Abstract
Background Providing adults with knowledge of results (KR) after each practice trial (100% KR) usually is found to be detrimental to motor skill learning compared with conditions in which feedback is less frequently provided. The effect of 100% KR on children's learning is less clear, with research showing that children with cerebral palsy benefit from less frequent KR, whereas children with typical development do not.
Objective This study was designed to examine the interaction of KR frequency and task complexity on the acquisition, retention, and transfer of a novel throwing skill in fourth- and fifth-grade children with typical development.
Design This was an observational study.
Methods Children threw beanbags for accuracy at an unseen target while walking or while standing still. These 2 levels of task complexity were crossed with 2 frequencies (33% and 100%) of KR provision. Following practice, retention tests without feedback were performed 5 minutes later and then 1 week later along with transfer tests to assess the generalizability of learning.
Results Analyses revealed that learning was improved on the easy version of the task when a 33% KR frequency was provided during practice. In contrast, in the difficult version, learning was facilitated by provision of a 100% KR frequency during practice.
Conclusions Structuring practice conditions for children should take into account task complexity and feedback frequency in determining the cognitive challenge necessary for optimal skill learning. More generally, the findings suggest that practitioners teaching motor skills should design practice conditions in accordance with the cognitive processing capacity of the learner.
In many clinical settings, one of the goals of the physical therapist will be to have the patient learn or relearn a motor skill. In the treatment of pediatric, geriatric, orthopedic, and neurologic patient populations, a therapist must adapt instructional techniques and structure practice conditions to increase the likelihood that the skill being taught will be remembered long after practice has finished. Such learning interventions must enable the patient to transfer learning from the clinical setting to his or her normal living environment. To ensure this transfer, a therapist must understand how motor skill learning can be influenced by the informational feedback that is provided to the patient and how such feedback interacts with the cognitive information-processing capabilities of that patient.
In the learning of a motor skill, information regarding the effectiveness of performance in achieving the task is vital for modifying motor output in subsequent attempts. Such informational feedback can arise from intrinsic sources, such as when a patient realizes she fell backward when attempting to stand from a chair, or from extrinsic sources, such as when a therapist tells a patient that her feet were too far in front of her. Examinations of the role of extrinsic knowledge of results (KR) feedback in motor learning proliferated following the influential review of Salmoni and colleagues.1 They concluded that that when KR is provided after every trial during acquisition, it excessively guides the learner to the goal response. This guidance, although beneficial for acquisition performance, can be detrimental to skill retention. According to this guidance hypothesis, during acquisition the learner's intrinsic error detection and correction processes remain undeveloped due to the readily available error information provided by KR. Thus, the individual's information-processing activities normally used to develop memory representations are supplanted by the guidance properties of KR. If KR is presented too frequently during skill acquisition, the learner may become dependent upon this information, and performance during no-KR retention is poor. Alternatively, if KR is less readily available during acquisition, less dependency develops, and in retention the learner is able to use intrinsic error detection and correction processes to achieve and maintain a higher level of performance.
A wealth of experiments examining the effect of KR frequency in adult learners who were able-bodied have generally supported the guidance hypothesis,2–5 with recent support for the hypothesis being found in patients with Parkinson disease,6 in individuals with developmental delay,7 and in children with cerebral palsy.8 Hemayattalab and Rostami8 found that children (7 to 15 years old) with cerebral palsy learned a dart-throwing task more proficiently when feedback was provided on 50% of the trials during practice than when it was provided on every trial (100% KR). Interestingly, an earlier examination of the role of KR frequency in children with typical development, in which young adults (mean age=25.6 years) and children (mean age=10.7 years) learned a computerized pattern-matching task with either 100% or 62% KR frequency, did not demonstrate support for the guidance hypothesis, with the greatest learning being found following practice with feedback provided after every response.9
Sullivan et al9 sought to explain the differences between the findings for the children and the adults by evoking the challenge point framework.10 In this framework, the degree of motor learning depends on the level of challenge emerging from the interaction of the information-processing capacity of the learner, task constraints, and practice conditions. This framework predicts that there is an optimal challenge point in terms of cognitive effort that yields maximum skill learning. This challenge is required to activate the cognitive processes associated with such skill learning. When the challenge imposed by a practice condition exceeds the information-processing capability of the learner or the level of challenge is too low to engage sufficient processing resources, motor skill learning is reduced. Thus, Sullivan et al9 suggested that the reason lower frequencies of KR did not improve children's learning in their study was because the cognitive effort required to learn the task with less KR was beyond the optimal challenge point.
