This article was downloaded by: [Memorial University of Newfoundland] On: 07 October 2014, At: 17:25 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Journal of Motor Behavior Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/vjmb20
Effects of Practice Variability on Unimanual Arm Rotation a
Eric G. James & Phillip Conatser
a
a
Department of Health and Human Performance, University of Texas at Brownsville Published online: 20 Mar 2014.
To cite this article: Eric G. James & Phillip Conatser (2014) Effects of Practice Variability on Unimanual Arm Rotation, Journal of Motor Behavior, 46:4, 203-210, DOI: 10.1080/00222895.2014.881314 To link to this article: http://dx.doi.org/10.1080/00222895.2014.881314
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Journal of Motor Behavior, Vol. 46, No. 4, 2014 C Taylor & Francis Group, LLC Copyright
RESEARCH ARTICLE
Effects of Practice Variability on Unimanual Arm Rotation Eric G. James, Phillip Conatser
Downloaded by [Memorial University of Newfoundland] at 17:25 07 October 2014
Department of Health and Human Performance, University of Texas at Brownsville.
(Schmidt, 1975) the parameters varied during practice have typically been confined to the scaling of invariant movement parameters as specified by Schema theory. Recent theoretical work (Sch¨ollhorn, Mayer-Kress, Newell, & Michelbrink, 2009; Sch¨ollhorn, Beckmann, Michelbrink, Sechelmann, Trockel, & Davids, 2006) has hypothesized that the rate of motor learning is a function of the magnitude and structure of practice variability. Differential learning theory hypothesizes that the optimal level of practice variability is higher than that hypothesized by prior motor learning theories (Adams, 1971; Gentile, 1972; Schmidt, 1975). In differential training high levels of variability are introduced into practice with respect to both the task goal and the execution redundancy (Frank, Michelbrink, Beckmann, & Sch¨ollhorn, 2008) and by randomizing the order (i.e., the structure) of practice trials (as in random practice; Lee & Genovese, 1988). Differential learning has been modeled (Frank et al., 2008) with the fourth-order polynomial:
ABSTRACT. High variability practice has been found to lead to a higher rate of motor learning than low variability practice in sports tasks. The authors compared the effects of low and high levels of practice variability on a simple unimanual arm rotation task. Participants performed rhythmic unimanual internal-external arm rotation as smoothly as possible before and after 2 weeks of low (LV) or high (HV) variability practice and after a 2-week retention interval. Compared to the pretest, the HV group significantly decreased hand, radioulnar, and shoulder rotation jerk on the retention test and shoulder jerk on the posttest. After training the LV group had lower radioulnar and shoulder jerk on the posttest but not the retention test. The results supported the hypothesis that high variability practice would lead to greater learning and reminiscence than low variability practice and the theoretical prediction of a bifurcation in the motor learning dynamics. Keywords: motor learning, practice, unimanual, variability
uch of the research comparing the effects of different levels of practice variability has examined predictions of Schema theory (Schmidt, 1975). Schema theory hypothesizes that the practice of the scaling of a generalized motor program (termed variable practice) enhances motor learning. In this theory practice variability should not be implemented by altering invariant characteristics of a motor program (e.g., relative timing, relative force). Research has found that variability in the sequencing of practice trials (random practice) can enhance the retention of motor learning compared to a blocked sequencing of practice trials (Lee & Genovese, 1988; Lee, Magill, & Weeks, 1985; Lintern, 1988; Magill, 1988; Newell, Antoniou, & Carlton, 1988). Practice variability can be introduced with respect to the task goal and the execution redundancy. The traditional approach to studying practice variability has been to introduce variability at the level of the task goal. Research including experiments that introduced variability in both the task goal and execution redundancy (Ranganathan & Newell, 2013) found that the benefit of constant or variable task goal practice was specific to each testing condition. Also, in this study the practice of redundant task solutions (movement trajectory) did not improve task performance. In prior research on motor learning the magnitude of practice variability has typically been relatively limited, both with respect to the number of practice variations and the parameters that are varied during practice. Typical motor learning studies have used a small number of variations of the target task (Lee et al., 1985; Shapiro & Schmidt, 1982). The effects of practicing these variations have also been studied when practiced in a randomized order versus in a blocked design (Shea & Morgan, 1979). As many studies of variable and random practice have used a motor programming approach
M
Q − Qc V (m) = − 2
c d m 2 − m3 + m4 3 4
(1)
where V(m) denotes the potential function of the learning rate m, Q is the level of practice variability, Qc is a critical threshold of practice variability, and c and d are positive constants. As shown in Figure 1A, when the level of practice variability Q is less than the critical threshold Qc the system is bistable, with stable attractor states for zero (m = 0) and positive (m > 0) learning rates. When practice variability exceeds the critical threshold (Q > Qc ) the system is monostable, with a single stable attractor state with a positive learning rate (m > 0; see Figure 1B). The Frank et al. (2008) model captures the occurrence of reminiscence. Reminiscence is the occurrence of learning, or an increase in performance, during a rest period (Eysenck, 1962, 1965; Hovland, 1951; Hull, 1943). Reminiscence is not fully understood, but has been attributed as being due to factors such as practice schedules, fatigue, motivation, consolidation, reactive inhibition (Ammons, 1988; Christina & Shea, 1988; Etnyre & Poindexter, 1995; Eysenck, 1965; Horn, 1975; Hovland, 1951; Lee & Genovese, 1988; Lintern, 1988; Magill, 1988), and the personality traits of extraversion and neuroticism (Eysenck, 1962, 1965; Horn, 1975). Correspondence address: Eric G. James, Department of Health and Human Performance, The University of Texas at Brownsville, REK Building 2.638, Brownsville, TX 78520, USA. e-mail: [email protected]
203
Downloaded by [Memorial University of Newfoundland] at 17:25 07 October 2014
E. G. James & P. Conatser
FIGURE 1. The fourth-order polynomial model (Equation 1) of differential learning with constant parameters c = 2.9 and d = 1. V is the order parameter potential and m is the order parameter. Panel A depicts the state of Q – Qc = –2. Prior to differential training the neuromotor system resides in the basin of attraction at a learning rate of m = 0 and after differential training at a learning rate of m = 1.77. Panel B depicts the state of Q – Qc = 0. This is the state of the system during differential training with a positive learning rate of m = 3. Adapted from Frank, Michelbrink, Beckmann, and Sch¨ollhorn (2008).
In the Frank et al. (2008) model reminiscence can occur when the level of practice variability increases from Q < Qc to Q > Qc and back again. When the level of practice variability is scaled up and down in this way the system moves from a stable attractor state of zero learning (m = 0; the solid circle in Figure 1A) to a stable state of learning (m > 0; the solid circle in Figure 1B) and remains in the stable attractor with a positive learning rate after the level of practice variability has decreased to Q < Qc (e.g., after practice has ended; the open circle in Figure 1A). The result is continued learning during a retention period (i.e., reminiscence). Reminiscence has previously been found to occur with high variability (HV) training (Beckmann & Sch¨ollhorn, 2003; Frank et al., 2008; Humpert, 2004; Humpert & Sch¨ollhorn, 2006; R¨omer, Sch¨ollhorn, & Jaitner, 2003). In differential training a high level of practice variability is created by giving learners multiple practice instructions in a randomized order, typically with only one trial performed for each instruction. The high number of task variations and short amount of time spent practicing each variation distinguishes differential training from typical random practice protocols, in which only a few task variations are practiced numerous times in a randomized order (Brady, 1998). Differential training instructions vary both the task goal and the execution redundancy characteristics such as the initial and/or ending conditions of a movement, the relative and/or 204
absolute duration of movement, joint angles, velocities, and acceleration profiles (Sch¨ollhorn et al., 2006). The Differential Learning Theory hypothesis that parameters such as relative timing and relative force should be varied during practice is contrary to Schema Theory (Schmidt, 1975). Prior research has shown HV (differential) practice to produce greater learning than traditional practice in soccer (Sch¨ollhorn et al., 2006), high jump (Sch¨ollhorn, Michelbrink, Welminsiki, & Davids, 2009), handball throwing (Wagner & M¨uller, 2008), shot put (Beckmann & Sch¨ollhorn, 2003), tennis (Humpert, 2004; Humpert & Sch¨ollhorn, 2006), and volleyball (R¨omer et al., 2003). Research on HV practice has typically examined the rate of motor learning in complex sports skills. Apart from the study of sports skills, one study found decreased postural sway after HV postural training (James, 2014). In the present study we compared the effects of low and high levels of practice variability on a simple rhythmic unimanual motor task. We hypothesized that HV practice would lead to greater motor learning than low variability (LV) practice. Method Participants Healthy young adult participants (N = 27; 12 men, 15 women) with a mean age of 23.93 years (SD = 3.79 years) Journal of Motor Behavior
Practice Variability and Arm Rotation
TABLE 1. The Arm Orientations and Training Movements Performed During High Variability Training Arm orientation Horizontal in front of the shoulder Horizontal directly to the right at shoulder height Vertical above the head
Downloaded by [Memorial University of Newfoundland] at 17:25 07 October 2014
FIGURE 2. Depiction of the experimental device used during the rhythmic arm internal-external rotation movements.
