Effects of different balance practice methods on motor learning in unstable environments: a randomized pre-post experimental design

Abstract

[Purpose] To verify and compare the effects of different training methods on balance retention during early motor learning in unstable standing environments. [Participants and Methods] Twenty-six healthy adults were randomly assigned to three groups for a balance maintenance task involving holding a tandem standing position on a slackline. The active assistance (AA) group held a cane in each hand, the passive assistance (PA) group had the participant manually assisted by a therapist, and the no assistance (NA) group had no canes or assistance. Tandem standing time (TST) was measured before, immediately after, and one week after the three weeks training period. [Results] Intra-group comparisons revealed that the AA and PA groups showed significant improvements in TST immediately after practice, with the PA group maintaining this significant improvement one week later. No significant differences were observed between groups at any time point. [Conclusion] Our findings suggest, that in the early stages of motor learning for balance in an unstable standing environment, practicing with manual assistance may be more effective for developing an accurate perception of optimal balance.

INTRODUCTION

In clinical physical therapy, balance training may be undertaken by the patient using devices such as a cane, walker, or parallel bars, through manual assistance by a therapist, or without any external assistance by independent patient practice. These clinical decisions are often guided by the therapist’s empirical experience rather than established theoretical frameworks. Motor learning theory offers valuable insights into these clinical judgements. Motor learning has been defined as a process incorporating skills and experience that results in relatively permanent changes in the ability to perform skilled actions¹⁾. Fitts’ model divides motor learning into three stages: cognitive, associative, and autonomous²⁾. In the cognitive stage, the therapist provides active instruction and manual assistance to maximize successful experiences. In the associative stage, active therapist involvement is gradually tapered to encourage learning through the patient’s own successes and failures. The autonomous stage is characterized by minimal feedback and assistance from the therapist for enhancing movement accuracy, thereby transitioning the learner towards independent control. In other words, the therapist’s role evolves from direct involvement to encouraging independence throughout the motor learning process.

Studies on motor learning and balance have explored the effects of verbal guidance³⁾, mechanical assistance^{4, 5)}, and active support⁶⁾. Sideway et al. reported that frequent manual assistance while training center-of-gravity symmetry might interfere with the patient’s automatic motor control process, resulting in no detectable learning effects³⁾. In contrast, interactive assistance in complex motor tasks may enable goal experiences that are unattainable by the patient alone^{4, 7,8,9,10)}. With the use of balancing poles in a skiing slalom task, Wulf et al. demonstrated that active support helped learners experience advanced control strategies by providing the opportunity to test balance and control on their own⁶⁾. Indeed, the effectiveness of assistance appears to depend on task complexity and the learner’s stage of motor learning^{11, 12)}. While simpler tasks may be prone to excessive assistance, challenging ones often benefit from both passive and active support.

To date, the majority of studies on the impact of active support on motor learning to improve balance have focused on non-clinical settings or specific tasks, such as skiing simulators, rather than real-world movements relevant to clinical physiotherapy. Moreover, few reports have directly compared the effects of active assistance, passive assistance, and independent practice on motor learning under similar conditions.

This study examined for effective strategies in motor learning during the cognitive stage of balance training. Participants practiced maintaining a tandem standing position on a slackline under three conditions: active assistance with canes, passive assistance from a therapist, and no support. The duration of standing was quantitatively analyzed to determine the effectiveness of each approach towards improving immediate and short-term balance.

PARTICIPANTS AND METHODS

The enrolled participants comprised 26 students aged 18 years or older from the Shinshu University Faculty of Medicine Department of Health Sciences. All participants provided informed consent after a detailed verbal explanation of the study’s purpose and methods. In the recruitment process, candidates were excluded if they had previous experience using a slackline or any medical condition causing pain or balance issues in the lower limbs. This study was approved by the Shinshu University Medical Ethics Committee (no. 1759).

Balance practice and measurement were performed using a slackline (15 m Gibbon Slackline Classic, Gibbon, Stuttgart, Germany) mounted on a 4-meter-long, 30-cm-high wooden stand (Gibbon Slackrack, Gibbon) (Fig. 1). This setup was chosen due to its novelty and challenging nature, with few reported studies even in healthy adults. The balance task involved maintaining a tandem standing position on the slackline.

Fig. 1.

Study set-up.

