appropriate compartment, and finally how to return as directly as possible to their hive entrance, open the entrance, and re-enter the hive. Our experiments were not designed to determine the specific nature of the cues the bees used, but our results strongly suggest that the tested bees developed a sense of subgoals that needed to be achieved by recognizing the array of elements in a pattern and possibly chain learning in order to achieve the ultimate goal of successfully foraging and returning to their colony. Our results also indicate that the bees had organized their learning by a hierarchy as evidenced by their proceeding to completion of the ultimate goal without reversing their foraging paths so as to return to the colony without food.

Evidence for age-associated cognitive decline from Internet game scores

Introduction

Lumosity's Memory Match (LMM) is an online game requiring visual working memory. Change in LMM scores may be associated with individual differences in age-related changes in working memory.

Methods

Effects of age and time on LMM learning and forgetting rates were estimated using data from 1890 game sessions for users aged 40 to 79 years.

Results

There were significant effects of age on baseline LMM scores (β = −.31, standard error or SE = .02, P < .0001) and lower learning rates (β = −.0066, SE = .0008, P < .0001). A sample size of 202 subjects/arm was estimated for a 1-year study for subjects in the lower quartile of game performance.

Discussion

Online memory games have the potential to identify age-related decline in cognition and to identify subjects at risk for cognitive decline with smaller sample sizes and lower cost than traditional recruitment methods.

Learning to explore the structure of kinematic objects in a virtual environment

The current study tested the quantity and quality of human exploration learning in a virtual environment. Given the everyday experience of humans with physical object exploration, we document substantial practice gains in the time, force, and number of actions needed to classify the structure of virtual chains, marking the joints as revolute, prismatic, or rigid. In line with current work on skill acquisition, participants could generalize the new and efficient psychomotor patterns of object exploration to novel objects. On the one hand, practice gains in exploration performance could be captured by a negative exponential practice function. On the other hand, they could be linked to strategies and strategy change. After quantifying how much was learned in object exploration and identifying the time course of practice-related gains in exploration efficiency (speed), we identified what was learned. First, we identified strategy components that were associated with efficient (fast) exploration performance: sequential processing, simultaneous use of both hands, low use of pulling rather than pushing, and low use of force. Only the latter was beneficial irrespective of the characteristics of the other strategy components. Second, we therefore characterized efficient exploration behavior by strategies that simultaneously take into account the abovementioned strategy components. We observed that participants maintained a high level of flexibility, sampling from a pool of exploration strategies trading the level of psycho-motoric challenges with exploration speed. We discuss the findings pursuing the aim of advancing intelligent object exploration by combining analytic (object exploration in humans) and synthetic work (object exploration in robots) in the same virtual environment.

The Now-or-Never bottleneck: A fundamental constraint on language

Memory is fleeting. New material rapidly obliterates previous material. How, then, can the brain deal successfully with the continual deluge of linguistic input? We argue that, to deal with this "Now-or-Never" bottleneck, the brain must compress and recode linguistic input as rapidly as possible. This observation has strong implications for the nature of language processing: (1) the language system must "eagerly" recode and compress linguistic input; (2) as the bottleneck recurs at each new representational level, the language system must build a multilevel linguistic representation; and (3) the language system must deploy all available information predictively to ensure that local linguistic ambiguities are dealt with "Right-First-Time"; once the original input is lost, there is no way for the language system to recover. This is "Chunk-and-Pass" processing. Similarly, language learning must also occur in the here and now, which implies that language acquisition is learning to process, rather than inducing, a grammar. Moreover, this perspective provides a cognitive foundation for grammaticalization and other aspects of language change. Chunk-and-Pass processing also helps explain a variety of core properties of language, including its multilevel representational structure and duality of patterning. This approach promises to create a direct relationship between psycholinguistics and linguistic theory. More generally, we outline a framework within which to integrate often disconnected inquiries into language processing, language acquisition, and language change and evolution.

