The Differences in Transient Characteristics of Postural Control between Young and Older Adults across Four Different Postural Tasks

Abstract

Recently, the transient characteristics of postural sway have been suggested as an alternative or an improvement to traditional whole-trial analyses, possibly reflecting the sensory reweighing process. The purpose of this study was to assess if the transient characteristics of postural sway are sensitive to age and postural tasks. Twenty young adults (18–27 years old) and fifteen older adults (67–80 years old) performed parallel (eyes open and eyes closed), tandem and single-leg postural tasks for 30 s. Centre of pressure (CoP) velocity, amplitude and frequency were derived from force plate data. In addition to whole-trial estimates, we calculated the relative differences between the 1st and the 2nd (DIF_21) and the 1st and the 3rd (DIF_31) 10 s time intervals. Postural sway increased with the difficulty of the task, and the increase was more pronounced in the older adult group (p < 0.001; η² = 0.31–0.38 for group × task interactions). Large and statistically significant age × task interactions were shown for both DIF_21 and DIF_31 for CoP anterior–posterior frequency (p < 0.001; η² = 0.17–0.18). Possible age differences were also indicated for the transient response in CoP medial–lateral velocity in the parallel stance task. Specifically, we found an exaggerated transient response (i.e., relatively higher COP velocity in the first 10 s) in older adults compared to young adults when visual information was restored during the parallel quiet stance. This study shows the potential for an application of measures of the transient behavior of postural sway after the removal or restoration of visual information.

1. Introduction

Postural stability is paramount for older adults in order to be able to perform the activities of daily living [1] and minimize the risk of falling [2,3]. While various methods are being used to assess postural stability [3], the analysis of center of pressure (CoP) movement remains the gold standard [4,5]. Several variables extracted from CoP movement during quiet stance have been shown to be linked to a risk of falling in older adults [3]. Using longer trials for the collection of CoP data improves the reliability of the outcomes [6]. However, averaging the data across a longer trial (e.g., 30 s or 60 s) may mask important information regarding the transient behavior of postural sway [7].

Postural stability is maintained based on sensory information from the vestibular, visual and somatosensory systems [8,9]. The relative contribution of each system is dependent on environmental conditions and the postural task [10]. When a different sensory condition is suddenly introduced (e.g., when closing the eyes), the central nervous system must change the relative contribution of each sensory system (i.e., the sensory reweighing) [9,10]. This results in a transient period of increased postural sway [7,11]. While this phenomenon has been known for a long time [12,13], Reed et al. [7] were the first to suggest that analyzing the transient behavior of postural sway after changes in postural conditions could offer additional information to traditional outcome variables, such as whole-trial estimates. In the context of this paper, whole-trial estimates refer to the traditional postural sway’s outcome variables, where the CoP movement (e.g., CoP movement velocity) is averaged over an entire trial (e.g., 30 s). The indexes of transient behavior represent the difference in CoP movement between specific time intervals within the trial (e.g., between the first 10 s and the last 10 s).

Researchers have focused on transient behavior after a transition from the parallel stance to the single-leg stance [11,14,15]. Specifically, it has been shown that the transient behavior of postural sway in the single-leg stance differs between ballet dancers and the general population [11] and between preferred and non-preferred legs [15], but it is similar in athletes with and without a history of ankle sprain injury [14]. In their study, Reed et al. [7] reported that the transient behavior of postural sway in the parallel stance after closing the eyes distinguished younger adults from older adults, supporting its potential clinical relevance. In addition, they reported that transient characteristics and whole-trial estimates exhibited negligible correlations [7], which means that the analysis of transient behavior could offer additional information to traditional whole-trial variables. This finding was also supported by studies using single-leg stance tasks [11,15]. Apart from the study by Reed et al. [7], the results by Parreira et al. [16] and Jonsson et al. [17] indicated that time interval choice within the single-leg postural sway trial importantly affects age-related differences, but they did not calculate any indexes of the transient behavior of postural sway.