There has been little empirical investigation of the challenge point framework, although the necessity of an optimal challenge to increase performance has been proposed for surgical training11 and for the management of neck pain.12 Onla-or and Winstein13 provided some experimental evidence that motor skill learning in people with Parkinson disease was a function of task difficulty and cognitive demand of the practice environment. Participants had to learn a computerized pattern-matching task while cognitive demand was manipulated by practice schedule (random versus blocked practice) and task difficulty was varied by imposing different movement time goals.
In the study by Sullivan et al,9 conclusions regarding learning were based on the children's recall of the task after only 1 day practicing. A stronger test of motor learning would evaluate the learners' performance a week or more following the termination of practice. Furthermore, a transfer test of learning was not included in Sullivan and colleagues' work. Such a test requires learners to perform a task similar to that practiced to ascertain the generalizability of any learning acquired during practice. It also is interesting to note that Sullivan and colleagues' work did not show a difference in learning between the 2 KR conditions in the young adults, contrary to the majority of previous work in this area. Unfortunately, the more recent work by Hemayattalab and Rostami8 did not include children with typical development as control participants, nor was the effect of KR frequency on learning in adults, with or without cerebral palsy, examined. Thus, it cannot be determined whether the differences in the results between Sullivan and colleagues' work9 and that of Hemayattalab and Rostami8 were due to the neurological status of the participants or the task itself, or perhaps some combination of both. With Sullivan and colleagues' finding that children with typical development showed the greatest learning with 100% KR9 and Hemayattalab and Rostami's finding that 100% KR resulted in less learning than 50% KR in children with cerebral palsy,8 it is clear that the effect of KR frequency on children's motor skill learning warrants further investigation.
The current experiment also examined the effect of manipulating the frequency of KR in motor skill learning in children but included long-term retention and transfer tests. Like the work of Sullivan et al,9 the theoretical framework for the proposed experiment borrowed from the optimal challenge point concept of Guadagnoli and Lee.10 Sullivan et al9 manipulated the relative cognitive challenge of the task by varying the frequency of KR provided to 2 different age groups of learners, reasoning that the young adults examined in their study likely had a greater cognitive processing capability than the children. When the adults were provided with KR after every trial (100% KR), a less-than-optimal cognitive effort was required because this availability of KR made the task, in essence, too easy. Consequently, learning was less than that engendered by the more challenging condition in which less frequent KR was provided to the adult learners.
Cognitive challenge also can be manipulated by varying the difficulty of the task itself. Rather than having individuals of different ages learn a single task under different conditions of KR availability, the current experiment crossed level of task difficulty with frequency of KR while keeping the learner's age, and thus cognitive ability, constant. Thus, in Guadagnoli and Lee's terminology,10 nominal difficulty was manipulated here while functional difficulty was held constant. Nominal difficulty refers to characteristics of the skill being learned irrespective of the learner or learning environment, whereas functional difficulty is defined relative to the skill level of learner. For example, take the act of performing a transfer between 2 surfaces. The nominal difficulty of such a task would change depending on the difference in height between the 2 surfaces, whereas the functional difficulty of the transfer would depend on the neurological status of the patient.
Operationally defining nominal difficulty (task complexity) is not a simple undertaking, as there are a wide variety of cognitive and perceptuo-motor constraints underpinning any skill that could form the basis for classifying the complexity of the skill. For the current experiment, the widely accepted motor skills taxonomy developed by Gentile14,15 was used. Gentile's taxonomy14,15 seeks to classify skills on a continuum of difficulty by an analysis of the perceptuo-motor demands placed on the performer. One demand category Gentile identified is that of body transport. She proposed that a task would be more difficult when it is performed while the body is moving overground during skill production. Thus 2 skills, identical in all other aspects, will vary in nominal difficulty if one skill is performed while the person is standing still and the other is performed while the person is walking.