We utilized an on-off switch and a versatile biological information analysis program for the experiment. Tandem Standing Time (TST) was defined as the duration when both feet were in contract with the slackline. It was measured to evaluate the participants’ ability to maintain balance while positioning their limbs on the slackline.

Several studies on slacklining have been published, particularly in the fields of kinesiology, sports science, psychology, and rehabilitation; however, the overall research is limited. We selected slacklining as our balancing task because it is a novel activity that none of the participants had previously experienced.

Two T-canes (Himawari, Yamaguchi, Japan) were used for active support exercises. Measurements were conducted using ON-OFF switches and a dedicated amplifier (M.E., Nagano, Japan), and data analysis was performed using a biometric analysis program (BIMUTAS-E ver. 2.0, Kissei Comtec, Nagano, Japan).

Participants provided general demographic information including age, gender, height, weight, dominant foot (i.e., the foot used to kick a ball¹³⁾, and prior experience with slacklining. Participants were then randomly assigned into the active assistance (AA) group that practiced balance using canes, the passive assistance (PA) group practicing balance with manual assistance from a therapist, or the no assistance (NA) group, which practiced balance without any active or passive assistance.

The Japan Slackline Federation lists the following six as fundamental exercises. 1. Climbing up and down (getting on and off the line), 2. Standing on one leg (balancing on one leg for 5 sec), 3. Moving forward, 4. Moving backward, 5. Exposure (balancing sideways on the line with both feet), 6. Turning (changing direction). In contrast, AA and PA are original methods that have been set up to simulate the situation of movement support in clinical practice of physical therapy.

Baseline tandem standing time (TST) was measured for all participants before the practice regimen. TST was defined as the duration of both feet simultaneously in contact with the slackline without any external support. Based on the training method of each group, balance exercises were conducted three times a week for three weeks, with each session lasting 14 minutes. Immediately post-practice TST was measured within one day after completing the 3-week training. Lastly, a retention test was conducted one week after the training period. The training schedule and measurement intervals were based on Astrid et al.’s review of balance training effects on neuromuscular control and functional performance¹⁴⁾.

This study employed a randomized pre-post experimental design with three test groups (AA, PA, and NA), three test points (pre-test, post-test, and retention test), and one measurement variable (TST).

Participants in the AA group adjusted the T-canes to achieve an elbow flexion angle of approximately 30° when the cane tips were positioned 15 cm ahead and 15 cm outside their feet on both sides of the slackline. Instructions included standing on the slackline for as long as possible with minimal reliance on the canes and the eyes focused forward. Participants in the PA group were first asked to stand on their non-dominant foot side on a 20-cm-high platform set up on both sides of the slackline. A physical therapist with 11 years of clinical experience provided fingertip assistance to the lateral shoulder area, gradually reducing support as balance improved. The participants were instructed to stand on the slackline for as long as possible with minimal assistance while crossing their arms in front of their chest, with the eyes focused forward. Participants in the NA group were tested similarly to those in the PA group, but without any guidance from the therapist. In the case that balance was lost, the participants were instructed to step down onto the adjacent platform and then immediately resume the task.

For TST measurements, two ON-OFF switches were taped to the slackline where the heel of the participant would be placed during balance testing. Participants were instructed to maintained a tandem standing position on the slackline, with the non-dominant foot placed ahead of the dominant foot. The analog signals from the switches were converted to A/D signals at a sampling frequency of 1 kHz and analyzed using a biometric analysis program on a laptop computer. TST was recorded as the time from establishing a tandem standing position to the time either heel left an ON-OFF switch. Three practice trials preceded formal testing. Each measurement was repeated three times, with a 30-sec interval between trials.

TST data were recorded in seconds and rounded to one decimal place. The highest value from the three trials was used for each of the pre-test, post-test, and retention test analyses. Normality testing by the Shapiro–Wilk test revealed non-normality in approximately 11% of tests. Accordingly, intra-group comparisons were performed using the Friedman test, with post-hoc analysis via the Wilcoxon signed-rank test. Inter-group comparisons employed the Kruskal–Wallis test, with the Mann–Whitney U test for post-hoc analysis. Bonferroni correction set the significance threshold at p≤0.017. All statistical analyses were conducted using SPSS ver. 27.0 J software (IBM, Tokyo, Japan).

