Performance and learning of generalized motor programs: relative (GMP) and absolute (parameter) errors

The effects of practice (Experiment 1) and parameter variability (Experiment 2) on the learning of generalized motor programs (GMPs) and movement parameterization were investigated. In each experiment, 2 tasks with different relative force-time structures were tested. Participants (N = 32, Experiment 1; N = 40, Experiment 2) attempted to exert a pattern of force that resembled in force and time a waveform that was displayed on a computer monitor. In both experiments, the analysis suggested that the GMP, although refined over practice, was relatively stable (i.e., resistant to decay and interference), even early in practice (after 20 trials). In addition, the results indicated that constant and variable parameter practice did not differentially affect GMP learning but did degrade the learning of the parameter that was not varied. The data provided additional evidence for the dissociation of the GMP and the parameterization processes proposed in GMP theory. Contrary to schema theory, the present data suggest an interdependence between the force and the time parameters: The manipulation of 1 of the parameters has a negative effect on the learning of the other parameter.     

Performance and Learning of Generalized Motor Programs: Relative (GMP) and Absolute (Parameter) Errors

The effects of practice (Experiment 1) and parameter variability (Experiment 2) on the learning of generalized motor programs (GMPs) and movement parameterization were investigated In each experiment, 2 tasks with different relative force-time structures were tested. Participants (N = 32, Experiment (N = 40, Experiment 2) attempted to exert a pattern of force that resembled in force and time a waveform that was displayed on a computer monitor. In both experiments, the analysis suggested that the GMP, although refined over practice, was relatively stable (i.e., resistant to decay and interference), even early in practice (after 20 trials). In addition, the results indicated that constant and variable parameter practice did not differentially affect GMP learning but did degrade the learning of the parameter that was not varied. The data provided additional evidence for the dissociation of the GMP and the parameterization processes proposed in GMP theory. Contrary to schema theory, the present data suggest an interdependence between the force and the time parameters: The manipulation of 1 of the parameters has a negative effect on the learning of the other parameter.     

Influence of Practice Schedule on Testing Schema Theory Predictions in Adults

Since the publication of Schmidt’s (1975) schema theory of motor learning, numerous studies have assessed the variability of practice hypothesis. Of these, only research using children as subjects has provided consistent results. Findings from adult studies have been equivocal. Two experiments were conducted to assess the possibility that the schedule of variable practice during acquisition may be a clue to this equivocality, since only experiments (using adults as subjects) that have provided random variability of practice conditions have consistently supported the tenets of schema theory. Using a two movement rapid timing task, subjects learned to control their actions under one (constant) or four (variable) parameters, with variable practice conducted under either a blocked or a randomized schedule. Results for variable error and absolute constant error showed that random-variable practice provided strong support for the schema theory prediction, whereas blocked-variable practice provided only relatively weak support. Two contrasting implications of these results are discussed with respect to support versus non-support for Schmidt’s schema theory.     

The schematic representation of effector function underlying perceptual-motor skill

Conceptual and methodological problems related to Schmidt' (1975) motor schema theory are discussed. In particular, the motor schema is interpreted as representing the dynamics of the system being controlled, which may or may not be associated with a referent movement pattern. Furthermore, it is suggested that prior familiarity with a control system's dynamics is a critical but uncontrolled factor in tests of the theory, and largely accounts for their equivocal findings. These ideas are examined by two experiments in which subjects had to bimanually control the movement of a computer-displayed cursor along a track on a CRT screen. Different track orientations required different patterns of movement not entailing a single generalized motor program. Experiment 1 shows that variable track performance with a given control system, results in better transfer to novel tracks than does fixed practice. Experiment 2 demonstrates that altering the control system disrupts performance whether or not the required movements remain the same. These results indicate the need for a fundamental modification of schema theory, such that a schematic representation of effector-environment relations (effector function) is available independently of particular movement patterns used in its acquisition.     

The Schematic Representation of Effector Function Underlying Perceptual-Motor Skill

Conceptual and methodological problems related to Schmidt’s (1975) motor schema theory are discussed. In particular, the motor schema is interpreted as representing the dynamics of the system being controlled, which may or may not be associated with a referent movement pattern. Furthermore, it is suggested that prior familiarity with a control system’s dynamics is a critical but uncontrolled factor in tests of the theory, and largely accounts for their equivocal findings. These ideas are examined by two experiments in which subjects had to bimanually control the movement of a computer-displayed cursor along a track on a CRT screen. Different track orientations required different patterns of movement not entailing a single generalized motor program. Experiment 1 shows that variable track performance with a given control system, results in better transfer to novel tracks than does fixed practice. Experiment 2 demonstrates that altering the control system disrupts performance whether or not the required movements remain the same. These results indicate the need for a fundamental modification of schema theory, such that a schematic representation of effector-environment relations (effector function) is available independently of particular movement patterns used in its acquisition.     