The current experiment sought to further explore the role of KR frequency in the motor skill learning of children by having them learn to throw beanbags at an unseen target while walking or standing still with KR either after every trial or after every third trial. We speculated that both decreasing KR frequency and having children walk while throwing would increase the cognitive challenge imposed on the learners. Unfortunately, it was not possible to predict the relative importance of these variables in increasing this challenge. Given this premise, there are a number of possible patterns of results that might emerge. If the groups receiving 100% KR (standing with 100% KR [S100], walking with 100% KR [W100]) perform similarly in retention and differently from the groups learning with 33% KR (standing with 33% KR [S33], walking with 33% KR [W33]), it would suggest that, in this experiment, KR frequency is more important than task complexity in determining cognitive challenge. If the two 33% KR groups were both more accurate in retention than the two 100% KR groups, it would imply that reducing the frequency of KR for the children increased cognitive challenge toward an optimal level. The relative level of retention performance within a given level of KR frequency also will be informative. If, for example, the S100 group outperforms the W100 group, this finding would indicate that the lowest level of cognitive challenge is optimal for learning, given the assumption that throwing from a standing position is less cognitively demanding than throwing while walking.
Alternatively, if the two 33% KR groups learn the task better than the two 100% KR groups, we could hypothesize that providing KR after every trial made the task insufficiently cognitively challenging. Such a result also would support a guidance hypothesis for the role of KR, as has been proposed for adult learners. Again, within a level of KR frequency, the comparative performance of walking and standing groups would indicate the effect of nominal task difficulty in determining cognitive challenge. If the 2 groups throwing from a standing position (S33, S100) are the most accurate in retention, it would suggest that task difficulty was more important in influencing level of learning than frequency of KR. Finally, we might envision a situation in which the W100 and S33 groups perform similarly and exhibit the greatest learning. This result would imply that the 2 variables used to manipulate cognitive challenge—KR frequency and task difficulty—have an equivalent impact on cognitive challenge. Thus, the increased cognitive challenge created by walking is mitigated by giving 100% KR, whereas the increased demands imposed by giving KR only every third trial are attenuated by allowing the children to throw from a stationary position.
Method
Participants
Forty-eight children (31 male, 17 female; mean age=10.7 years, SD=0.6) from fourth- and fifth-grade classes at a local community school volunteered to participate. Exclusion criteria for participants included a known history of a learning disability or any current musculoskeletal or neurological dysfunctions. Signed informed parental consent and child assent forms were obtained prior to participation. An institutional human participants review board approved all experimental procedures.
Task and Apparatus
The task required the children to throw 100-g cloth beanbags over a barrier onto an unseen target on the floor 6 m away. The target was a 14-cm-wide × 2.5-m-long strip that ran orthogonally to the direction of the throw. On both sides of this target zone were a series of 14-cm error zones that ran parallel to the target zone. The error zones past the target were labeled +1, +2, +3, and so on, and those zones short of the target were marked −1, −2, −3, and so on. The solid barrier extended 1.25 m above the floor and prevented the participants from viewing the target zone while throwing. A line 2 m in front of the barrier indicated the point from which the children had to throw the beanbags. During practice and retention phases, all throws were made with an overhand action with the self-selected dominant hand.
Procedure
Male and female children were quasi-randomly assigned to 1 of 4 groups to ensure that the male:female ratio was consistent across groups. Groups were differentiated on the basis of the frequency at which KR was provided (100% or 33%) and whether the children walked or stood during the throw. The 2 walking groups started 2 m away from the throwing line and walked toward the barrier, throwing the beanbag when they crossed the throwing line 2 m from the barrier. The 2 stationary groups walked up to the throwing line, but then stood still before throwing. An experimenter provided feedback indicating on which zone the beanbag landed after either every trial or after every third trial depending on group assignment. Combining KR frequency and task difficulty resulted in 4 acquisition groups: walking with 100% KR (W100), walking with 33% KR (W33), standing with 100% KR (S100), and standing with 33% KR (S33).