volunteered for this study. All participants signed an informed consent form that was approved by the local Institutional Review Board. All participants identified themselves as right-handed and as having had no prior surgery or injury to the arm, wrist, or shoulder. Task and Procedure Participants were seated comfortably and grasped the handle of a custom-made device (see Figure 2) with their right hand. The participant’s right arm was positioned horizontally at shoulder height in front of the body and participants were instructed to maintain the elbow in the extended position at all times during motor performance. The experimental task consisted of rhythmically rotating the experimental handle by internally and externally rotating the right arm (internal and external rotation of the shoulder and radioulnar joints) through a range of approximately 180◦ . The participants were given 2 min to familiarize themselves with the task and apparatus. The participants were instructed to perform the rotational movements as smoothly as possible and to synchronize the points of maximum internal and external rotation with auditory metronome tones that paced movement frequency in separate trials at 1 and 2 Hz. One trial was performed at each movement frequency on each test. Movement trials were of 30 s duration with 1 min of rest provided between trials. A pretest was performed, as well as a posttest after two weeks of practice and a twoweek retention test after practice had ended. Two Biometrics (Ladysmith, VA) ACL300 3-D accelerometers were attached to the lateral surface of the wrist and of the upper arm midway between the elbow and shoulder. DataLINK (Biometrics, Ladysmith, VA) software was used to collect acceleration data at a sampling rate of 1000 Hz during movement trials. The participants were randomly assigned to a low (LV; n = 14) or high (HV; n = 13) variability training group. During training the LV group repeated the arm rotation movements, with the instruction to make the movements as smooth and continuous as possible. During training the HV group performed arm rotation movements 2014, Vol. 46, No. 4
Training movement In-phase shoulder-radioulnar rotation Antiphase shoulder-radioulnar rotation Shoulder protraction-retraction with trunk fixed Shoulder protraction-retraction with arm fixed Shoulder flexion-extension Shoulder abduction-adduction Arm circumduction
with multiple instructions, with one trial performed with each instruction. For the HV training group the movement training included the use of three arm orientations and seven practice movements (see Table 1). The movements performed on each trial during training were randomly selected from a list that consisted of the seven individual movements and all possible (nonexclusive) pairings of these movements (e.g., in-phase shoulder-radioulnar rotation and shoulder protraction-retraction with the trunk fixed). One of the three arm orientations was randomly selected as the orientation in which to perform the selected movement. The participants were given instructions to perform each randomly selected movement in the selected arm orientation. When performing movement in the first two arm orientations listed in Table 1 the experimental device depicted in Figure 2 was used. When performing movements in the third arm orientation (vertical orientation) the arm was held above the head and the experimental device was not used. The in-phase mode was practiced by moving the upper arm and hand (shoulder and radioulnar joints) in the same direction. This movement has been shown to typically involve an in-phase coordination pattern (James, 2012). The antiphase coordination mode was practiced by maintaining the hand in a fixed position while internally and externally rotating the upper arm. While this movement initially appeared to be impossible to some participants they were readily able to perform it after a short amount of practice. Shoulder protraction and retraction were practiced with the trunk fixed (and the extended arm moving) and with the extended arm fixed (and the trunk moving). Shoulder flexion-extension, shoulder abduction-adduction, and arm circumduction were also practiced. Consistent with prior research on differential training participants in both groups performed the practice movements at a self-paced movement frequency and used a self-selected range of movement (Beckmann, Winkel, & Sch¨ollhorn, 2008; James, 2014; Sch¨ollhorn et al., 2006; Sch¨ollhorn et al., 2010). 205
Downloaded by [Memorial University of Newfoundland] at 17:25 07 October 2014
E. G. James & P. Conatser
No feedback regarding task performance was given to participants during data collection or during training. For both theoretical and methodological reasons no corrective feedback is typically provided during differential training. The theoretical objective of differential training instructions is to perturb the neuromotor system out of a habitual behavioral attractor, rather than to require the production of a specific prescribed movement pattern (Frank et al., 2008; James, 2014; Sch¨ollhorn et al., 2006; Sch¨ollhorn et al., 2010). In practical terms, this method of performing only a single trial with each movement instruction limits the possible use of corrective feedback during training. To maintain consistent feedback conditions across groups the LV training group also did not receive feedback during training. Each group practiced 2 times per week for two weeks and performed 20 practice trials of 1 min duration with 1 min of rest between trials on each day of practice. The arm rotation testing protocol pretest was performed within 48 hr before practice sessions began. The posttest was performed 24 hr after the last practice session and the retention test was performed two weeks after the posttest. Data Analysis Tangential acceleration data collected from the accelerometers were filtered with a ninth-order 30 Hz low-pass Butterworth filter. Hand and shoulder movement jerk were calculated from the accelerometry data. The jerk for the radioulnar joint was calculated as the difference between the hand and shoulder movement jerk data. The root mean square jerk (RMSJ) of the hand was calculated as an index of task performance and the jerk for the shoulder and radioulnar joints was calculated to examine the level of jerk at each effector used in this task. Jerk was calculated as Jerkt =
ACCLt − ACCLt−1 dt
where t is each data point in the time series, ACCL is the acceleration signal, and dt is the derivative of time. The RMSJ was calculated as RMSJ =
(Jerkt − M)2 /n
where t is each point in the time series from 1 to n (the number of data points) and M is the mean jerk. Three 2 Group × 2 Frequency × 3 Test repeated measures analysis of variance were used to analyze the RMSJ of the radioulnar and shoulder joints and of the hand. The calculation of all dependent variables was performed with coded MATLAB (The MathWorks, Natick, MA) programs. Inferential statistical analyses were performed using the SPSS software package (version 19.0) with a type-I error of .05 used to determine statistical significance. Fisher’s least significant difference test was used for post hoc analysis. 206
Results Hand Root Mean Square Jerk The RMSJ at the 1 Hz movement frequency was significantly lower than at 2 Hz, F(1, 25) = 190.393, p < .001. There was a significant group by test interaction, F(2, 50) = 4.740, p = .016 (see Figure 3C). In post hoc analysis for the HV group the retention test RMSJ was significantly lower than the pretest (p = .038) and the posttest (p = .001). There was no significant difference between the pre- and posttests (p = .528). For the LV group the posttest RMSJ was significantly lower than the retention test (p = .046). There were no significant differences between the preand posttests (p = .109) or the pre- and retention tests (p = .822). Also, on the retention test the HV group RMSJ was significantly lower than the LV group (p = .035) while there was no significant difference between groups on the pretest (p = .882) or posttest (p = .059). There was a significant group by frequency by test interaction, F(2, 50) = 3.459, p = .039 (see Figure 3). For the HV group at the 1 Hz movement frequency there were no significant differences between the pre- and posttests (p = .563), the pre- and retention tests (p = .631) or the post- and retention tests (p = .198). For the HV group at the 2 Hz movement frequency the retention test RMSJ was significantly lower than the pretest (p = .023) and posttest (p < .001). There was no significant difference between the pre- and posttests (p = .570). For the LV group at the 1 Hz movement frequency there were no significant differences between the pre- and posttests (p = .106), the pre- and retention tests (p = .728) or the postand retention tests (p = .107). At the 2 Hz movement frequency there were no significant differences between the pre- and posttests (p = .178), the pre- and retention tests (p = .911) or the post- and retention tests (p = .111). Also, at the 1 Hz movement frequency there were no significant differences between groups on the pretest (p = .572), posttest (p = .131), or retention test (p = .347). At the 2 Hz movement frequency the HV group RMSJ was significantly lower than LV group on the retention test (p = .018), while there was no significant difference between groups on the pretest (p = .889) or posttest (p = .069). The main effects for group, F(1, 25) = 0.001, p = .973, and test, F(2, 50) = 1.402, p = .256, were not significant. The group by frequency, F(1, 25) = 0.005, p = .942, and frequency by test, F(2, 50) = 2.411, p = .100, interactions were not significant. Radioulnar Root Mean Square Jerk The radioulnar RMSJ was significantly lower at the 1 Hz movement frequency than at the 2 Hz frequency, F(1, 25) = 156.040, p < .001. There was also a significant group by test interaction, F(2, 50) = 5.868, p = .006 (see Figure 4A). For the HV group the retention test RMSJ was significantly lower than on the pretest (p = .041) and posttest (p = .001). There was no significant difference between the pre- and posttests (p = .580). For the LV group the posttest RMSJ was significantly lower than the pretest (p = .036) and retention Journal of Motor Behavior
Downloaded by [Memorial University of Newfoundland] at 17:25 07 October 2014
Practice Variability and Arm Rotation
FIGURE 3. Representation of hand root mean square jerk data as a function of: group, test, and frequency. A = 1 Hz movement frequency; B = 2 Hz movement frequency; C = group and test. HV = High practice variability group. LV = Low variability practice group. Standard errors are represented with error bars. Asterisk indicates significant differences.