Learning-induced autonomy of sensorimotor systems

Distributed networks of brain areas interact with one another in a time-varying fashion to enable complex cognitive and sensorimotor functions. Here we used new network-analysis algorithms to test the recruitment and integration of large-scale functional neural circuitry during learning. Using functional magnetic resonance imaging data acquired from healthy human participants, we investigated changes in the architecture of functional connectivity patterns that promote learning from initial training through mastery of a simple motor skill. Our results show that learning induces an autonomy of sensorimotor systems and that the release of cognitive control hubs in frontal and cingulate cortices predicts individual differences in the rate of learning on other days of practice. Our general statistical approach is applicable across other cognitive domains and provides a key to understanding time-resolved interactions between distributed neural circuits that enable task performance.

Bayesian estimation of multinomial processing tree models with heterogeneity in participants and items

Multinomial processing tree (MPT) models are theoretically motivated stochastic models for the analysis of categorical data. Here we focus on a crossed-random effects extension of the Bayesian latent-trait pair-clustering MPT model. Our approach assumes that participant and item effects combine additively on the probit scale and postulates (multivariate) normal distributions for the random effects. We provide a WinBUGS implementation of the crossed-random effects pair-clustering model and an application to novel experimental data. The present approach may be adapted to handle other MPT models.

Learning to tie well with others: bimanual versus intermanual performance of a highly practised skill

Studies indicate that novices are faster in manual tasks when performing with a partner ('intermanual') than with their own two hands ('bimanual'). The generality of this 'mode effect' was examined using a highly practised bimanual task, shoe tying, at which participants were experts. Speed-variability correlations confirmed participants were bimanually skilled but not intermanually skilled. Contrary to results using novices, intermanual was slower, such that prior skill reverses the effect. Analyses incorporating the similarity of shoe-tying strategies across dyads implicated a perceptual rather than shared knowledge/representation basis for intermanual performance. Practice effects indicated that intermanual performance built upon prior bimanual skill, such that novel relative timings between dyads' hands must be acquired. Motor transfer effects provided support for this conclusion. During shoe tying, hands were tightly coupled in the intermanual mode due to the perceptual coupling constraints of intermanual performance. Increased coupling was correlated with slower performance. Implications for real-world tasks (e.g. surgical knot tying) are described.

A Bayesian hierarchical diffusion model decomposition of performance in Approach-Avoidance Tasks

Common methods for analysing response time (RT) tasks, frequently used across different disciplines of psychology, suffer from a number of limitations such as the failure to directly measure the underlying latent processes of interest and the inability to take into account the uncertainty associated with each individual's point estimate of performance. Here, we discuss a Bayesian hierarchical diffusion model and apply it to RT data. This model allows researchers to decompose performance into meaningful psychological processes and to account optimally for individual differences and commonalities, even with relatively sparse data. We highlight the advantages of the Bayesian hierarchical diffusion model decomposition by applying it to performance on Approach-Avoidance Tasks, widely used in the emotion and psychopathology literature. Model fits for two experimental data-sets demonstrate that the model performs well. The Bayesian hierarchical diffusion model overcomes important limitations of current analysis procedures and provides deeper insight in latent psychological processes of interest.

Multifaceted aspects of chunking enable robust algorithms

Sequence production tasks are a standard tool to analyze motor learning, consolidation, and habituation. As sequences are learned, movements are typically grouped into subsets or chunks. For example, most Americans memorize telephone numbers in two chunks of three digits, and one chunk of four. Studies generally use response times or error rates to estimate how subjects chunk, and these estimates are often related to physiological data. Here we show that chunking is simultaneously reflected in reaction times, errors, and their correlations. This multimodal structure enables us to propose a Bayesian algorithm that better estimates chunks while avoiding overfitting. Our algorithm reveals previously unknown behavioral structure, such as an increased error correlations with training, and promises a useful tool for the characterization of many forms of sequential motor behavior.

Learning curves in highly skilled chess players: a test of the generality of the power law of practice

The power law of practice holds that a power function best interrelates skill performance and amount of practice. However, the law's validity and generality are moot. Some researchers argue that it is an artifact of averaging individual exponential curves while others question whether the law generalizes to complex skills and to performance measures other than response time. The present study tested the power law's generality to development over many years of a very complex cognitive skill, chess playing, with 387 skilled participants, most of whom were grandmasters. A power or logarithmic function best fit grouped data but individuals showed much variability. An exponential function usually was the worst fit to individual data. Groups differing in chess talent were compared and a power function best fit the group curve for the more talented players while a quadratic function best fit that for the less talented. After extreme amounts of practice, a logarithmic function best fit grouped data but a quadratic function best fit most individual curves. Individual variability is great and the power law or an exponential law are not the best descriptions of individual chess skill development.