During the practice phase of the experiment, all participants completed 6 blocks of 12 throws with a 3-minute rest between blocks. Following completion of the acquisition phase, a 5-minute rest was given followed by the first retention test in which all participants completed one block of 12 trials without the provision of KR. A transfer test followed in which another 12 trials were completed without KR, but in this test the children were required to throw the beanbags underhand. One week later, delayed retention and transfer tests were completed in the same manner as the first tests. Children completed retention and transfer tests in the same walking or standing manner that they practiced during the acquisition phase. All children completed all of the trials in their respective conditions.
Absolute error (AE, average error without regard to sign) and variable error (VE, standard deviation about the participant's mean score) were calculated for each block of 12 trials during the acquisition, retention, and transfer phases of the experiment. These data were analyzed using analysis of variance (ANOVA) procedures.
Results
Acquisition
Absolute error data during acquisition were analyzed using a 2 × 2 × 6 (task difficulty × KR frequency × block) ANOVA with repeated measures on the last factor. The main effects of KR frequency and task difficulty were not significant (F<1), but there was a significant main effect for block (F5,280=25.9, P<.001, Fig. 1), with mean AE decreasing from 5.11 (SD=1.33) in block 1 to 3.45 (SD=1.13) in block 6. Somewhat complicating this main effect was a significant interaction of task difficulty × block (F5,280=3.1, P<.05). This interaction was caused by greater errors being produced by the groups of children who stood still to throw the beanbags during the first 2 blocks of practice, but from then on these groups performed with AE similar to that of the walking groups.
Mean absolute error as a function of acquisition, retention, and transfer. W33=walking with 33% knowledge of results (KR), S33=standing with 33% KR, W100=walking with 100% KR, S100=standing with 100% KR, Ret 1=retention test 1, Ret 2=retention test 2, Tran 1=transfer test 1, Tran 2=transfer test 2.
The variability in the children's performance during practice was examined by analyzing VE using a 2 × 2 × 6 (task difficulty × KR frequency × block) ANOVA with repeated measures on the last factor. The main effects of KR frequency (F1,56=3.2, P>.05) and task difficulty (F1,56=1.9, P>.1) were not significant, nor was the interaction of the 2 factors (F<1). The effect of block was significant (F5,280=3.4, P<.01, Fig. 2). Mean VE decreased from 4.47 (SD=1.22) at the start of practice to 3.89 (SD=1.15) at the end of practice. No other interactions were significant (P>.05).
Mean variable error as a function of acquisition, retention, and transfer. W33=walking with 33% knowledge of results (KR), S33=standing with 33% KR, W100=walking with 100% KR, S100=standing with 100% KR, Ret 1=retention test 1, Ret 2=retention test 2, Tran 1=transfer test 1, Tran 2=transfer test 2.
Retention
The throwing accuracy of the children in the immediate and delayed retention tests was analyzed using a 2 × 2 × 2 (task difficulty × KR frequency × retention test) ANOVA with repeated measures on the last factor. This analysis revealed significant main effects for task difficulty (F1,56=13.8, P<.001) and retention test (F1,56=8.5, P<.01) but not for KR frequency (F<1) (Fig. 1). The main effect for task difficulty revealed that the children in the 2 walking groups were more accurate in both retention tests (X̅=3.29, SD=0.69) than the children in the 2 groups who stood still to throw (X̅=3.95, SD=1.10), whereas the main effect for retention test indicated that children were more accurate during the immediate retention test (X̅=3.41, SD=0.76) than during the 1-week delayed retention test (X̅=3.84, SD=1.11). Complicating the main effect of task difficulty was a significant interaction of task difficulty × KR frequency (F1,56=4.9, P<.05). As Figure 3 shows, when the children threw while walking, they were more accurate after receiving a KR frequency of 33% (X̅=3.06, SD=0.66) during practice than after receiving a KR frequency of 100% (X̅=3.54, SD=0.65). In contrast, when throwing from a standing position, the children were more accurate after being given 100% KR (X̅=3.79, SD=0.96) than after 33% KR (X̅=4.10, SD=1.22).