test (p = .034). There was no significant difference between the pre- and retention tests (p = .629). Also, the HV group RMSJ was significantly higher than the LV group on the posttest (p = .050), and significantly lower than the LV group on the retention test (p = .047). There was no significant difference between groups on the pretest (p = .914). The main effects for group, F(1, 25) = 0.012, p = .914, and test, F(2, 50) = 2.321, p = .112, were not significant. The group by frequency, F(1, 25) = 0.016, p = .899; frequency by test, F(2, 50) = 2.498, p = .092; and group by frequency by test, F(2, 50) = 2.809, p = .070, interactions were not significant. Shoulder RMSJ The RMSJ at the 1 Hz movement frequency was significantly lower than at the 2 Hz movement frequency, F(1, 25) = 84.587, p < .001. There was also a significant main effect for test, F(2, 50) = 8.041, p = .001. In post hoc analysis the pretest RMSJ was significantly higher than on the posttest 2014, Vol. 46, No. 4
(p < .001) or retention test (p = .004). There was no significant difference between the post- and retention tests (p = .893). There was a significant group by test interaction, F(2, 50) = 5.262, p = 008 (see Figure 4B). In post hoc analysis for the HV group the pretest RMSJ was significantly higher than the posttest (p = .028) and retention test (p = .001) while there was no significant difference between the post- and retention tests (p = .066). For the LV group the posttest RMSJ was significantly lower than the pre- (p = .001) and retention tests (p = .038). There was no significant difference between the pre- and posttests (p = .485). There was a significant frequency by test interaction, F(2, 50) = 5.081, p = .010. In post hoc analysis at the 1 Hz movement frequency the pretest RMSJ was significantly higher than on the posttest (p = .01) while there was no significant difference between the retention test and the pre- (p = .397) or posttests (p = .270). At the 2 Hz movement frequency the pretest RMSJ was significantly higher than on the
207
Downloaded by [Memorial University of Newfoundland] at 17:25 07 October 2014
E. G. James & P. Conatser
FIGURE 4. Representation of (A) radioulnar and (B) shoulder root mean square jerk data as a function of group and test. HV = high variability practice group; LV = low variability practice group. Standard errors are represented with error bars. Asterisk indicates significant differences.
post- (p < .001) or retention tests (p = .001), while there was no significant difference between the post- and retention tests (p = .632). The main effect for group was not significant, F(1, 25) = 0.139, p = .712. The group by frequency, F(1, 25) = 0.211, p = .650, and group by frequency by test, F(2, 50) = 1.963, p = .152, interactions were not significant. Discussion Prior research has shown that HV practice can lead to greater learning and retention on sports skills than lower variability practice (Beckmann & Sch¨ollhorn, 2003; Frank et al., 2008; Humpert, 2004; Humpert & Sch¨ollhorn, 2006; R¨omer et al., 2003). In the present study HV practice led to significant task improvement on the retention test. LV practice did not lead to a significant improvement on the task criterion (hand RMSJ). These findings supported the hypothesis that HV practice would provide greater learning and retention than LV practice of a simple movement task. 208
This indicated that HV training can provide greater learning and retention not only in sports skills but also in a simple motor task such as unimanual arm rotation. The LV practice led to a decrease in the jerk of individual effectors (radioulnar and shoulder joints) on the posttest. This form of training led to a small but not statistically significant improvement on the task criterion (hand rotation jerk). The decrease in the jerk of the two effectors was temporary and was not maintained on the retention test. While the LV training led to a nonsignificant decrease in hand jerk and a lack of retention in radioulnar and shoulder rotation jerk it may well be that with a greater number of subjects (or more training time) significant learning and retention may have occurred. The present results indicated greater learning and retention with HV training. The HV practice led to decreased shoulder jerk on the posttest that was maintained on the retention test. The HV practice also led to a delayed improvement in radioulnar joint jerk that occurred on the retention test. The improvement on the task criterion at the 2 Hz, but not the 1 Hz, movement frequency may have been due to a floor effect. The 1 Hz movement frequency may have been sufficiently easy that performance was already at a high level (i.e., the RMSJ was already low) on the pretest and so did not improve significantly on the post- and retention tests. The results supported the theoretical prediction of reminiscence (Frank et al., 2008) as significant learning took place during the retention period for the HV practice group. This effect was theoretically caused by the neuromotor system remaining within a stable attractor with a learning rate m > 0 (the open circle in Figure 1A) rather than returning to the stable attractor state with a learning rate of m = 0 where the system was located prior to training (the solid circle in Figure 1A). Improvement on the task criterion did not occur until the retention test. This indicated that the theoretical bifurcation from bistability to monostability did not occur until after the posttest. The finding of reminiscence after differential training is consistent with previous theoretical and experimental work on differential learning (Beckmann & Sch¨ollhorn, 2003; Frank et al., 2008; Humpert, 2004; Humpert & Sch¨ollhorn, 2006; R¨omer et al., 2003). The repetitive practice failed to lead to significant improvement on the task criterion. This indicated that the LV practice did not lead to learning (i.e., a relatively permanent improvement in performance; Guthrie, 1952). Theories of reminiscence have hypothesized that reminiscence is due to reactive or conditioned inhibition (Hull, 1943; Kimble, 1949) and/or consolidation (Eysenck, 1965; Eysenck & Frith, 1977). Massed practice can lead to shortterm reactive inhibition, due to mental (Walker, 1958) or physical work (Hull, 1943) that depresses the level of performance during practice. This effect diminishes with rest, leading to the occurrence of reminiscence. Several types of consolidation exist (e.g., synaptic, sleep, systems) that can also lead to greater performance after a retention period. Journal of Motor Behavior
Downloaded by [Memorial University of Newfoundland] at 17:25 07 October 2014
Practice Variability and Arm Rotation
The finding of decreased shoulder, radioulnar and hand jerk (though the latter was a nonsignificant trend) on the posttest for the LV training group suggests that this form of training did not lead to reactive inhibition on the posttest (Eysenck, 1965). It is possible that the absence of a posttest improvement on the task criterion for the HV training group was due to reactive inhibition. It could be the case that the HV training leads to greater reactive inhibition than LV training. If this is the case the reminiscence effect found on the retention test after HV training could have been due, in part, to a depression of performance on the posttest caused by reactive inhibition. However, reactive inhibition has generally been found to occur after massed practice (Eysenck, 1965), which was not performed in the present experiment. Also, the posttest was given after reactive inhibition should have ended (24 hr after the last day of practice). These facts do not support the interpretation of the lack of posttest improvement after the HV training as due to reactive inhibition. According to prior theoretical work (Eysenck, 1965; Eysenck & Frith, 1977) if the HV training did not lead to reactive inhibition the reminiscence effect that occurred after this form of training would be due to the occurrence of consolidation during the retention period. Synaptic consolidation typically occurs within a period of hours (Dudai, 2004), sleep consolidation within a few days (Walker et al., 2003), and systems consolidation over weeks or years (Squire & Alvarez, 1995). Based on these approximate time scales the reminiscence that occurred after HV training in the present study may have been due to sleep and/or systems consolidation. The present and prior studies of differential learning (Beckmann et al., 2008; Frank et al., 2008) have examined performance after retention periods of two or more weeks. To determine the