The purpose of this study was to compare young and older adults in terms of whole-trial estimates as well as the transient behavior of postural sway in four different postural tasks. We hypothesized that postural sway (measured by whole-trial estimates) would increase with task difficulty and that the increase would be more pronounced in older adults than in younger adults. We also hypothesized that the two age groups would exhibit different transient behaviors, with the relative decrease in postural sway being larger in the older group.

2. Materials and Methods

2.1. Participants

The sample comprised 20 young adults (18–27 years old, 10 females) and 15 older adults (67–80 years old, 10 females). Participant demographics are available in

Table 1. Basic participant characteristics.

Group	N	Age (Years)		Body Mass (kg)		Body Height (m)
		Mean	SD	Mean	SD	Mean	SD
Younger males	10	23.4	3.2	76.4	8.2	1.82	0.04
Younger females	10	24.1	2.5	58.6	5.5	1.71	0.06
Older males	5	70.2	3.3	85.2	8.2	1.77	0.09
Older females	10	74.5	6.6	64.5	9.1	1.59	0.06

SD—Standard deviation.

. The exclusion criteria included the history of neurological diseases and any musculoskeletal injuries in the past 6 months. Older adults were also assessed with the Berg balance scale [18], which indicated that they were able to safely and efficiently maintain balance throughout most functional tasks (mean: 53.2 ± 2.1 points, range: 50–54). Participants were given detailed information about the testing procedures and were required to sign a written informed consent form prior to the measurements. The study was approved by the National Committee for Medical Ethics of the Republic of Slovenia (Approval Number 0120-321/2017/4) and was conducted in accordance with the latest revision of the Declaration of Helsinki.

2.2. Study Design and Procedures

The study was conducted as a single-visit, cross-sectional experiment. After the explanation of the procedures and obtaining Berg Balance Scale scores, the participants performed four quiet stance tasks on a force plate in a randomized order: parallel stance, tandem stance and single-leg stance, with the parallel stance being performed separately with eyes open (EO) and eyes closed (EC). All repetitions for one task were completed before moving onto the next task. For each task, three 30 s repetitions were performed for each leg, with 120 s breaks between repetitions and 180 s breaks between the tasks. All tasks involved 30 s of quiet standing. The acquisition of the data began right after a transition (e.g., from EO to EC or from parallel to single-leg stance). These transitions were the transition to a new body posture in the tandem and single-leg tasks, closing the eyes in the parallel EC task and opening the eyes in the parallel EO task. Therefore, the parallel EO task was preceded by 15 s of standing with EC, which was not taken into the analysis. In all tasks except parallel EC, the participants were instructed to focus on a black dot on a white background, which was placed at approximately eye level and ∼4 m away. A detailed description of the tasks can be found in

Table 2. An overview of postural tasks performed in the study.

Task	Descriptions of the Tasks and the Transition to the Task Positions
Parallel EO	Participants assumed the parallel stance position on the force plate and closed their eyes. The eyes were opened (as instructed by the examiner) after 15 s, simultaneously with the start of the data recording. The arms were hanging loosely at the sides.
Parallel EC	Participants assumed the parallel stance position on the force plate. After that, they closed their eyes (as instructed by the examiner) simultaneously with the start of the data recording. The arms were hanging loosely at the sides.
Tandem	Participants stood with their feet parallel, at the posterior edge of the force plate. When ready, they transitioned the dominant leg forward and assumed the tandem stance. The data recording started manually 1 s after they assumed the tandem position. The hands were placed on the hips.
Single-leg	Participants stood with their feet parallel, at the middle of the force plate. When ready, they transitioned the weight to their dominant leg, assuming a single-leg stance. The data recording started ~1 s after they assumed this position. The hip of the free leg was at 0°, and the thigh was parallel to the standing leg, while the knee was flexed at ~90°. The hands were placed on the hips.