Interaction of task difficulty and knowledge of results (KR) frequency for accuracy in retention.
The variability of throwing performance during the retention tests was subjected to a 2 × 2 × 2 (task difficulty × KR frequency × retention test) ANOVA with repeated measures on the last factor. Only the main effect of KR frequency approached traditional levels of significance (F1,56=3.9, P=.055), with the 33% KR frequency groups being more consistent in throwing (X̅=3.68, SD=1.25) than the 100% KR frequency groups (X̅=4.11, SD=1.06) (Fig. 2). No other main effects or interactions approached significance (P>.05).
Transfer
A 2 × 2 × 2 (task difficulty × KR frequency × transfer test) ANOVA of the AE data from the transfer tests revealed a pattern of results similar to that for retention, with significant main effects for task difficulty (F1,56=6.5, P<.05), transfer test (F1,56=9.3, P<.01), and an interaction of task difficulty × KR frequency (F1,56=4.1, P<.05) (Fig. 1). The main effect of task difficulty was caused by the standing groups exhibiting greater error (X̅=5.16, SD=1.34) than the walking groups (X̅=4.51, SD=1.13), and the effect of transfer test was caused by children throwing more accurately during the delayed transfer test (X̅=4.55, SD=1.35) than during the immediate transfer test (X̅=5.12, SD=1.14). Superseding the main effect of task difficulty was the significant interaction with KR frequency. Similar to the pattern for AE during retention, when the children stood still to throw the beanbags, they were more accurate following practice with 100% KR frequency (X̅=4.86, SD=1.12) than with 33% KR frequency (X̅=5.45, SD=1.48). In comparison, in the throwing-while-walking groups, greater accuracy was exhibited following 33% KR practice (X̅=4.30, SD=1.11) than following 100% KR practice (X̅=4.73, SD=1.12).
Finally, the 2 × 2 × 2 (task difficulty × KR frequency × transfer test) ANOVA on the VE data during the transfer tests showed no significant main effects or interactions for any factor (P>.1) (Fig. 2).
Discussion
Providing adults who are cognitively intact with a 100% KR frequency while practicing a motor skill usually is found to be detrimental to learning in comparison with learning environments in which such feedback is less readily available.16 In children, however, Sullivan et al9 recently reported contrary findings, with the greatest learning of a computerized pattern-matching task being demonstrated by children who received KR after every response attempt (100% KR frequency). Our current findings add to the understanding of the role of KR in motor skill learning by showing that in children the effect of KR frequency is mediated by the difficulty of the motor skill being learned. More generally, our findings emphasize the need to take into account the cognitive processing ability of the learner, regardless of whether he or she is a patient with neurological involvement or learner who is able-bodied, when planning the conditions of practice vis-à-vis task complexity and frequency of feedback.
Analysis of acquisition data indicated that all groups of children improved in both the accuracy and consistency of responses with practice. The manipulations of KR frequency and task difficulty had little differential influence on the rate at which the children improved in throwing performance. The lack of an effect for KR frequency during acquisition stands in contrast to the findings of Sullivan et al,9 who found that reduced KR frequency decreased both accuracy and consistency of performance in comparison with performance with 100% KR frequency. The fact that, in the present study, KR frequency did not significantly affect variability during practice does not support the notion that maladaptive short-term corrections are made when KR is given after every trial (100% KR). Schmidt16 and Schmidt and Bjork17 have suggested that high frequencies of KR might cause learners to continually attempt to correct their performance on every trial even though the perceptuo-motor system is unable to detect and correct the minor errors being reported by the extrinsic KR. This process of error correction potentially increases the variability of performance during acquisition and consequently leads to inferior retention performance. We found no evidence for this contention, as the 100% KR groups did not exhibit higher VE during acquisition than the 33% KR groups. In retention, there was a trend (P=.055) for the 33% KR frequency groups to perform with less variability. With an increase in the sample size, this trend might have been statistically significant. However, in order to support a maladaptive short-term corrections explanation for these results, in contrast to the findings of the current study, a significant increase in the variability of performance of the 100% KR groups in practice must be evident.