possible types of consolidation processes that may occur after differential training future research is needed that includes multiple retention tests over a range of time intervals. It is also possible that the reminiscence that occurs after HV training may be due to a process or processes other than inhibition and consolidation. Existing modeling of differential learning (Frank et al., 2008) captures learning dynamics that differ from traditional theories of motor learning and reminiscence. In the Frank et al. (2008) model of differential learning a stable attractor state for a positive rate of motor learning exists before practice has begun. The neuromotor system can remain in this stable state after practice has ended, leading to reminiscence. The existence of a stable state for a positive rate of learning prior to the onset of practice is inconsistent with prior theories of reminiscence. In prior theories reminiscence is due to the dissipation of inhibition (Hull, 1943; Kimble, 1949) or consolidation of memory (Eysenck, 1965; Eysenck & Frith, 1977) that has accrued during training. Neither of these can theoretically exist before training has begun, as does the stable state for a positive rate of learning in the Frank et al. (2008) model. Future research is needed to test the predictions of traditional 2014, Vol. 46, No. 4
theories of reminiscence vs. the Frank et al. model and to determine if the reminiscence that can occur after HV training differs from the reminiscence that follows more repetitive forms of practice. REFERENCES Adams, J. A. (1971). A closed-loop theory of motor learning. Journal of Motor Behavior, 2, 111–149. Ammons, R. B. (1988). Distribution of practice in motor skill acquisition: A few questions. Research Quarterly for Exercise and Sport, 59, 288–290. Beckmann, H., & Sch¨ollhorn, W. I. (2003). Differential Learning in shot put. In W. I. Sch¨ollhorn, C. Bohn, J. M. J¨ager, H. Schaper, & M. Alichmann (Eds.), European workshop on movement science (pp. 68–69). K¨oln, Germany: Sport&Buch Strauβ. Beckmann, H., Winkel, C., & Sch¨ollhorn, W. I. (2008). Optimal range of variation in hockey technique training. Paper presented at the 2nd International Congress of Complex Systems in Sport, 10th European Workshop of Ecological Psychology, Madeira, Portugal. Brady, F. (1998). A theoretical and empirical review of the contextual interference effect and the learning of motor skills. Quest, 50, 266–293. Christina, R. W., & Shea, J. B. (1988). The limitations of generalization based on restricted information. Research Quarterly for Exercise and Sport, 59, 291–297. Dudai, Y. (2004). The neurobiology of consolidations, or, how stable is the engram? Annual Review of Psychology, 55, 51–86. Etnyre, B. R., & Poindexter, H. B. (1995). Characteristics of motor performance, learning, warm-up decrement, and reminiscence during a balancing task. Perceptual and Motor Skills, 80, 1027–1030. Eysenck, H. J. (1962). Reminiscence, drive and personality – revision and extension of a theory. British Journal of Social and Clinical Psychology, 1, 127–140. Eysenck, H. J. (1965). A three-factor theory of reminiscence. British Journal of Psychology, 56, 163–181. Eysenck, H. J., & Frith, C. D. (1977). Reminiscence, motivation, and personality. London, England: Plenum. Frank, T. D., Michelbrink, M., Beckmann, H., & Sch¨ollhorn, W. I. (2008). A quantitative dynamical systems approach to differential learning: Self-organization principle and order parameter equations. Biological Cybernetics, 98, 19–31. Gentile, A. M. (1972). A working model of skill acquisition with application to teaching. Quest, 17, 3–23. Guthrie, E. R. (1952). The psychology of learning. New York, NY: Harper & Row. Horn, P. W. (1975). Evidence for the generality of reminiscence as a function of extraversion and neuroticism. Journal of Psychology, 90, 41–44. Hovland, C. I. (1951). Human learning and retention. In S. S. Stevens (Ed.), Handbook of experimental psychology (pp. 613–689). New York, NY: Wiley. Hull, C. L. (1943). Principles of behavior. New York, NY: Appleton-Century-Crofts. Humpert, V. (2004). Vergleichende Analyse von Techniktrainingans¨atzen zum Tennisaufschlag [Comparative analysis of tennis serve training]. M. S. thesis. M¨unster, Germany: University of M¨unster. Humpert, V., & Sch¨ollhorn, W. I. (2006). Vergleich von Techniktrainingans¨antzen zum Tennisaufschlag [Comparison of tennis technique training]. In A. Ferrouti & H. Remmert (Eds.), Trainingswissenschaften im Freizeitsport – Symposium der dvsSektion Trainingswissenschaften vommm 7: 9 April 2005; Hamburg [Training sciences in recreational sports: symposium section 209
Downloaded by [Memorial University of Newfoundland] at 17:25 07 October 2014
E. G. James & P. Conatser training sciences VOMMM] (pp. 121–124). Hamburg, Germany: Bochum Czwalina. James, E. G. (2012). Fixed-point drift and hysteresis in frequencyscaled unimanual coordination. Journal of Motor Behavior, 44, 281–288. James, E. G. (2014). Short-term differential training decreases postural sway. Gait & Posture, 39, 172–176. Kimble, G. A. (1949). An experimental test of a two-factor theory of inhibition. Journal of Experimental Psychology, 39, 15–23. Lee, T. D., & Genovese, E. D. (1988). Distribution of practice in motor skill acquisition: learning and performance effects reconsidered. Research Quarterly for Exercise and Sport, 59, 277–287. Lee, T. D., Magill, R. A., & Weeks, D. J. (1985). Influence of practice schedule on testing schema theory predictions in adults. Journal of Motor Behavior, 17, 283–299. Lintern, G. (1988). Distributed practice: are there useful insights for application or theory? Research Quarterly for Exercise and Sport, 59, 298–302. Magill, R. A. (1988). The many faces of practice distribution in motor learning. Research Quarterly for Exercise and Sport, 59, 303–307. Newell, K. M., Antoniou, A., & Carlton, L. G. (1988). Massed and distributed practice effects: Phenomena in search of a theory. Research Quarterly of Exercise and Sport, 59, 308–313. Ranganathan, R., & Newell, K. M. (2013). Changing up the routine: intervention-induced variability in motor learning. Exercise and Sport Sciences Reviews, 41, 64–70. R¨omer, J., Sch¨ollhorn, W. I., & Jaitner, T. (2003). Differentielles lernen bei der Aufschlagannahme im Volleyball. In J. Krug & T. M¨uller (Eds.), Messpl¨atze, Messtraining, Motorisches Lernen [Measuring stations, measuring station training, motor learning] (pp. 129–133). Sankt Augustin, Germany: Academia Verlag. Schmidt, R. A. (1975). A schema theory of discrete motor skill learning. Psychological Review, 82, 225–260. Sch¨ollhorn, W. I., Beckmann, H., & Davids, K. (2010). Exploiting system fluctuations. Differential training in physical prevention and rehabilitation programs for health and exercise. Medicina, 46, 365–373.
210
Sch¨ollhorn, W. I., Beckmann, H., Michelbrink, M., Sechelmann, M., Trockel, M., & Davids, K. (2006). Does noise provide a basis for the unification of motor learning theories? International Journal of Sport Psychology, 37, 186–206. Sch¨ollhorn, W. I., Mayer-Kress, G., Newell, K. M., & Michelbrink, M. (2009). Time scales of adaptive behavior and motor learning in the presence of stochastic perturbations. Human Movement Science, 28, 319–333. Sch¨ollhorn, W., Michelbrink, M., Welminsiki, D., & Davids, K. (2009). Increasing stochastic perturbations enhances acquisition and learning of complex sport movements. In D. Araujo, H. Ripoll, & M. Raab (Eds.), Perspectives on cognition and action in sport (pp. 59–73). Hauppage, NY: Nova Science. Shapiro, D. C., Schmidt, R. A. (1982). The schema theory: Recent evidence and developmental implications. In J. A. S. Kelso & J. E. Clark (Eds.), The development of movement control and co-ordination (pp. 113–150). New York, NY: Wiley. Shea, J. B., & Morgan, R. L. (1979). Contextual interference effects on the acquisition, retention, and transfer of a motor skill. Journal of Experimental Psychology: Human Learning and Memory, 5, 179–187. Squire, L. R., & Alvarez, P. (1995). Retrograde amnesia and memory consolidation: A neurobiological perspective. Current Opinion in Neurobiology, 5, 169–177. Wagner, H., & M¨uller, E. (2008). The effects of differential and variable training on the quality of parameters of a handball throw. Sports Biomechanics, 7, 54–71. Walker, E. L. (1958). Action decrement and its relation to learning. Psychological Review, 65, 129–142. Walker, M. P., Brakefield, T., Seidman, J., Morgan, A., Hobson, J. A., & Stickgold, R. (2003). Sleep and the time course of motor skill learning. Learning & Memory, 10, 275–284.