Playing off the curve - testing quantitative predictions of skill acquisition theories in development of chess performance

Learning curves have been proposed as an adequate description of learning processes, no matter whether the processes manifest within minutes or across years. Different mechanisms underlying skill acquisition can lead to differences in the shape of learning curves. In the current study, we analyze the tournament performance data of 1383 chess players who begin competing at young age and play tournaments for at least 10 years. We analyze the performance development with the goal to test the adequacy of learning curves, and the skill acquisition theories they are based on, for describing and predicting expertise acquisition. On the one hand, we show that the skill acquisition theories implying a negative exponential learning curve do a better job in both describing early performance gains and predicting later trajectories of chess performance than those theories implying a power function learning curve. On the other hand, the learning curves of a large proportion of players show systematic qualitative deviations from the predictions of either type of skill acquisition theory. While skill acquisition theories predict larger performance gains in early years and smaller gains in later years, a substantial number of players begin to show substantial improvements with a delay of several years (and no improvement in the first years), deviations not fully accounted for by quantity of practice. The current work adds to the debate on how learning processes on a small time scale combine to large-scale changes.

Effects of video-game play on information processing: a meta-analytic investigation

Do video games enhance cognitive functioning? We conducted two meta-analyses based on different research designs to investigate how video games impact information-processing skills (auditory processing, executive functions, motor skills, spatial imagery, and visual processing). Quasi-experimental studies (72 studies, 318 comparisons) compare habitual gamers with controls; true experiments (46 studies, 251 comparisons) use commercial video games in training. Using random-effects models, video games led to improved information processing in both the quasi-experimental studies, d = 0.61, 95% CI [0.50, 0.73], and the true experiments, d = 0.48, 95% CI [0.35, 0.60]. Whereas the quasi-experimental studies yielded small to large effect sizes across domains, the true experiments yielded negligible effects for executive functions, which contrasted with the small to medium effect sizes in other domains. The quasi-experimental studies appeared more susceptible to bias than were the true experiments, with larger effects being reported in higher-tier than in lower-tier journals, and larger effects reported by the most active research groups in comparison with other labs. The results are further discussed with respect to other moderators and limitations in the extant literature.

Dissociable mechanisms of speed-accuracy tradeoff during visual perceptual learning are revealed by a hierarchical drift-diffusion model

Two phenomena are commonly observed in decision-making. First, there is a speed-accuracy tradeoff (SAT) such that decisions are slower and more accurate when instructions emphasize accuracy over speed, and vice versa. Second, decision performance improves with practice, as a task is learnt. The SAT and learning effects have been explained under a well-established evidence-accumulation framework for decision-making, which suggests that evidence supporting each choice is accumulated over time, and a decision is committed to when the accumulated evidence reaches a decision boundary. This framework suggests that changing the decision boundary creates the tradeoff between decision speed and accuracy, while increasing the rate of accumulation leads to more accurate and faster decisions after learning. However, recent studies challenged the view that SAT and learning are associated with changes in distinct, single decision parameters. Further, the influence of speed-accuracy instructions over the course of learning remains largely unknown. Here, we used a hierarchical drift-diffusion model to examine the SAT during learning of a coherent motion discrimination task across multiple training sessions, and a transfer test session. The influence of speed-accuracy instructions was robust over training and generalized across untrained stimulus features. Emphasizing decision accuracy rather than speed was associated with increased boundary separation, drift rate and non-decision time at the beginning of training. However, after training, an emphasis on decision accuracy was only associated with increased boundary separation. In addition, faster and more accurate decisions after learning were due to a gradual decrease in boundary separation and an increase in drift rate. The results suggest that speed-accuracy instructions and learning differentially shape decision-making processes at different time scales.