Despite the general lack of difference among the groups in performance accuracy and variability during practice, manipulations of KR frequency and task difficulty during practice did cause differences in retention performance. The pattern of retention results did not match any of our a priori predictions. The results for the stationary throwing groups (S100, S33) were similar to those reported by Sullivan et al9 in that the children who received KR after every throw performed more accurately than those who had been given KR after every third trial. However, this pattern of results was reversed for the groups of children who threw while walking (W100, W33). Here the children who had received KR at a 33% frequency exhibited the most accurate retention performance, a pattern of results that would be expected by the guidance hypothesis of Salmoni et al.1
Taken together, these unexpected results can be explained by a reevaluation of the categorization of difficulty in the throwing task. In designing the study, we borrowed from Gentile's14,15 motor skills taxonomy, reasoning that throwing while walking would be more difficult than throwing from a standing position. However, in hindsight, we believe we inadvertently made the stationary throwing condition more difficult than the walking condition. We originally had intended to have the children simply stand at the line 2 m from the barrier and throw. However, we made a last-minute change to the procedures to ensure a constant inter-trial interval between the walking and standing groups. Thus, the standing groups started at the line 2 m from the throwing line, walked up to the throwing line, and then stopped before throwing the beanbag over the barrier. We speculate that this requirement to control momentum and adjust posture before releasing the beanbag added to the complexity of the motor planning process involved in executing the throw and thus actually made the task more difficult than simply walking through the line and throwing without having to come to a stop.
If the premise is accepted that the throwing-while-walking task was less cognitively challenging than the walking-then-stopping throwing task, the pattern of results can be explained by the challenge point framework.10 In the more difficult (standing) version of task, children showed the greatest learning when they were provided with a 100% KR frequency. The increased motor planning required imposed a cognitive challenge greater than the optimal challenge point. Thus, providing the children with KR after every trial made the task less cognitively challenging, bringing the challenge back closer to the optimal challenge point. In contrast, given that throwing while walking was less cognitively taxing, and less than optimal, reducing the frequency of KR provision increased the cognitive challenge to a more optimal level than when KR was given after every trial. Providing KR at a 100% frequency decreased the cognitive challenge on this easier version of the task and thus moved the cognitive challenge below the optimal level.
Prior to the current study, the only other work examining the effect of KR frequency on motor learning in children with typical development was that of Sullivan et al.9 In their work, children showed the greatest learning following 100% KR frequency in practice. Such findings were replicated by the standing condition results but not by the walking condition results in the present work. Interpreting these findings from the challenge point framework would suggest that both the pattern-matching task of Sullivan et al9 and the standing condition in the current work presented the children with a relatively difficult cognitive challenge, necessitating KR after every trial to reduce the cognitive challenge. It is important to note that Sullivan and colleagues' study9 tested retention only on the day after practice. The current work examined skill retention immediately after finishing practice and 1 week later. Although there was a significant decline in throwing performance from the immediate retention test to the delayed retention test, the same pattern of relative differences among the groups remained. That is, there was no differential forgetting as a function acquisition condition.
The design of the current study also differed from that of Sullivan et al9 by including a transfer test in which the children attempted to hit the target with an underhand rather than an overhand throwing action. This test assesses the generalizability of the skill learning created by the various acquisition conditions. The pattern of coordination used to throw the beanbags was very different in the transfer test, and we were interested in determining whether the calibration of the target space acquired during overhand throwing could be transferred to a new coordinative pattern. Although the children were less accurate when asked to throw underhand, the same relative level of performance that was evident in the retention tests was found in the 4 groups. This finding implies that the differential learning engendered by the combination of task difficulty and KR frequency could be generalized to a new coordinative pattern.
The pattern of results for transfer does differ from that observed during retention with respect to time. In retention, all groups decreased in accuracy from the immediate retention test to the delayed test, indicating that forgetting had occurred during the 1-week interval. In contrast, in transfer the accuracy of the children's underhand throwing improved from the immediate test to the delayed test. There also appeared to be a trend for throwing consistency to improve for all groups across time (Fig. 2) but this trend was not statistically significant. The improvement in accuracy from the first transfer test to the second transfer test probably was due to the fact that the first transfer test essentially served as practice. Even though no KR was given regarding accuracy during this test, the children were able to use the previously acquired awareness of the spatial location of the target to hone their underhand throwing performance.