Received August 23, 2013 Revised November 25, 2013 Accepted January 3, 2014
Journal of Motor Behavior
Journal of Motor Behavior Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/vjmb20
Effects of Practice Variability on Unimanual Arm Rotation a
Eric G. James & Phillip Conatser
a
a
Department of Health and Human Performance, University of Texas at Brownsville Published online: 20 Mar 2014.
To cite this article: Eric G. James & Phillip Conatser (2014) Effects of Practice Variability on Unimanual Arm Rotation, Journal of Motor Behavior, 46:4, 203-210, DOI: 10.1080/00222895.2014.881314 To link to this article: http://dx.doi.org/10.1080/00222895.2014.881314
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Journal of Motor Behavior, Vol. 46, No. 4, 2014 C Taylor & Francis Group, LLC Copyright
RESEARCH ARTICLE
Effects of Practice Variability on Unimanual Arm Rotation Eric G. James, Phillip Conatser
Downloaded by [Memorial University of Newfoundland] at 17:25 07 October 2014
Department of Health and Human Performance, University of Texas at Brownsville.
(Schmidt, 1975) the parameters varied during practice have typically been confined to the scaling of invariant movement parameters as specified by Schema theory. Recent theoretical work (Sch¨ollhorn, Mayer-Kress, Newell, & Michelbrink, 2009; Sch¨ollhorn, Beckmann, Michelbrink, Sechelmann, Trockel, & Davids, 2006) has hypothesized that the rate of motor learning is a function of the magnitude and structure of practice variability. Differential learning theory hypothesizes that the optimal level of practice variability is higher than that hypothesized by prior motor learning theories (Adams, 1971; Gentile, 1972; Schmidt, 1975). In differential training high levels of variability are introduced into practice with respect to both the task goal and the execution redundancy (Frank, Michelbrink, Beckmann, & Sch¨ollhorn, 2008) and by randomizing the order (i.e., the structure) of practice trials (as in random practice; Lee & Genovese, 1988). Differential learning has been modeled (Frank et al., 2008) with the fourth-order polynomial:
ABSTRACT. High variability practice has been found to lead to a higher rate of motor learning than low variability practice in sports tasks. The authors compared the effects of low and high levels of practice variability on a simple unimanual arm rotation task. Participants performed rhythmic unimanual internal-external arm rotation as smoothly as possible before and after 2 weeks of low (LV) or high (HV) variability practice and after a 2-week retention interval. Compared to the pretest, the HV group significantly decreased hand, radioulnar, and shoulder rotation jerk on the retention test and shoulder jerk on the posttest. After training the LV group had lower radioulnar and shoulder jerk on the posttest but not the retention test. The results supported the hypothesis that high variability practice would lead to greater learning and reminiscence than low variability practice and the theoretical prediction of a bifurcation in the motor learning dynamics. Keywords: motor learning, practice, unimanual, variability
uch of the research comparing the effects of different levels of practice variability has examined predictions of Schema theory (Schmidt, 1975). Schema theory hypothesizes that the practice of the scaling of a generalized motor program (termed variable practice) enhances motor learning. In this theory practice variability should not be implemented by altering invariant characteristics of a motor program (e.g., relative timing, relative force). Research has found that variability in the sequencing of practice trials (random practice) can enhance the retention of motor learning compared to a blocked sequencing of practice trials (Lee & Genovese, 1988; Lee, Magill, & Weeks, 1985; Lintern, 1988; Magill, 1988; Newell, Antoniou, & Carlton, 1988). Practice variability can be introduced with respect to the task goal and the execution redundancy. The traditional approach to studying practice variability has been to introduce variability at the level of the task goal. Research including experiments that introduced variability in both the task goal and execution redundancy (Ranganathan & Newell, 2013) found that the benefit of constant or variable task goal practice was specific to each testing condition. Also, in this study the practice of redundant task solutions (movement trajectory) did not improve task performance. In prior research on motor learning the magnitude of practice variability has typically been relatively limited, both with respect to the number of practice variations and the parameters that are varied during practice. Typical motor learning studies have used a small number of variations of the target task (Lee et al., 1985; Shapiro & Schmidt, 1982). The effects of practicing these variations have also been studied when practiced in a randomized order versus in a blocked design (Shea & Morgan, 1979). As many studies of variable and random practice have used a motor programming approach
M
Q − Qc V (m) = − 2
c d m 2 − m3 + m4 3 4
(1)
where V(m) denotes the potential function of the learning rate m, Q is the level of practice variability, Qc is a critical threshold of practice variability, and c and d are positive constants. As shown in Figure 1A, when the level of practice variability Q is less than the critical threshold Qc the system is bistable, with stable attractor states for zero (m = 0) and positive (m > 0) learning rates. When practice variability exceeds the critical threshold (Q > Qc ) the system is monostable, with a single stable attractor state with a positive learning rate (m > 0; see Figure 1B). The Frank et al. (2008) model captures the occurrence of reminiscence. Reminiscence is the occurrence of learning, or an increase in performance, during a rest period (Eysenck, 1962, 1965; Hovland, 1951; Hull, 1943). Reminiscence is not fully understood, but has been attributed as being due to factors such as practice schedules, fatigue, motivation, consolidation, reactive inhibition (Ammons, 1988; Christina & Shea, 1988; Etnyre & Poindexter, 1995; Eysenck, 1965; Horn, 1975; Hovland, 1951; Lee & Genovese, 1988; Lintern, 1988; Magill, 1988), and the personality traits of extraversion and neuroticism (Eysenck, 1962, 1965; Horn, 1975). Correspondence address: Eric G. James, Department of Health and Human Performance, The University of Texas at Brownsville, REK Building 2.638, Brownsville, TX 78520, USA. e-mail: [email protected]
203
Downloaded by [Memorial University of Newfoundland] at 17:25 07 October 2014
E. G. James & P. Conatser
FIGURE 1. The fourth-order polynomial model (Equation 1) of differential learning with constant parameters c = 2.9 and d = 1. V is the order parameter potential and m is the order parameter. Panel A depicts the state of Q – Qc = –2. Prior to differential training the neuromotor system resides in the basin of attraction at a learning rate of m = 0 and after differential training at a learning rate of m = 1.77. Panel B depicts the state of Q – Qc = 0. This is the state of the system during differential training with a positive learning rate of m = 3. Adapted from Frank, Michelbrink, Beckmann, and Sch¨ollhorn (2008).
In the Frank et al. (2008) model reminiscence can occur when the level of practice variability increases from Q < Qc to Q > Qc and back again. When the level of practice variability is scaled up and down in this way the system moves from a stable attractor state of zero learning (m = 0; the solid circle in Figure 1A) to a stable state of learning (m > 0; the solid circle in Figure 1B) and remains in the stable attractor with a positive learning rate after the level of practice variability has decreased to Q < Qc (e.g., after practice has ended; the open circle in Figure 1A). The result is continued learning during a retention period (i.e., reminiscence). Reminiscence has previously been found to occur with high variability (HV) training (Beckmann & Sch¨ollhorn, 2003; Frank et al., 2008; Humpert, 2004; Humpert & Sch¨ollhorn, 2006; R¨omer, Sch¨ollhorn, & Jaitner, 2003). In differential training a high level of practice variability is created by giving learners multiple practice instructions in a randomized order, typically with only one trial performed for each instruction. The high number of task variations and short amount of time spent practicing each variation distinguishes differential training from typical random practice protocols, in which only a few task variations are practiced numerous times in a randomized order (Brady, 1998). Differential training instructions vary both the task goal and the execution redundancy characteristics such as the initial and/or ending conditions of a movement, the relative and/or 204
absolute duration of movement, joint angles, velocities, and acceleration profiles (Sch¨ollhorn et al., 2006). The Differential Learning Theory hypothesis that parameters such as relative timing and relative force should be varied during practice is contrary to Schema Theory (Schmidt, 1975). Prior research has shown HV (differential) practice to produce greater learning than traditional practice in soccer (Sch¨ollhorn et al., 2006), high jump (Sch¨ollhorn, Michelbrink, Welminsiki, & Davids, 2009), handball throwing (Wagner & M¨uller, 2008), shot put (Beckmann & Sch¨ollhorn, 2003), tennis (Humpert, 2004; Humpert & Sch¨ollhorn, 2006), and volleyball (R¨omer et al., 2003). Research on HV practice has typically examined the rate of motor learning in complex sports skills. Apart from the study of sports skills, one study found decreased postural sway after HV postural training (James, 2014). In the present study we compared the effects of low and high levels of practice variability on a simple rhythmic unimanual motor task. We hypothesized that HV practice would lead to greater motor learning than low variability (LV) practice. Method Participants Healthy young adult participants (N = 27; 12 men, 15 women) with a mean age of 23.93 years (SD = 3.79 years) Journal of Motor Behavior
Practice Variability and Arm Rotation
TABLE 1. The Arm Orientations and Training Movements Performed During High Variability Training Arm orientation Horizontal in front of the shoulder Horizontal directly to the right at shoulder height Vertical above the head
Downloaded by [Memorial University of Newfoundland] at 17:25 07 October 2014
FIGURE 2. Depiction of the experimental device used during the rhythmic arm internal-external rotation movements.