EO—eyes open; EC—eyes closed.

. Prior to the main trials, one familiarization trial was performed for each task. If the participant lost balance, the trial was repeated.

2.3. Data Acquisition and Analysis

We used a single force plate (model 9260AA; Kistler, Winterthur, Switzerland) to collect ground reaction force data at 1000 Hz. The data were processed automatically in the manufacturer’s MARS software (version 4.0; Kistler, Winterthur, Switzerland). After low-pass filtering (Butterworth, 2nd order, 10 Hz), the data were further automatically processed in the same software to obtain the mean CoP velocity (total, anterior–posterior (AP) and medial–lateral (ML)), CoP amplitude (AP and ML) and CoP frequency (AP and ML). The CoP amplitude was defined as the average amount of CoP sway in the AP or ML direction, which is calculated as the common length of the trajectory of the CoP sway (in AP and ML direction) and divided by the number of changes in movement direction. The CoP frequency was defined as the frequency of oscillations of the CoP calculated as the number of CoP signal peaks in the AP or ML direction (i.e., changes in the direction of CoP movement) divided by the measurement time. Firstly, we calculated traditional whole-trial estimates by averaging CoP characteristics over the entire trial. Additionally, we calculated the relative differences between the 1st and the 2nd (DIF_21) and the 1st and the 3rd (DIF_31) 10 s time intervals. These relative differences were expressed as percentages, with 100% representing no change, >100% indicating an increase between the 1st and 2nd (or the 1st and 3rd) time intervals and <100% indicating a decrease between the 1st and 2nd (or the 1st and 3rd) time intervals. The DIF_21 and DIF_31, using 10 s time intervals, were chosen in accordance to previous studies that showed that these indexes potentially bear some information independent from the whole-trial estimates [11,15]. Our preliminary analyses (not published) also indicated that 10 s interval data generally exhibit acceptable reliability (ICC = 0.6–0.8), while 5 s data do not.

2.4. Statistical Analysis

Statistical analyses were performed with SPSS (version 25.0, SPSS Inc., Chicago, IL, USA). Descriptive statistics are reported as mean ± standard deviation. The normality of the data distribution was verified with Shapiro–Wilk tests. The associations between the whole-trial outcome variables and corresponding transient characteristic indexes (DIF_21 and DIF_31) were analyzed by Pearson’s product–moment correlation (r) and interpreted as negligible (<0.1), weak (0.1–0.4), moderate (0.4–0.7), strong (0.7–0.9) and very strong (>0.9) [19]. We performed a mixed-model analysis of variance (ANOVA), with the group (young adults, older adults) as the between-subject factor and the postural task as the within-subject factor. In case of statistically significant main effects or interactions, we additionally performed pairwise (between pairs of tasks) or separate (difference between groups for each task individually) t-tests with Bonferroni corrections. The threshold for statistical significance was set at α < 0.05.

3. Results

3.1. Correlations

We assessed the associations between whole-trial outcome variables (CoP velocity, CoP amplitude and CoP frequency) with their corresponding indexes of transient behavior (DIF_21 and DIF_31). There were no statistically significant correlations between the whole-trial outcome variables and corresponding transient characteristic indexes (DIF_21 and DIF_31) within parallel EO (r ≤ 0.31; p > 0.05), parallel EC (r ≤ 0.39; p > 0.05) and tandem stance tasks (r ≤ 0.45; p > 0.05). For the single-leg task, whole-trial CoP ML velocity (r = 0.63–0.66; p < 0.05) and CoP ML amplitude (r = 0.57–0.68) were moderately correlated with their corresponding DIF_21 and DIF_31. No statistically significant correlations were present when the age groups were analyzed separately (r ≤ 0.19; p < 0.05).