Tracing the trajectory of skill learning with a very large sample of online game players

In the present study, we analyzed data from a very large sample (N = 854,064) of players of an online game involving rapid perception, decision making, and motor responding. Use of game data allowed us to connect, for the first time, rich details of training history with measures of performance from participants engaged for a sustained amount of time in effortful practice. We showed that lawful relations exist between practice amount and subsequent performance, and between practice spacing and subsequent performance. Our methodology allowed an in situ confirmation of results long established in the experimental literature on skill acquisition. Additionally, we showed that greater initial variation in performance is linked to higher subsequent performance, a result we link to the exploration/exploitation trade-off from the computational framework of reinforcement learning. We discuss the benefits and opportunities of behavioral data sets with very large sample sizes and suggest that this approach could be particularly fecund for studies of skill acquisition.

Revisiting the learning curve (once again)

The vast majority of published work in the field of associative learning seeks to test the adequacy of various theoretical accounts of the learning process using average data. Of course, averaging hides important information, but individual departures from the average are usually designated "error" and largely ignored. However, from the perspective of an individual differences approach, this error is the data of interest; and when associative models are applied to individual learning curves the error is substantial. To some extent individual differences can be reasonably understood in terms of parametric variations of the underlying model. Unfortunately, in many cases, the data cannot be accomodated in this way and the applicability of the underlying model can be called into question. Indeed several authors have proposed alternatives to associative models because of the poor fits between data and associative model. In the current paper a novel associative approach to the analysis of individual learning curves is presented. The Memory Environment Cue Array Model (MECAM) is described and applied to two human predictive learning datasets. The MECAM is predicated on the assumption that participants do not parse the trial sequences to which they are exposed into independent episodes as is often assumed when learning curves are modeled. Instead, the MECAM assumes that learning and responding on a trial may also be influenced by the events of the previous trial. Incorporating non-local information the MECAM produced better approximations to individual learning curves than did the Rescorla-Wagner Model (RWM) suggesting that further exploration of the approach is warranted.

Statistical word learning at scale: the baby's view is better

A key question in early word learning is how children cope with the uncertainty in natural naming events. One potential mechanism for uncertainty reduction is cross-situational word learning - tracking word/object co-occurrence statistics across naming events. But empirical and computational analyses of cross-situational learning have made strong assumptions about the nature of naming event ambiguity, assumptions that have been challenged by recent analyses of natural naming events. This paper shows that learning from ambiguous natural naming events depends on perspective. Natural naming events from parent-child interactions were recorded from both a third-person tripod-mounted camera and from a head-mounted camera that produced a 'child's-eye' view. Following the human simulation paradigm, adults were asked to learn artificial language labels by integrating across the most ambiguous of these naming events. Significant learning was found only from the child's perspective, pointing to the importance of considering statistical learning from an embodied perspective.

Individual differences in task-specific paired associates learning in older adults: the role of processing speed and working memory

BACKGROUND/STUDY CONTEXT: The role of processing speed and working memory was investigated in terms of individual differences in task-specific paired associates learning in a sample of older adults. Task-specific learning, as distinct from content-oriented item-specific learning, refers to gains in performance due to repeated practice on a learning task in which the to-be-learned material changes over trials.

METHODS

Learning trajectories were modeled within an intensive repeated-measures design based on participants obtained from an opt-in Internet-based sampling service (M(age) = 65.3, SD = 4.81). Participants completed an eight-item paired associates task daily over a 7-day period.

RESULTS

Results indicated that a three-parameter hyperbolic model (i.e., initial level, learning rate, and asymptotic performance) best described learning trajectory. After controlling for age-related effects, both higher working memory and higher processing speed had a positive effect on all three learning parameters.

CONCLUSION

These results emphasize the role of cognitive abilities for individual differences in task-specific learning of older adults.