volunteered for this study. All participants signed an informed consent form that was approved by the local Institutional Review Board. All participants identified themselves as right-handed and as having had no prior surgery or injury to the arm, wrist, or shoulder. Task and Procedure Participants were seated comfortably and grasped the handle of a custom-made device (see Figure 2) with their right hand. The participant’s right arm was positioned horizontally at shoulder height in front of the body and participants were instructed to maintain the elbow in the extended position at all times during motor performance. The experimental task consisted of rhythmically rotating the experimental handle by internally and externally rotating the right arm (internal and external rotation of the shoulder and radioulnar joints) through a range of approximately 180◦ . The participants were given 2 min to familiarize themselves with the task and apparatus. The participants were instructed to perform the rotational movements as smoothly as possible and to synchronize the points of maximum internal and external rotation with auditory metronome tones that paced movement frequency in separate trials at 1 and 2 Hz. One trial was performed at each movement frequency on each test. Movement trials were of 30 s duration with 1 min of rest provided between trials. A pretest was performed, as well as a posttest after two weeks of practice and a twoweek retention test after practice had ended. Two Biometrics (Ladysmith, VA) ACL300 3-D accelerometers were attached to the lateral surface of the wrist and of the upper arm midway between the elbow and shoulder. DataLINK (Biometrics, Ladysmith, VA) software was used to collect acceleration data at a sampling rate of 1000 Hz during movement trials. The participants were randomly assigned to a low (LV; n = 14) or high (HV; n = 13) variability training group. During training the LV group repeated the arm rotation movements, with the instruction to make the movements as smooth and continuous as possible. During training the HV group performed arm rotation movements 2014, Vol. 46, No. 4
Training movement In-phase shoulder-radioulnar rotation Antiphase shoulder-radioulnar rotation Shoulder protraction-retraction with trunk fixed Shoulder protraction-retraction with arm fixed Shoulder flexion-extension Shoulder abduction-adduction Arm circumduction
with multiple instructions, with one trial performed with each instruction. For the HV training group the movement training included the use of three arm orientations and seven practice movements (see Table 1). The movements performed on each trial during training were randomly selected from a list that consisted of the seven individual movements and all possible (nonexclusive) pairings of these movements (e.g., in-phase shoulder-radioulnar rotation and shoulder protraction-retraction with the trunk fixed). One of the three arm orientations was randomly selected as the orientation in which to perform the selected movement. The participants were given instructions to perform each randomly selected movement in the selected arm orientation. When performing movement in the first two arm orientations listed in Table 1 the experimental device depicted in Figure 2 was used. When performing movements in the third arm orientation (vertical orientation) the arm was held above the head and the experimental device was not used. The in-phase mode was practiced by moving the upper arm and hand (shoulder and radioulnar joints) in the same direction. This movement has been shown to typically involve an in-phase coordination pattern (James, 2012). The antiphase coordination mode was practiced by maintaining the hand in a fixed position while internally and externally rotating the upper arm. While this movement initially appeared to be impossible to some participants they were readily able to perform it after a short amount of practice. Shoulder protraction and retraction were practiced with the trunk fixed (and the extended arm moving) and with the extended arm fixed (and the trunk moving). Shoulder flexion-extension, shoulder abduction-adduction, and arm circumduction were also practiced. Consistent with prior research on differential training participants in both groups performed the practice movements at a self-paced movement frequency and used a self-selected range of movement (Beckmann, Winkel, & Sch¨ollhorn, 2008; James, 2014; Sch¨ollhorn et al., 2006; Sch¨ollhorn et al., 2010). 205
Downloaded by [Memorial University of Newfoundland] at 17:25 07 October 2014
E. G. James & P. Conatser
No feedback regarding task performance was given to participants during data collection or during training. For both theoretical and methodological reasons no corrective feedback is typically provided during differential training. The theoretical objective of differential training instructions is to perturb the neuromotor system out of a habitual behavioral attractor, rather than to require the production of a specific prescribed movement pattern (Frank et al., 2008; James, 2014; Sch¨ollhorn et al., 2006; Sch¨ollhorn et al., 2010). In practical terms, this method of performing only a single trial with each movement instruction limits the possible use of corrective feedback during training. To maintain consistent feedback conditions across groups the LV training group also did not receive feedback during training. Each group practiced 2 times per week for two weeks and performed 20 practice trials of 1 min duration with 1 min of rest between trials on each day of practice. The arm rotation testing protocol pretest was performed within 48 hr before practice sessions began. The posttest was performed 24 hr after the last practice session and the retention test was performed two weeks after the posttest. Data Analysis Tangential acceleration data collected from the accelerometers were filtered with a ninth-order 30 Hz low-pass Butterworth filter. Hand and shoulder movement jerk were calculated from the accelerometry data. The jerk for the radioulnar joint was calculated as the difference between the hand and shoulder movement jerk data. The root mean square jerk (RMSJ) of the hand was calculated as an index of task performance and the jerk for the shoulder and radioulnar joints was calculated to examine the level of jerk at each effector used in this task. Jerk was calculated as Jerkt =
ACCLt − ACCLt−1 dt
where t is each data point in the time series, ACCL is the acceleration signal, and dt is the derivative of time. The RMSJ was calculated as RMSJ =
(Jerkt − M)2 /n
where t is each point in the time series from 1 to n (the number of data points) and M is the mean jerk. Three 2 Group × 2 Frequency × 3 Test repeated measures analysis of variance were used to analyze the RMSJ of the radioulnar and shoulder joints and of the hand. The calculation of all dependent variables was performed with coded MATLAB (The MathWorks, Natick, MA) programs. Inferential statistical analyses were performed using the SPSS software package (version 19.0) with a type-I error of .05 used to determine statistical significance. Fisher’s least significant difference test was used for post hoc analysis. 206
Results Hand Root Mean Square Jerk The RMSJ at the 1 Hz movement frequency was significantly lower than at 2 Hz, F(1, 25) = 190.393, p < .001. There was a significant group by test interaction, F(2, 50) = 4.740, p = .016 (see Figure 3C). In post hoc analysis for the HV group the retention test RMSJ was significantly lower than the pretest (p = .038) and the posttest (p = .001). There was no significant difference between the pre- and posttests (p = .528). For the LV group the posttest RMSJ was significantly lower than the retention test (p = .046). There were no significant differences between the preand posttests (p = .109) or the pre- and retention tests (p = .822). Also, on the retention test the HV group RMSJ was significantly lower than the LV group (p = .035) while there was no significant difference between groups on the pretest (p = .882) or posttest (p = .059). There was a significant group by frequency by test interaction, F(2, 50) = 3.459, p = .039 (see Figure 3). For the HV group at the 1 Hz movement frequency there were no significant differences between the pre- and posttests (p = .563), the pre- and retention tests (p = .631) or the post- and retention tests (p = .198). For the HV group at the 2 Hz movement frequency the retention test RMSJ was significantly lower than the pretest (p = .023) and posttest (p < .001). There was no significant difference between the pre- and posttests (p = .570). For the LV group at the 1 Hz movement frequency there were no significant differences between the pre- and posttests (p = .106), the pre- and retention tests (p = .728) or the postand retention tests (p = .107). At the 2 Hz movement frequency there were no significant differences between the pre- and posttests (p = .178), the pre- and retention tests (p = .911) or the post- and retention tests (p = .111). Also, at the 1 Hz movement frequency there were no significant differences between groups on the pretest (p = .572), posttest (p = .131), or retention test (p = .347). At the 2 Hz movement frequency the HV group RMSJ was significantly lower than LV group on the retention test (p = .018), while there was no significant difference between groups on the pretest (p = .889) or posttest (p = .069). The main effects for group, F(1, 25) = 0.001, p = .973, and test, F(2, 50) = 1.402, p = .256, were not significant. The group by frequency, F(1, 25) = 0.005, p = .942, and frequency by test, F(2, 50) = 2.411, p = .100, interactions were not significant. Radioulnar Root Mean Square Jerk The radioulnar RMSJ was significantly lower at the 1 Hz movement frequency than at the 2 Hz frequency, F(1, 25) = 156.040, p < .001. There was also a significant group by test interaction, F(2, 50) = 5.868, p = .006 (see Figure 4A). For the HV group the retention test RMSJ was significantly lower than on the pretest (p = .041) and posttest (p = .001). There was no significant difference between the pre- and posttests (p = .580). For the LV group the posttest RMSJ was significantly lower than the pretest (p = .036) and retention Journal of Motor Behavior
Downloaded by [Memorial University of Newfoundland] at 17:25 07 October 2014
Practice Variability and Arm Rotation
FIGURE 3. Representation of hand root mean square jerk data as a function of: group, test, and frequency. A = 1 Hz movement frequency; B = 2 Hz movement frequency; C = group and test. HV = High practice variability group. LV = Low variability practice group. Standard errors are represented with error bars. Asterisk indicates significant differences.