Influence of practice schedule on testing schema theory predictions in adults

Since the publication of Schmidt's (1975) schema theory of motor learning, numerous studies have assessed the variability of practice hypothesis. Of these, only research using children as subjects has provided consistent results. Findings from adult studies have been equivocal. Two experiments were conducted to assess the possibility that the schedule of variable practice during acquisition may be a clue to this equivocality, since only experiments (using adults as subjects) that have provide random variability of practice conditions have consistently supported the tenets of schema theory. Using a two movement rapid timing task, subjects learned to control their actions under one (constant) or four (variable) parameters, with variable practice conducted under either a blocked or a randomized schedule. Results for variable error and absolute constant error showed that random-variable practice provided strong support for the schema theory prediction, whereas blocked-variable practice provided only relatively weak support. Two contrasting implications of these results are discussed with respect to support versus non-support for Schmidt' schema theory.     

Variability of Practice and Contextual Interference in Motor Skill Learning

The purpose of this study was to investigate whether learning benefits in multiple-task learning situations are a result of contextual interference or of schema enhancement related to the amount of variability in the practice session. Two experiments were designed that replicated and extended the experiment reported by Wulf and Schmidt (1988). In a 2 (same vs. different relative time) x 2 (blocked vs. random practice schedule) design, 48 right-handed subjects were randomly assigned to one of four experimental conditions. A tapping task was employed that required a right-handed tap of three small brass plates arranged in a diamond pattern. Each segment had a specific time requirement. Target times and response times were provided on a computer screen directly in front of the subject. Each subject participated in two acquisition sessions (i.e., 198 practice trials) and was tested for learning on several different retention and transfer tests. In Experiment 2, a control group was added that received no acquisition phase. Results of both experiments showed a typical contextual interference effect, with depressed scores by the random groups during acquisition but significantly better scores than the blocked groups on several retention and transfer tests. Certain characteristics of the tests were found to influence the demonstration of the practice schedule effects. These results were consistent with predictions from Magill and Hall (1990) that the learning benefits of contextual interference are more likely to occur when skill variations are from different classes of movement and that the amount of variability in practice is more influential when the to-be-learned tasks are parameter modifications of the same generalized motor program.     

Average-KR schedule benefits generalized motor program learning

The main purpose of this study was to examine the effects of average Knowledge of Results (KR) on generalized motor program learning and parameter learning. Two groups of participants (n = 15 per group) performed 80 acquisition trials of sequential timing tasks. All participants were asked to depress sequentially four keys (2, 4, 8, and 6) on the numeric pad portion of the computer keyboard with the index finger of the right hand. The author presented average feedback on timing errors based on 5-trial blocks and compared this feedback schedule with every-trial feedback. Analysis of the delayed no-feedback retention test indicated a strong advantage for the average KR compared with the every-trial condition in both generalized motor program learning and parameter learning. The current results suggest that the average KR schedule may have positive effects on generalized motor program learning and parameter learning.     

Applying Deutsch’s concept of good explanations to artificial intelligence and neuroscience – An initial exploration

Artificial intelligence has made great strides since the deep learning revolution, but AI systems remain incapable of learning principles and rules which allow them to extrapolate outside of their training data to new situations. For inspiration we look to the domain of science, where scientists have been able to develop theories which show remarkable ability to extrapolate and sometimes even predict the existence of phenomena which have never been observed before. According to David Deutsch, this type of extrapolation, which he calls “reach”, is due to scientific theories being hard to vary. In this work we investigate Deutsch’s hard-to-vary principle and how it relates to more formalized principles in deep learning such as the bias-variance trade-off and Occam’s razor. We distinguish internal variability, how much a model/theory can be varied internally while still yielding the same predictions, with external variability, which is how much a model must be varied to predict new, out-of-distribution data. We discuss how to measure internal variability using the notion of the Rashomon set and how to measure external variability using Kolmogorov complexity. We explore what role hard-to-vary explanations play in intelligence by looking at the human brain, the only example of highly general purpose intelligence known. We distinguish two learning systems in the brain – the first operates similar to deep learning and likely underlies most of perception while the second is a more creative system capable of generating hard-to-vary models and explanations of the world. We make contact with Popperian epistemology which suggests that the generation of scientific theories is a not an inductive process but rather an evolutionary process which proceeds through conjecture and refutation. We argue that figuring out how replicate this second system, which is capable of generating hard-to-vary explanations, is a key challenge which needs to be solved in order to realize artificial general intelligence.     