Task-based core-periphery organization of human brain dynamics

As a person learns a new skill, distinct synapses, brain regions, and circuits are engaged and change over time. In this paper, we develop methods to examine patterns of correlated activity across a large set of brain regions. Our goal is to identify properties that enable robust learning of a motor skill. We measure brain activity during motor sequencing and characterize network properties based on coherent activity between brain regions. Using recently developed algorithms to detect time-evolving communities, we find that the complex reconfiguration patterns of the brain's putative functional modules that control learning can be described parsimoniously by the combined presence of a relatively stiff temporal core that is composed primarily of sensorimotor and visual regions whose connectivity changes little in time and a flexible temporal periphery that is composed primarily of multimodal association regions whose connectivity changes frequently. The separation between temporal core and periphery changes over the course of training and, importantly, is a good predictor of individual differences in learning success. The core of dynamically stiff regions exhibits dense connectivity, which is consistent with notions of core-periphery organization established previously in social networks. Our results demonstrate that core-periphery organization provides an insightful way to understand how putative functional modules are linked. This, in turn, enables the prediction of fundamental human capacities, including the production of complex goal-directed behavior.

When someone learns a new skill, his/her brain dynamically alters individual synapses, regional activity, and larger-scale circuits. In this paper, we capture some of these dynamics by measuring and characterizing patterns of coherent brain activity during the learning of a motor skill. We extract time-evolving communities from these patterns and find that a temporal core that is composed primarily of primary sensorimotor and visual regions reconfigures little over time, whereas a periphery that is composed primarily of multimodal association regions reconfigures frequently. The core consists of densely connected nodes, and the periphery consists of sparsely connected nodes. Individual participants with a larger separation between core and periphery learn better in subsequent training sessions than individuals with a smaller separation. Conceptually, core-periphery organization provides a framework in which to understand how putative functional modules are linked. This, in turn, enables the prediction of fundamental human capacities, including the production of complex goal-directed behavior.

Learning to never forget-time scales and specificity of long-term memory of a motor skill

Despite anecdotal reports that humans retain acquired motor skills for many years, if not a lifetime, long-term memory of motor skills has received little attention. While numerous neuroimaging studies showed practice-induced cortical plasticity, the behavioral correlates, what is retained and also what is forgotten, are little understood. This longitudinal case study on four subjects presents detailed kinematic analyses of humans practicing a bimanual polyrhythmic task over 2 months with retention tests after 6 months and, for two subjects, after 8 years. Results showed that individuals not only retained the task, but also reproduced their individual "style" of performance, even after 8 years. During practice, variables such as the two hands' frequency ratio and relative phase, changed at different rates, indicative of multiple time scales of neural processes. Frequency leakage across hands, reflecting intermanual crosstalk, attenuated at a significantly slower rate and was the only variable not maintained after 8 years. Complementing recent findings on neuroplasticity in gray and white matter, our study presents new behavioral evidence that highlights the multi-scale process of practice-induced changes and its remarkable persistence. Results suggest that motor memory may comprise not only higher-level task variables but also individual kinematic signatures.

Network dysfunction during associative learning in schizophrenia: Increased activation, but decreased connectivity: an fMRI study

Schizophrenia (SCZ) is characterized by disordered activation and disordered connectivity, yet few fMRI studies have convergently investigated both. Here, we compared differences in activation and connectivity between SCZ and controls (HC). Twenty-two subjects (18≤age≤35yrs) participated in a paired-associative learning task, a behavioral domain particularly dependent on fronto-hippocampal connectivity and of relevance to the schizophrenia diathesis. Activation differences were assessed using standard approaches. Seed-based connectivity differences were compared using Psychophysiological Interaction (PPI) with a hippocampus-based seed. SCZ evinced significantly increased activation, but significantly decreased connectivity with the hippocampus across a cortical-striatal learning network. These results assess potentially complementary patterns of network dysfunction in schizophrenia: increased activation suggests inefficient responses relating to functional specialization; decreased connectivity suggests impaired integration of functional signals between regions. Inefficiency and dysconnection appear to collectively characterize functional deficits in schizophrenia.

Skill acquisition with text-entry interfaces: particularly older users benefit from minimized information-processing demands

Operating information technology challenges older users if it requires executive control, which generally declines with age. Especially for novel and occasional tasks, cognitive demands can be high. We demonstrate how interface design can reduce cognitive demands by studying skill acquisition with the destination entry interfaces of two customary route guidance systems. Young, middle-aged, and older adults performed manual destination entry either with a system operated with multiple buttons in a dialogue encompassing spelling and list selection, or with a system operated by a single rotary encoder, in which an intelligent speller constrained destination entry to a single line of action. Each participant performed 100 training trials. A retention test after at least 10 weeks encompassed 20 trials. The same task was performed faster, more accurately, and produced much less age-related performance differences especially at the beginning of training if interface design reduced demand for executive control, perceptual processing, and motor control.