RESULTS

Table 1 summarizes the mean and standard deviation values for age, height, weight, body mass index, sex ratio, and dominant foot for each group. One participant in the NA group suffered a slackline-unrelated injury and was excluded from the analysis.

Table 1. Participant characteristics.

	AA group	PA group	NA group
Age (years)	22 ± 4.1	24 ± 5.9	22 ± 3.4
Height (cm)	168.8 ± 5.6	172.1 ± 5.6	168.3 ± 6.5
Weight (kg)	60.7 ± 6.0	62.6 ± 5.2	65.6 ± 11.8
BMI (kg/m2)	21 ± 1.0	21 ± 3.4	23.5 ± 3.9
Gender ratio (male/female)	6/3	8/1	6/1
Dominant foot (right/left)	9/0	5/4	6/1

Data are presented as the mean ± standard deviation or number.

BMI: body mass index; AA: active assistance; PA: passive assistance; NA: no assistance.

The mean and standard deviation (SD) of age, height, weight, and BMI for each group of participants in this study, as well as the sex ratio and dominant foot are shown in Table 1.

As a result of the within-group comparison in the TST, significant differences were observed in the respective AA and PA groups (AA: p=0.015, PA: p=0.013, NA: p=0.368). Next, the post-test results showed that the PA group showed a significant prolongation of TST in the retention test compared to the pre-test (AA pre-post: p=0.066, AA post-retention: p=0.944, AA pre-retention: p=0.086, PA pre-post: p=0.018, PA post-retention: p=0.312, PA pre-retention: p=0.015).

As the results of the between-group comparisons, there were no significant differences between groups for all tests (pre-test p=0.733, post-test p=0.575, retention-test p=0.221).

Effect sizes and power for between-group comparisons are shown in Table 2.

Table 2. Effect sizes and testing power for between-group comparisons.

		Effect size	Testing power
Pre	PA group -SA group	0.23	0.07
	PA group-NA group	0.28	0.08
	SA group-NA group	0.01	0.03
Post	PA group -SA group	0.64	0.25
	PA group-NA group	0.46	0.13
	SA group-NA group	0.24	0.06
Retention	PA group SA group	1.01	0.52
	PA group-NA group	0.85	0.35
	SA group-NA group	0.30	0.08

AA: active assistance; PA: passive assistance; NA: no assistance.

The change in TST over time for each test is shown in Tables 3, 4, 5.

Table 3. Changes in TST over time for each participant characteristics in Group AA.

	Pre (sec)	Post (sec)	Retention (sec)
AA1	1.0	2.9	4.5
AA2	2.2	2.8	2.8
AA3	1.5	1.6	2.5
AA4	1.8	2.6	2.0
AA5	1.1	1.9	3.2
AA6	0.4	1.5	1.7
AA7	0.8	5.5	2.7
AA8	3.5	2.3	1.3
AA9	2.0	3.0	2.1

TST: tandem standing time; AA: active assistance.

Table 4. Changes in TST over time for each participant in Group PA.

	Pre (sec)	Post (sec)	Retention (sec)
PA1	1.4	2.5	2.9
PA2	2.1	2.6	8.9
PA3	1.7	7.1	7.5
PA4	0.8	4.7	2.5
PA5	0.2	6.4	8.3
PA6	1.0	2.0	3.5
PA7	3.1	2.6	2.1
PA8	1.0	1.1	1.5
PA9	1.5	4.8	4.6

TST: tandem standing time; PA: passive assistance.

Table 5. Changes in TST over time for each participant in Group NA.

	Pre (sec)	Post (sec)	Retention (sec)
NA1	2.0	4.4	1.9
NA2	2.1	2.0	1.8
NA3	1.7	4.6	2.6
NA4	2.2	1.3	3.6
NA5	1.3	3.3	3.6
NA6	1.4	3.9	4.6
NA7	0.5	1.8	1.7

TST: tandem standing time; NA: no assistance.

DISCUSSION

This study evaluated the effects of different learning methods (active, passive, and no assistance) on motor learning in the early stages of balance acquisition under challenging standing conditions in healthy young participants. Participants in the AA and PA groups demonstrated significant improvements in balance ability after a training regimen, with the PA group displaying significant retention of enhanced TST in a post-practice retention test. These results offer novel insights into the effect of learning methods on motor-learning ability.