Induction of Relational Schemas: Common Processes in Reasoning and Complex Learning

Five experiments were performed to test whether participants induced a coherent representation of the structure of a task, called a relational schema, from specific instances. Properties of a relational schema include: An explicit symbol for a relation, a binding that preserves the truth of a relation, potential for higher-order relations, omnidirectional access, potential for transfer between isomorphs, and ability to predict unseen items in isomorphic problems. However relational schemas are not necessarily coded in abstract form. Predictions from relational schema theory were contrasted with predictions from configural learning and other nonstructural theories in five experiments in which participants were taught a structure comprised of a set of initial-state,operator → end-state instances. The initial-state,operator pairs were presented and participants had to predict the correct end-state. Induction of a relational schema was achieved efficiently by adult participants as indicated by ability to predict items of a new isomorphic problem. The relational schemas induced showed the omnidirectional access property, there was efficient transfer to isomorphs, and structural coherence had a powerful effect on learning. The “learning to learn” effect traditionally associated with the learning set literature was observed, and the long-standing enigma of learning set acquisition is explained by a model composed of relational schema induction and structure mapping. Performance was better after reversal of operators than after shift to an alternate structure, even though the latter entailed more overlap with previously learned tasks in terms of the number of configural associations that were preserved. An explanation for the reversal shift phenomenon in terms of induction and mapping of a relational schema is proposed. The five experiments provided evidence supporting predictions from relational schema theory, and no evidence was found for configural or nonstructural learning theories.     

The pragmatic aspects of linguistic negation: Speech act,argumentation and pragmatic inference

This paper is an attempt to give a general explanation of pragmatic aspects of linguistic negation. After a brief survey of classical accounts of negation within pragmatic theories (as speech act theory, argumentation theory and polyphonic theory), the main pragmatic uses of negation (illocutionary negation, external negation, lowering and majoring negation) are discussed within relevance theory. The question of the relevance of negative utterance is raised, and a general inferential schema (based on the so-called invited inference) is proposed and tested for the main uses of negation discussed in the paper.Translated from French by Anne Reboul and Lyne DaSylva.     

Variability of prior practice in learning and retention of a novel motor response

The schema theory prediction that variability of practice would influence the strength of the motor response schema was investigated. Three groups of subjects, defined by much, little, or no variability of distances moved in prior practice on a linear-slide apparatus, completed 12 initial practice trials. Subjects then had 10 trials to estimate a novel distance from a new starting point. Knowledge of results was provided on the first three trials and withdrawn on the last seven trials. No differences in accuracy of production of the novel distance were noted although evidence for a stronger-schema by the group with little variability was noted on the second transfer trial. On the trials on which knowledge of results was withdrawn this stronger schema group maintained performance while the other two groups showed a decrement in performance.     

The effects of relational structure on analogical learning

Relational structure is important for various cognitive tasks, such as analogical transfer, but its role in learning of new relational concepts is poorly understood. This article reports two experiments testing people’s ability to learn new relational categories as a function of their relational structure. In Experiment 1, each stimulus consisted of 4 objects varying on 2 dimensions. Each category was defined by two binary relations between pairs of objects. The manner in which the relations were linked (i.e., by operating on shared objects) varied between subjects, producing 3 logically different conditions. In Experiment 2, each stimulus consisted of 4 objects varying on 3 dimensions. Categories were defined by three binary relations, leading to six logically different conditions. Various learning models were compared to the behavioral data, based on the theory of schema refinement. The results highlight several shortcomings of schema refinement as a model of relational learning: (1) it can make unreasonable demands on working memory, (2) it does not allow schemas to grow in complexity, and (3) it incorrectly predicts learning is insensitive to relational structure. We propose schema elaboration as an additional mechanism that provides a more complete account, and we relate this mechanism to previous proposals regarding interactions between analogy and representation construction. The current findings may advance understanding of the cognitive mechanisms involved in learning and representing relational concepts.     