The present findings are some of the first to provide empirical support for the challenge point framework of Guadagnoli and Lee.10 Despite the framework's intuitive appeal, little experimental work investigating the notion has been published, perhaps because the concept of task difficulty is tricky to define, as is clearly illustrated by our attempts to manipulate difficulty in the current work. Onla-or and Winstein13 did provide some support for the framework in people with Parkinson disease, but the findings were somewhat muddied by a powerful specificity of practice effect. Clearly, there is still much to be learned about how cognitive challenge influences the learning of motor skills in adults and children with and without neurological involvement.
Potential Limitations
The current study was limited by the fact that children with neurological involvement did not participate. How such children would have compared in performance with the children with typical development we examined cannot be determined. The goal of this study, however, was not to make direct comparisons regarding the effect of KR frequency and task difficulty between children with and without neurological involvement in the task we examined. Rather, the theoretical rationale behind the design of the study was to examine how task complexity and the frequency of feedback presentation interact with cognitive processing capability in determining the relative level of motor learning engendered.
As in many motor skill–learning studies, sample size could have limited the ability to find statistical significance in certain analyses. For example, the effect of KR frequency on the variability of performance seen in retention might have been significant had additional children been tested. However, this significance would not have changed any of the important conclusions generated by this work.
Clinical Implications
From a clinical perspective, our findings highlight the important role that a determination of the cognitive status of a patient and an assessment of the functional task difficulty play in designing optimal learning environments. Recall that under Guadagnoli and Lee's optimal challenge point framework,10 nominal difficulty refers to an objective assessment of the complexity of the task regardless of who is attempting to perform the task, whereas functional difficulty attempts to quantify the subjective difficulty involved in task production and thus takes into account, among other things, the cognitive ability of the performer. A task with a certain nominal difficulty, therefore, will be functionally more difficult to a child, to an elderly individual, or to someone with neurological involvement.
Sullivan et al9 varied the level of nominal task difficulty in their study by manipulating KR frequency and used age as an index of cognitive processing capability to determine the relative functional difficulty of the task for the young adults and children in the study. The differences found as a function of KR frequency between the 2 age groups thus were explained by the different levels of functional difficulty imposed. Similarly, at the other end of the life span, cognitive processing is typically diminished in the elderly population18,19; therefore, the same level of nominal task difficulty will present greater functional difficulty to a geriatric population than to young adult learners. Likewise, individuals with certain brain lesions might be expected to have cognitive processing deficits, and therapists must take into account such deficits in structuring the learning environment. Our findings strongly suggest that seemingly small changes in task difficulty can interact with the frequency of therapist-provided feedback in determining the amount of skill carryover from the clinic to the patient's everyday environment. To optimize motor learning in the clinical environment, therapists should manipulate nominal difficulty by varying task parameters, such as feedback frequency, in order to generate an appropriate level of functional difficulty for the patient. If functional task difficulty is too high or too low, skill learning will not be maximized.
Footnotes
-
All authors provided concept/idea/research design and data collection and analysis. Dr Sidaway provided writing, project management, participants, facilities/equipment, and institutional liaisons.
-
The authors thank the children of Dedham School in Maine, who enthusiastically participated in the study, and their parents, who allowed their children to take part in the study. They extend their gratitude to the faculty at Dedham School, especially to Kathy Lawson (principal) and to Tim Pearson, who allowed the authors to interrupt his physical education class on numerous occasions to collect data.
-
This study was approved by the Husson University's Institutional Review Board.
-
This work was presented at the North American Society for the Psychology of Sport and Physical Activity (NASPSPA) Conference; June 7–9, 2012; Honolulu, Hawaii.
- Received November 4, 2011.
- Accepted March 8, 2012.
- © 2012 American Physical Therapy Association
References
Issue highlights