Intra-group comparisons revealed significant improvements in TST for the AA and PA groups, but not for the NA group. These findings suggest that using active support (e.g., canes) or passive assistance (e.g., manual support) may enhance the initial stages of motor learning in challenging balance tasks for healthy young individuals. Domingo and Ferris⁴⁾ explored motor learning during treadmill tasks of varying track widths with and without assistance. They reported that while passive assistance did not significantly enhance motor execution ability compared with unassisted practice, it reduced the number of failures as the task difficulty increased, thereby facilitating goal-oriented learning. Similarly, the effects of interactive assistance on complex tasks were demonstrated to aid motor learning by enabling successful experiences that would otherwise be unattainable without assistance^{4, 9,10,11,12)}. Therefore, the prolonged TST observed in the PA group during the retention test may have reflected the benefits of passive assistance in providing opportunities to experience successful goal-oriented movement without numerous failures. Through the enablement of participants to perceive the correct movement patterns, passive assistance likely contributed to more effective motor learning in this challenging balance task.

In their study, Wulf et al.⁶⁾ demonstrated the value of active support in a slalom task using a ski simulator. Participants using balancing poles experienced the sensation of skilled movement normally felt by athletes and were able to test balance control strategies independently. Similarly, the AA group in our cohort employed canes to stabilize themselves when needed, which facilitated the feeling of success and reinforced the merit of goal-oriented experiences in the challenging slackline task.

According to Bernstein¹⁵⁾ and Verheijen et al.¹⁶⁾, early motor learning should involve reducing the degree of freedom in body movements to simplify the task. In a similar way, active and passive assistance in our study may have helped participants manage joint movements when standing in a tandem position, thus improving balance ability during the initial stages of motor learning. Taken together, the above findings provide evidence that appropriate active or passive assistance can enhance motor learning by providing the sensation of successful movements and limiting confounding body movements.

The PA group showed significant improvement across all three tests, with remarkable improvement maintained in the retention test. Although the AA group exhibited significant improvements during training, retention effects were less pronounced. It was noteworthy that the NA group showed no significant changes across tests. The TST performance curves showed a tendency for the PA group to achieve relatively better results despite greater individual variability. Comparable gradual improvements were noted for the AA group. In contrast, performance in the NA group often exhibited a decline between the post-test and retention test, suggesting that a lack of assistance limited effective motor learning.

Postural stability has been defined as the ability to maintain balance in challenging conditions by keeping the body’s center of gravity within the stability zone, i.e., the range in which the body can be held without changing the base support surface¹⁷⁾. Fujisawa et al.¹⁸⁾ demonstrated significant non-linear correlations between standing balance and walking ability in stroke patients, while Mochizuki and Mineshima¹⁹⁾ reported greater postural stability in participants who were better able to maintain the body’s center of gravity within the stability zone. By receiving physical support, the PA group was likely able to experience and internalize the goal of stabilizing their center of gravity, leading to improved outcomes. These findings underscore the value of passive assistance in promoting balance acquisition, especially in tasks requiring significant postural control. In clinical physical therapy, such dynamic assistance may enhance motor learning by enabling patients to repeatedly experience and internalize goal-directed movement strategies.

This study had several limitations. First, the slackline environment, while novel and challenging, differs from balance training scenarios for older adults or individuals with disabilities. Future research should explore more clinically relevant contexts. Second, the sample size was relatively small, which increased the risk of Type II error. Larger studies are needed to confirm our findings and examine for differences across assistive methods more robustly. Third, the training duration in this study may have been insufficient. In their review, Zech et al.¹⁴⁾ found that balance training typically spanned 4–12 weeks, with 2–7 weekly sessions lasting 5–90 minutes. As effective results for postural sway often required regimens of 6 weeks or more, extending the duration of slackline training could yield more pronounced improvements in motor learning. Lastly, factors such as upper limb positioning and eye focus during training may have influenced balance control; future studies should consider these variables to optimize training protocols.

This study revealed differences among practice conditions on the learning and retention results of a balancing motor function task. Specifically, short-term motor learning gains were most pronounced with passive assistance by a therapist in the early stages of balance training. In clinical physical therapy, assisting individuals to maintain proper alignment and experience successful goal-directed movements may enhance motor learning and treatment outcomes.

Conflicts of interest

The authors declare no conflicts of interest.