How Mean is the Mean?

In this paper we voice concerns about the uncritical manner in which the mean is often used as a summary statistic in psychological research. We identify a number of implicit assumptions underlying the use of the mean and argue that the fragility of these assumptions should be more carefully considered. We examine some of the ways in which the potential violation of these assumptions can lead us into significant theoretical and methodological error. Illustrations of alternative models of research already extant within Psychology are used to explore methods of research less mean-dependent and suggest that a critical assessment of the assumptions underlying its use in research play a more explicit role in the process of study design and review.

Two routes to expertise in mental rotation

The ability to imagine objects undergoing rotation (mental rotation) improves markedly with practice, but an explanation of this plasticity remains controversial. Some researchers propose that practice speeds up the rate of a general-purpose rotation algorithm. Others maintain that performance improvements arise through the adoption of a new cognitive strategy-repeated exposure leads to rapid retrieval from memory of the required response to familiar mental rotation stimuli. In two experiments we provide support for an integrated explanation of practice effects in mental rotation by combining behavioral and EEG measures in a way that provides more rigorous inference than is available from either measure alone. Before practice, participants displayed two well-established signatures of mental rotation: Both response time and EEG negativity increased linearly with rotation angle. After extensive practice with a small set of stimuli, both signatures of mental rotation had all but disappeared. In contrast, after the same amount of practice with a much larger set both signatures remained, even though performance improved markedly. Taken together, these results constitute a reversed association, which cannot arise from variation in a single cause, and so they provide compelling evidence for the existence of two routes to expertise in mental rotation. We also found novel evidence that practice with the large but not the small stimulus set increased the magnitude of an early visual evoked potential, suggesting increased rotation speed is enabled by improved efficiency in extracting three-dimensional information from two-dimensional stimuli.

Task difficulty and the time scales of warm-up and motor learning

The authors investigated the influence of task difficulty on warm-up decrement and learning across practice sessions. Three groups of participants practiced a star-tracing task over 3 consecutive days with different levels (e.g., easy, medium, hard) of task difficulty. The performance data were modeled with a 2 time scale function that represented the transient, fast time scale process of warm-up decrement superimposed with the persistent, slow time scale process of learning. Movement time decreased as a function of practice with the most difficult condition exhibiting the greatest reduction though still the longest movement time. The 2 time scale model provided a better fit to the data than an exponential or power law function and showed that the 3 difficulty conditions exhibited similar rates of change for the respective slow (i.e., learning) and fast (i.e., warm-up decrement) time scale processes that varied by an order of magnitude. Task difficulty was inversely related to the initial level of warm-up decrement but not the rate of performance recovery early in a practice session. The findings support the postulation that there is a persistent learned component to the initial conditions in subsequent practice sessions but that there is a common time scale of accommodating the transient process of warm-up decrement.

The "pure-study" learning curve: the learning curve without cumulative testing

The customary assumption in the study of human learning using alternating study and test trials is that learning occurs during study trials and that test trials are useful only to measure learning. In fact, tests seem to play little role in the development of learning, because the learning curve is similar even when the number of test trials varies widely (Tulving, Journal of verbal learning and verbal behavior 6:175-184, 1967). However, this outcome seems odd, because other research has shown that testing fosters greater long-term learning than does studying. We report three experiments addressing whether tests affect the shape of the learning curve. In two of the experiments, we examined this issue by varying the number of spaced study trials in a sequence and examining performance on only a single test trial at the end of the series (a "pure-study" learning curve). We compared these pure-study learning curves to standard learning curves and found that the standard curves increase more rapidly and reach a higher level in both free recall (exp. 1) and paired-associate learning (exp. 2). In Experiment 3, we provided additional study trials in the "pure-study" condition to determine whether the standard (study-test) condition would prove superior to a study-study condition. The standard condition still produced better retention on both immediate and delayed tests. Our experiments show that test trials play an important role in the development of learning using both free-recall (exps. 1 and 3) and paired-associate (exp. 2) procedures. Theories of learning have emphasized processes that occur during study, but our results show that processes engaged during tests are also critical.