3.2. Difference between Young and Older Adults in Traditional Whole-Trial Estimates

The two-way ANOVA showed a large and statistically significant effect of task for all whole-trial outcomes (p < 0.001; η² = 0.34–0.85). For all CoP velocity and amplitude outcomes, there were also statistically significant age × task interactions (p < 0.001; η² = 0.31–0.38) and main effects of age (p ≤ 0.003; η² = 0.25–0.40). There were no statistically significant age effects on CoP frequency (p = 0.126–0.384) and no age × task interactions (p = 0.755–0.784). A closer examination of CoP velocity and amplitude data (Figure 1, top and middle panel) reveals that (a) the postural sway increased with task difficulty, (b) both age groups had similar values for parallel EO and parallel EC tasks and (c) postural sway was statistically significantly larger (all p < 0.001) in older compared to young adults in tandem and single-leg stances.

3.3. Difference between Young and Older Adults in Transient Characteristics of Postural Control

The two-way ANOVA showed age × task interactions for both DIF_21 and DIF_31 for CoP AP frequency (p < 0.001; η² = 0.17–0.18). There were no other statistically significant interactions (p = 0.054–0.852) and no statistically significant main effects of age (p = 0.074–0.918). The task had a statistically significant effect (p = 0.001–0.014; η² = 0.11–0.37) on 11 out of 14 outcome measures (7 CoP variables and 2 different indexes of transient behavior). Specifically, only DIF_31 indexes for CoP total velocity (p = 0.075), CoP AP velocity (p = 0.771) and CoP AP amplitude (p = 0.468) were not influenced by task. Pairwise comparisons showed that, regarding CoP frequency, there were statistically significant differences between age groups in parallel EC, tandem and single-leg stance for DIF_21 and tandem and single-leg stance for DIF_31 (Figure 2 and Figure 3). Despite the lack of the main effect of age on the CoP ML velocity according to the ANOVA (p = 0.056–0.074), post hoc tests indicated a potential difference in DIF_21 and DIF_31 between the age groups regarding CoP ML velocity in the parallel EO task (p = 0.046–0.048). Namely, both indexes were closer to 100% in young adults (99.1 ± 11.0% and 94.1 ± 14.7% for DIF_21 and DIF_31, respectively) than in older adults (88.5 ± 8.2% and 85.0 ± 10.6% for DIF_21 and DIF_31, respectively) (Figure 2 and Figure 3, upper right chart).

4. Discussion

The purpose of this study was to compare whole-trial estimates as well as the transient behavior of postural sway across four different postural tasks between young and older adults. Our first hypothesis was confirmed as the postural sway increased with task difficulty, and the increase was more pronounced in older adults than younger adults (as confirmed by statistically significant task × group interactions). There were only a few variables related to transient behavior that indicated a difference between the age groups, which were mostly limited to CoP frequency. Interestingly, there were some trends in the easiest task (parallel EO) for both indexes that were closer to 100% in young adults than in older adults for CoP ML velocity. This would indicate the presence of larger transient changes in postural sway in older adults after vison is returned (note again that the parallel EO task was preceded by 15 s of standing with eyes closed). However, the main effects were not significant, making it impossible to draw any conclusion on the significance of these observations.

Several previous studies suggested that long trials (e.g., 60 s or more) are needed for a valid assessment of postural sway [6,20,21]. However, the transient behavior of postural sway that is masked by whole-trial averaging could also contribute important information regarding the individual’s postural control. For instance, falls in older adults could occur during sudden changes in sensory conditions (e.g., lights turning off) or body position. Such conditions entail efficient sensory reweighing [8]. Reed et al. [7] suggested that the difference between CoP outcomes obtained right after closing the eyes and those obtained later in the trial could reflect the efficiency of sensory reweighing. Moreover, they reported that older adults show an exaggerated transient response after closing the eyes compared to young adults [7]. The removal of visual information compels the central nervous system to quickly adjust the relative sensory weights for vestibular and proprioceptive systems [8,9,10]. Although purely speculative at this point, it could be that the transient behavior of postural sway after a change in sensory conditions (presumably reflecting sensory reweighing processes) is as if not more important for preserving stability and preventing falls in older adults than whole-trial, averaged postural sway outcomes. As in previous studies [7,11,15], the indexes of the transient behavior of postural sway were generally not correlated with whole-trial outcomes, supporting the notion that additional information may be obtained by calculating the indexes of transient behavior.