test (p = .034). There was no significant difference between the pre- and retention tests (p = .629). Also, the HV group RMSJ was significantly higher than the LV group on the posttest (p = .050), and significantly lower than the LV group on the retention test (p = .047). There was no significant difference between groups on the pretest (p = .914). The main effects for group, F(1, 25) = 0.012, p = .914, and test, F(2, 50) = 2.321, p = .112, were not significant. The group by frequency, F(1, 25) = 0.016, p = .899; frequency by test, F(2, 50) = 2.498, p = .092; and group by frequency by test, F(2, 50) = 2.809, p = .070, interactions were not significant. Shoulder RMSJ The RMSJ at the 1 Hz movement frequency was significantly lower than at the 2 Hz movement frequency, F(1, 25) = 84.587, p < .001. There was also a significant main effect for test, F(2, 50) = 8.041, p = .001. In post hoc analysis the pretest RMSJ was significantly higher than on the posttest 2014, Vol. 46, No. 4
(p < .001) or retention test (p = .004). There was no significant difference between the post- and retention tests (p = .893). There was a significant group by test interaction, F(2, 50) = 5.262, p = 008 (see Figure 4B). In post hoc analysis for the HV group the pretest RMSJ was significantly higher than the posttest (p = .028) and retention test (p = .001) while there was no significant difference between the post- and retention tests (p = .066). For the LV group the posttest RMSJ was significantly lower than the pre- (p = .001) and retention tests (p = .038). There was no significant difference between the pre- and posttests (p = .485). There was a significant frequency by test interaction, F(2, 50) = 5.081, p = .010. In post hoc analysis at the 1 Hz movement frequency the pretest RMSJ was significantly higher than on the posttest (p = .01) while there was no significant difference between the retention test and the pre- (p = .397) or posttests (p = .270). At the 2 Hz movement frequency the pretest RMSJ was significantly higher than on the
207
Downloaded by [Memorial University of Newfoundland] at 17:25 07 October 2014
E. G. James & P. Conatser
FIGURE 4. Representation of (A) radioulnar and (B) shoulder root mean square jerk data as a function of group and test. HV = high variability practice group; LV = low variability practice group. Standard errors are represented with error bars. Asterisk indicates significant differences.
post- (p < .001) or retention tests (p = .001), while there was no significant difference between the post- and retention tests (p = .632). The main effect for group was not significant, F(1, 25) = 0.139, p = .712. The group by frequency, F(1, 25) = 0.211, p = .650, and group by frequency by test, F(2, 50) = 1.963, p = .152, interactions were not significant. Discussion Prior research has shown that HV practice can lead to greater learning and retention on sports skills than lower variability practice (Beckmann & Sch¨ollhorn, 2003; Frank et al., 2008; Humpert, 2004; Humpert & Sch¨ollhorn, 2006; R¨omer et al., 2003). In the present study HV practice led to significant task improvement on the retention test. LV practice did not lead to a significant improvement on the task criterion (hand RMSJ). These findings supported the hypothesis that HV practice would provide greater learning and retention than LV practice of a simple movement task. 208
This indicated that HV training can provide greater learning and retention not only in sports skills but also in a simple motor task such as unimanual arm rotation. The LV practice led to a decrease in the jerk of individual effectors (radioulnar and shoulder joints) on the posttest. This form of training led to a small but not statistically significant improvement on the task criterion (hand rotation jerk). The decrease in the jerk of the two effectors was temporary and was not maintained on the retention test. While the LV training led to a nonsignificant decrease in hand jerk and a lack of retention in radioulnar and shoulder rotation jerk it may well be that with a greater number of subjects (or more training time) significant learning and retention may have occurred. The present results indicated greater learning and retention with HV training. The HV practice led to decreased shoulder jerk on the posttest that was maintained on the retention test. The HV practice also led to a delayed improvement in radioulnar joint jerk that occurred on the retention test. The improvement on the task criterion at the 2 Hz, but not the 1 Hz, movement frequency may have been due to a floor effect. The 1 Hz movement frequency may have been sufficiently easy that performance was already at a high level (i.e., the RMSJ was already low) on the pretest and so did not improve significantly on the post- and retention tests. The results supported the theoretical prediction of reminiscence (Frank et al., 2008) as significant learning took place during the retention period for the HV practice group. This effect was theoretically caused by the neuromotor system remaining within a stable attractor with a learning rate m > 0 (the open circle in Figure 1A) rather than returning to the stable attractor state with a learning rate of m = 0 where the system was located prior to training (the solid circle in Figure 1A). Improvement on the task criterion did not occur until the retention test. This indicated that the theoretical bifurcation from bistability to monostability did not occur until after the posttest. The finding of reminiscence after differential training is consistent with previous theoretical and experimental work on differential learning (Beckmann & Sch¨ollhorn, 2003; Frank et al., 2008; Humpert, 2004; Humpert & Sch¨ollhorn, 2006; R¨omer et al., 2003). The repetitive practice failed to lead to significant improvement on the task criterion. This indicated that the LV practice did not lead to learning (i.e., a relatively permanent improvement in performance; Guthrie, 1952). Theories of reminiscence have hypothesized that reminiscence is due to reactive or conditioned inhibition (Hull, 1943; Kimble, 1949) and/or consolidation (Eysenck, 1965; Eysenck & Frith, 1977). Massed practice can lead to shortterm reactive inhibition, due to mental (Walker, 1958) or physical work (Hull, 1943) that depresses the level of performance during practice. This effect diminishes with rest, leading to the occurrence of reminiscence. Several types of consolidation exist (e.g., synaptic, sleep, systems) that can also lead to greater performance after a retention period. Journal of Motor Behavior
Downloaded by [Memorial University of Newfoundland] at 17:25 07 October 2014
Practice Variability and Arm Rotation
The finding of decreased shoulder, radioulnar and hand jerk (though the latter was a nonsignificant trend) on the posttest for the LV training group suggests that this form of training did not lead to reactive inhibition on the posttest (Eysenck, 1965). It is possible that the absence of a posttest improvement on the task criterion for the HV training group was due to reactive inhibition. It could be the case that the HV training leads to greater reactive inhibition than LV training. If this is the case the reminiscence effect found on the retention test after HV training could have been due, in part, to a depression of performance on the posttest caused by reactive inhibition. However, reactive inhibition has generally been found to occur after massed practice (Eysenck, 1965), which was not performed in the present experiment. Also, the posttest was given after reactive inhibition should have ended (24 hr after the last day of practice). These facts do not support the interpretation of the lack of posttest improvement after the HV training as due to reactive inhibition. According to prior theoretical work (Eysenck, 1965; Eysenck & Frith, 1977) if the HV training did not lead to reactive inhibition the reminiscence effect that occurred after this form of training would be due to the occurrence of consolidation during the retention period. Synaptic consolidation typically occurs within a period of hours (Dudai, 2004), sleep consolidation within a few days (Walker et al., 2003), and systems consolidation over weeks or years (Squire & Alvarez, 1995). Based on these approximate time scales the reminiscence that occurred after HV training in the present study may have been due to sleep and/or systems consolidation. The present and prior studies of differential learning (Beckmann et al., 2008; Frank et al., 2008) have examined performance after retention periods of two or more weeks. To determine the possible types of consolidation processes that may occur after differential training future research is needed that includes multiple retention tests over a range of time intervals. It is also possible that the reminiscence that occurs after HV training may be due to a process or processes other than inhibition and consolidation. Existing modeling of differential learning (Frank et al., 2008) captures learning dynamics that differ from traditional theories of motor learning and reminiscence. In the Frank et al. (2008) model of differential learning a stable attractor state for a positive rate of motor learning exists before practice has begun. The neuromotor system can remain in this stable state after practice has ended, leading to reminiscence. The existence of a stable state for a positive rate of learning prior to the onset of practice is inconsistent with prior theories of reminiscence. In prior theories reminiscence is due to the dissipation of inhibition (Hull, 1943; Kimble, 1949) or consolidation of memory (Eysenck, 1965; Eysenck & Frith, 1977) that has accrued during training. Neither of these can theoretically exist before training has begun, as does the stable state for a positive rate of learning in the Frank et al. (2008) model. Future research is needed to test the predictions of traditional 2014, Vol. 46, No. 4