Effective part-task training as evidence of distinct adaptive processes with different time scales

For some types of visuo-motor transformations like large visuo-motor rotations or the complex transformation of a sliding first-order lever, distinct adaptive processes have been hypothesized that produce a rapid, discrete approximation of the transformation and a slow, graded fine tuning, respectively. Here we investigate whether part-task training of only the second of these processes, namely the fine tuning, transfers to the subsequent performance in a condition with the full transformation of the sliding first-order lever. Therefore, we compared performance of three groups with different practice conditions during transfer to the full transformation. While two groups only practiced the fine tuning without the right-left inversion of the lever prior to transfer, a third group practiced the full lever transformation. Our results show a positive, but less than perfect transfer of the isolated practice of the fine tuning on performance with the full transformation. For the fine tuning itself, transfer was not reliably different from being perfect. The observation that the fine tuning can be acquired separately and added to the later adaptation to the left-right inversion of the lever supports the notion that these slow and fast processes progress rather independently. The additional finding that the preceding acquisition of the fine tuning also facilitates the subsequent rapid process could be due to generalized learning-to-learn or to a more precise assignment of movement errors to the process from which they originate.

Explicit pre-training instruction does not improve implicit perceptual-motor sequence learning

Memory systems theory argues for separate neural systems supporting implicit and explicit memory in the human brain. Neuropsychological studies support this dissociation, but empirical studies of cognitively healthy participants generally observe that both kinds of memory are acquired to at least some extent, even in implicit learning tasks. A key question is whether this observation reflects parallel intact memory systems or an integrated representation of memory in healthy participants. Learning of complex tasks in which both explicit instruction and practice is used depends on both kinds of memory, and how these systems interact will be an important component of the learning process. Theories that posit an integrated, or single, memory system for both types of memory predict that explicit instruction should contribute directly to strengthening task knowledge. In contrast, if the two types of memory are independent and acquired in parallel, explicit knowledge should have no direct impact and may serve in a "scaffolding" role in complex learning. Using an implicit perceptual-motor sequence learning task, the effect of explicit pre-training instruction on skill learning and performance was assessed. Explicit pre-training instruction led to robust explicit knowledge, but sequence learning did not benefit from the contribution of pre-training sequence memorization. The lack of an instruction benefit suggests that during skill learning, implicit and explicit memory operate independently. While healthy participants will generally accrue parallel implicit and explicit knowledge in complex tasks, these types of information appear to be separately represented in the human brain consistent with multiple memory systems theory.

From universal laws of cognition to specific cognitive models

The remarkable successes of the physical sciences have been built on highly general quantitative laws, which serve as the basis for understanding an enormous variety of specific physical systems. How far is it possible to construct universal principles in the cognitive sciences, in terms of which specific aspects of perception, memory, or decision making might be modelled? Following Shepard (e.g., 1987), it is argued that some universal principles may be attainable in cognitive science. Here, 2 examples are proposed: the simplicity principle (which states that the cognitive system prefers patterns that provide simpler explanations of available data); and the scale-invariance principle, which states that many cognitive phenomena are independent of the scale of relevant underlying physical variables, such as time, space, luminance, or sound pressure. This article illustrates how principles may be combined to explain specific cognitive processes by using these principles to derive SIMPLE, a formal model of memory for serial order (Brown, Neath, & Chater, 2007), and briefly mentions some extensions to models of identification and categorization. This article also considers the scope and limitations of universal laws in cognitive science.

Practice and forgetting effects on vocabulary memory: an activation-based model of the spacing effect

An experiment was performed to investigate the effects of practice and spacing on retention of Japanese-English vocabulary paired associates. The relative benefit of spacing increased with increased practice and with longer retention intervals. Data were fitted with an activation-based memory model, which proposes that each time an item is practiced it receives an increment of strength but that these increments decay as a power function of time. The rate of decay for each presentation depended on the activation at the time of the presentation. This mechanism limits long-term benefits from further practice at higher levels of activation and produces the spacing effect and its observed interactions with practice and retention interval. The model was compared with another model of the spacing effect (Raaijmakers, 2003) and was fit to some results from the literature on spacing and memory.