Motivation of extended behaviors by anterior cingulate cortex

Intense research interest over the past decade has yielded diverse and often discrepant theories about the function of anterior cingulate cortex (ACC). In particular, a dichotomy has emerged between neuropsychological theories suggesting a primary role for ACC in motivating or 'energizing' behavior, and neuroimaging-inspired theories emphasizing its contribution to cognitive control and reinforcement learning. To reconcile these views, we propose that ACC supports the selection and maintenance of 'options' - extended, context-specific sequences of behavior directed toward particular goals - that are learned through a process of hierarchical reinforcement learning. This theory accounts for ACC activity in relation to learning and control while simultaneously explaining the effects of ACC damage as disrupting the motivational context supporting the production of goal-directed action sequences.     

Negative Transfer or Cognitive Confusion

We examined the occurrence of negative transfer in the learning of a novel motor task with regard to the schema theory (Schmidt, 1975) premise that practicing similar movements should have a facilitatory effect on learning. Given that negative transfer is often difficult to produce, an attempt was made to use a design similar to that of Lewis, McAllister, and Adarns (1951), in which the interpolated task was a complete reversal of the original task. Forty-eight university students were tested on a discrete pursuit-tracking task. They completed trials on both the original task and an interpolated reversed task before relearning the original task. The decrement between the first re- learning trial and the last original learning trial was analyzed for a negative transfer effect. Schmidt's (1982) hypothesis that negative transfer is due to temporary cognitive or decision confusion, not the result of a motor control problem, was supported. All evident performance decrements were observed only in the first relearning trial and were not present on subsequent trials. The data support a schema theory view of transfer in motor learning.     

Impulse-variability theory: implications for ballistic, multijoint motor skill performance

Impulse-variability theory (R. A. Schmidt, H. N. Zelaznik, B. Hawkins, J. S. Frank, & J. T. Quinn, 1979) accounts for the curvilinear relationship between the magnitude and resulting variability of the muscular forces that influence the success of goal-directed limb movements. The historical roots of impulse-variability theory are reviewed in the 1st part of this article, including the relationship between movement speed and spatial error. The authors then address the relevance of impulse-variability theory for the control of ballistic, multijoint skills, such as throwing, striking, and kicking. These types of skills provide a stark contrast to the relatively simple, minimal degrees of freedom movements that characterized early research. However, the inherent demand for ballistic force generation is a strong parallel between these simple laboratory tasks and multijoint motor skills. Therefore, the authors conclude by recommending experimental procedures for evaluating the adequacy of impulse variability as a theoretical model within the context of ballistic, multijoint motor skill performance.     

Contextual Interference in Movements of the Same Class: Differential Effects on Program and Parameter Learning

Contextual interference effects in motor learning usually were not found when the tasks to be learned presumably required the same generalized motor program (GMP) and differed only with regard to the movement parameters (see Lee, Wulf, & Schmidt, 1992; Magill & Hall, 1990). Thus, tasks requiring different motor programs (e.g., different relative timings) seemed to be a prerequisite for random practice to be more effective than blocked practice. However, the previous studies (that did not find random/blocked differences) used global error measures that confounded errors in relative timing and errors in absolute timing. In the present study, subjects practiced three movement patterns that had the same relative timing (requiring the same GMP) but different overall durations (requiring different parameters). Errors in relative timing and in absolute timing were assessed separately. The results indicate that random practice is more effective for the learning of relative timing (GMP learning) and less effective for the learning of absolute timing (parameter learning) than blocked practice. Preliminary ideas as to the reasons for this effect are discussed.     

Contextual interference in movements of the same class: differential effects on program and parameter learning

Contextual interference effects in motor learning usually were not found when the tasks to be learned presumably required the same generalized motor program (GMP) and differed only with regard to the movement parameters (see Lee, Wulf, & Schmidt, 1992; Magill & Hall, 1990). Thus, tasks requiring different motor programs (e.g., different relative timings) seemed to be a prerequisite for random practice to be more effective than blocked practice. However, the previous studies (that did not find random/blocked differences) used global error measures that confounded errors in relative timing and errors in absolute timing. In the present study, subjects practiced three movement patterns that had the same relative timing (requiring the same GMP) but different overall durations (requiring different parameters). Errors in relative timing and in absolute timing were assessed separately. The results indicate that random practice is more effective for the learning of relative timing (GMP learning) and less effective for the learning of absolute timing (parameter learning) than blocked practice. Preliminary ideas as to the reasons for this effect are discussed.