theories of reminiscence vs. the Frank et al. model and to determine if the reminiscence that can occur after HV training differs from the reminiscence that follows more repetitive forms of practice. REFERENCES Adams, J. A. (1971). A closed-loop theory of motor learning. Journal of Motor Behavior, 2, 111–149. Ammons, R. B. (1988). Distribution of practice in motor skill acquisition: A few questions. Research Quarterly for Exercise and Sport, 59, 288–290. Beckmann, H., & Sch¨ollhorn, W. I. (2003). Differential Learning in shot put. In W. I. Sch¨ollhorn, C. Bohn, J. M. J¨ager, H. Schaper, & M. Alichmann (Eds.), European workshop on movement science (pp. 68–69). K¨oln, Germany: Sport&Buch Strauβ. Beckmann, H., Winkel, C., & Sch¨ollhorn, W. I. (2008). Optimal range of variation in hockey technique training. Paper presented at the 2nd International Congress of Complex Systems in Sport, 10th European Workshop of Ecological Psychology, Madeira, Portugal. Brady, F. (1998). A theoretical and empirical review of the contextual interference effect and the learning of motor skills. Quest, 50, 266–293. Christina, R. W., & Shea, J. B. (1988). The limitations of generalization based on restricted information. Research Quarterly for Exercise and Sport, 59, 291–297. Dudai, Y. (2004). The neurobiology of consolidations, or, how stable is the engram? Annual Review of Psychology, 55, 51–86. Etnyre, B. R., & Poindexter, H. B. (1995). Characteristics of motor performance, learning, warm-up decrement, and reminiscence during a balancing task. Perceptual and Motor Skills, 80, 1027–1030. Eysenck, H. J. (1962). Reminiscence, drive and personality – revision and extension of a theory. British Journal of Social and Clinical Psychology, 1, 127–140. Eysenck, H. J. (1965). A three-factor theory of reminiscence. British Journal of Psychology, 56, 163–181. Eysenck, H. J., & Frith, C. D. (1977). Reminiscence, motivation, and personality. London, England: Plenum. Frank, T. D., Michelbrink, M., Beckmann, H., & Sch¨ollhorn, W. I. (2008). A quantitative dynamical systems approach to differential learning: Self-organization principle and order parameter equations. Biological Cybernetics, 98, 19–31. Gentile, A. M. (1972). A working model of skill acquisition with application to teaching. Quest, 17, 3–23. Guthrie, E. R. (1952). The psychology of learning. New York, NY: Harper & Row. Horn, P. W. (1975). Evidence for the generality of reminiscence as a function of extraversion and neuroticism. Journal of Psychology, 90, 41–44. Hovland, C. I. (1951). Human learning and retention. In S. S. Stevens (Ed.), Handbook of experimental psychology (pp. 613–689). New York, NY: Wiley. Hull, C. L. (1943). Principles of behavior. New York, NY: Appleton-Century-Crofts. Humpert, V. (2004). Vergleichende Analyse von Techniktrainingans¨atzen zum Tennisaufschlag [Comparative analysis of tennis serve training]. M. S. thesis. M¨unster, Germany: University of M¨unster. Humpert, V., & Sch¨ollhorn, W. I. (2006). Vergleich von Techniktrainingans¨antzen zum Tennisaufschlag [Comparison of tennis technique training]. In A. Ferrouti & H. Remmert (Eds.), Trainingswissenschaften im Freizeitsport – Symposium der dvsSektion Trainingswissenschaften vommm 7: 9 April 2005; Hamburg [Training sciences in recreational sports: symposium section 209
Downloaded by [Memorial University of Newfoundland] at 17:25 07 October 2014
E. G. James & P. Conatser training sciences VOMMM] (pp. 121–124). Hamburg, Germany: Bochum Czwalina. James, E. G. (2012). Fixed-point drift and hysteresis in frequencyscaled unimanual coordination. Journal of Motor Behavior, 44, 281–288. James, E. G. (2014). Short-term differential training decreases postural sway. Gait & Posture, 39, 172–176. Kimble, G. A. (1949). An experimental test of a two-factor theory of inhibition. Journal of Experimental Psychology, 39, 15–23. Lee, T. D., & Genovese, E. D. (1988). Distribution of practice in motor skill acquisition: learning and performance effects reconsidered. Research Quarterly for Exercise and Sport, 59, 277–287. Lee, T. D., Magill, R. A., & Weeks, D. J. (1985). Influence of practice schedule on testing schema theory predictions in adults. Journal of Motor Behavior, 17, 283–299. Lintern, G. (1988). Distributed practice: are there useful insights for application or theory? Research Quarterly for Exercise and Sport, 59, 298–302. Magill, R. A. (1988). The many faces of practice distribution in motor learning. Research Quarterly for Exercise and Sport, 59, 303–307. Newell, K. M., Antoniou, A., & Carlton, L. G. (1988). Massed and distributed practice effects: Phenomena in search of a theory. Research Quarterly of Exercise and Sport, 59, 308–313. Ranganathan, R., & Newell, K. M. (2013). Changing up the routine: intervention-induced variability in motor learning. Exercise and Sport Sciences Reviews, 41, 64–70. R¨omer, J., Sch¨ollhorn, W. I., & Jaitner, T. (2003). Differentielles lernen bei der Aufschlagannahme im Volleyball. In J. Krug & T. M¨uller (Eds.), Messpl¨atze, Messtraining, Motorisches Lernen [Measuring stations, measuring station training, motor learning] (pp. 129–133). Sankt Augustin, Germany: Academia Verlag. Schmidt, R. A. (1975). A schema theory of discrete motor skill learning. Psychological Review, 82, 225–260. Sch¨ollhorn, W. I., Beckmann, H., & Davids, K. (2010). Exploiting system fluctuations. Differential training in physical prevention and rehabilitation programs for health and exercise. Medicina, 46, 365–373.
210
Sch¨ollhorn, W. I., Beckmann, H., Michelbrink, M., Sechelmann, M., Trockel, M., & Davids, K. (2006). Does noise provide a basis for the unification of motor learning theories? International Journal of Sport Psychology, 37, 186–206. Sch¨ollhorn, W. I., Mayer-Kress, G., Newell, K. M., & Michelbrink, M. (2009). Time scales of adaptive behavior and motor learning in the presence of stochastic perturbations. Human Movement Science, 28, 319–333. Sch¨ollhorn, W., Michelbrink, M., Welminsiki, D., & Davids, K. (2009). Increasing stochastic perturbations enhances acquisition and learning of complex sport movements. In D. Araujo, H. Ripoll, & M. Raab (Eds.), Perspectives on cognition and action in sport (pp. 59–73). Hauppage, NY: Nova Science. Shapiro, D. C., Schmidt, R. A. (1982). The schema theory: Recent evidence and developmental implications. In J. A. S. Kelso & J. E. Clark (Eds.), The development of movement control and co-ordination (pp. 113–150). New York, NY: Wiley. Shea, J. B., & Morgan, R. L. (1979). Contextual interference effects on the acquisition, retention, and transfer of a motor skill. Journal of Experimental Psychology: Human Learning and Memory, 5, 179–187. Squire, L. R., & Alvarez, P. (1995). Retrograde amnesia and memory consolidation: A neurobiological perspective. Current Opinion in Neurobiology, 5, 169–177. Wagner, H., & M¨uller, E. (2008). The effects of differential and variable training on the quality of parameters of a handball throw. Sports Biomechanics, 7, 54–71. Walker, E. L. (1958). Action decrement and its relation to learning. Psychological Review, 65, 129–142. Walker, M. P., Brakefield, T., Seidman, J., Morgan, A., Hobson, J. A., & Stickgold, R. (2003). Sleep and the time course of motor skill learning. Learning & Memory, 10, 275–284.
Received August 23, 2013 Revised November 25, 2013 Accepted January 3, 2014
Journal of Motor Behavior