Word learning emerges from the interaction of online referent selection and slow associative learning

Classic approaches to word learning emphasize referential ambiguity: In naming situations, a novel word could refer to many possible objects, properties, actions, and so forth. To solve this, researchers have posited constraints, and inference strategies, but assume that determining the referent of a novel word is isomorphic to learning. We present an alternative in which referent selection is an online process and independent of long-term learning. We illustrate this theoretical approach with a dynamic associative model in which referent selection emerges from real-time competition between referents and learning is associative (Hebbian). This model accounts for a range of findings including the differences in expressive and receptive vocabulary, cross-situational learning under high degrees of ambiguity, accelerating (vocabulary explosion) and decelerating (power law) learning, fast mapping by mutual exclusivity (and differences in bilinguals), improvements in familiar word recognition with development, and correlations between speed of processing and learning. Together it suggests that (a) association learning buttressed by dynamic competition can account for much of the literature; (b) familiar word recognition is subserved by the same processes that identify the referents of novel words (fast mapping); (c) online competition may allow the children to leverage information available in the task to augment performance despite slow learning; (d) in complex systems, associative learning is highly multifaceted; and (e) learning and referent selection, though logically distinct, can be subtly related. It suggests more sophisticated ways of describing the interaction between situation- and developmental-time processes and points to the need for considering such interactions as a primary determinant of development.

A hierarchical model for estimating response time distributions

We present a statistical model for inference with response time (RT) distributions. The model has the following features. First, it provides a means of estimating the shape, scale, and location (shift) of RT distributions. Second, it is hierarchical and models between-subjects and within-subjects variability simultaneously. Third, inference with the model is Bayesian and provides a principled and efficient means of pooling information across disparate data from different individuals. Because the model efficiently pools information across individuals, it is particularly well suited for those common cases in which the researcher collects a limited number of observations from several participants. Monte Carlo simulations reveal that the hierarchical Bayesian model provides more accurate estimates than several popular competitors do. We illustrate the model by providing an analysis of the symbolic distance effect in which participants can more quickly ascertain the relationship between nonadjacent digits than that between adjacent digits.

The effects of evidence bounds on decision-making: theoretical and empirical developments

Converging findings from behavioral, neurophysiological, and neuroimaging studies suggest an integration-to-boundary mechanism governing decision formation and choice selection. This mechanism is supported by sequential sampling models of choice decisions, which can implement statistically optimal decision strategies for selecting between multiple alternative options on the basis of sensory evidence. This review focuses on recent developments in understanding the evidence boundary, an important component of decision-making raised by experimental findings and models. The article starts by reviewing the neurobiology of perceptual decisions and several influential sequential sampling models, in particular the drift-diffusion model, the Ornstein–Uhlenbeck model and the leaky-competing-accumulator model. In the second part, the article examines how the boundary may affect a model’s dynamics and performance and to what extent it may improve a model’s fits to experimental data. In the third part, the article examines recent findings that support the presence and site of boundaries in the brain. The article considers two questions: (1) whether the boundary is a spontaneous property of neural integrators, or is controlled by dedicated neural circuits; (2) if the boundary is variable, what could be the driving factors behind boundary changes? The review brings together studies using different experimental methods in seeking answers to these questions, highlights psychological and physiological factors that may be associated with the boundary and its changes, and further considers the evidence boundary as a generic mechanism to guide complex behavior.

Task demands determine the specificity of the search template

When searching for an object, an observer holds a representation of the target in mind while scanning the scene. If the observer repeats the search, performance may become more efficient as the observer hones this target representation, or "search template," to match the specific demands of the search task. An effective search template must have two characteristics: It must reliably discriminate the target from the distractors, and it must tolerate variability in the appearance of the target. The present experiment examined how the tolerance of the search template is affected by the search task. Two groups of 18 observers trained on the same set of stimuli blocked either by target image (block-by-image group) or by target category (block-by-category group). One or two days after training, both groups were tested on a related search task. The pattern of test results revealed that the two groups of observers had developed different search templates, and that the templates of the block-by-category observers better captured the general characteristics of the category. These results demonstrate that observers match their search templates to the demands of the search task.