In this study, postural sway increased with task difficulty in both groups, and the increase was exaggerated in older adults compared to young adults, which is well in line with previous studies [22,23]. In addition, older adults exhibited different transient behavior (larger changes throughout the trial) in postural sway as per CoP frequency. These results are difficult to interpret, as the differences between age groups were not consistent across tasks. This was likely observed because the older adults were relatively fit and their postural control was comparable to the young adults in the easier tasks (as also reflected in the whole-trial variables.) The difference between the groups in the parallel EC task could imply that older adults show a larger variation in CoP frequency across the trial and need more time to stabilize their postural control after closing their eyes. However, the CoP frequency seems to be the least sensitive among the CoP variables for assessing the risk of falling in older adults [3], making further inferences questionable. On the other hand, older adults tended to exhibit different transient behavior for CoP ML velocity in the parallel EO task (note that this task was preceded with 15 s of standing with closed eyes). Namely, both indexes of transient behavior were closer to 100% in young adults (94–99%) than in older adults (85–88%), indicating that transient behavior is larger in older adults after visual information is restored. Indeed, sensory reweighing also needs to arise when a sensory source is re-established and not only after its removal [8,9]. As mentioned before, Reed et al. [7] reported that the indexes of transient behavior after closing the eyes (such as in the parallel EC task in our study) distinguished between older and young adults with moderate to large effect sizes (Cohen’s d = 0.45–0.71). We did not observe statistically significant differences in terms of the transient behavior between the age groups in parallel EC task, although some outcomes (CoP ML velocity and CoP ML amplitude) approached statistical significance (p = 0.064–0.078). The older adults in our study were relatively fit and physically active, as reflected in high Berg balance scale scores (51–54 points). Therefore, it is reasonable to expect that we would observe even larger transient behavior in older adults at a high risk of falling; however, this needs further investigation. Larger responses to the removal/restoration of vision are perhaps expected, given the increased weight placed on visual information for postural control in older compared to young adults [24]. As the sensory weights for the vestibular and somatosensory systems change with visual manipulations, the transient behavior of postural sway could reflect the deterioration of these two systems, which is frequently observed in older adults [24,25,26].

Regarding the transition to the single-leg stance, it was found that transient behavior is different between ballet dancers and the general population [11] but not between athletes with and without ankle sprain history [14]. Tasks involving postural change (e.g., transition from parallel to single-leg stance) are methodologically problematic, because it is difficult to distinguish between the transient effects of the postural change itself (i.e., the CoP moving under the stance leg) and the supposed subsequent effects of sensory reweighting. As in previous studies [11,14,15], we started the data acquisition ~1 s after the participant assumed the new position; however, the effect of CoP movements arising from the change in body position itself remains unknown. Due to these issues, we suggest that investigating transient behavior after the removal or restoration of vision (parallel EC and EO tasks) is methodologically a sounder option. In addition, a transient response in postural sway was also observed after the addition of a cognitive task [27], which is another promising avenue, especially in light of the well-known increased cognitive cost in older adults [28] and possible links of dual-task ability to the risk of falling [29,30]. Finally, extending the application of analyzing the transient behavior of postural sway to patients with neurological diseases is also intriguing. For instance, when vision is removed or disturbed, patients with Parkinson’s disease seem to be able to reweigh sensory information appropriately but not as quickly as age-matched controls [31]. Indeed, Brown et al. [13] reported a larger transient increase in postural sway in Parkinson’s disease patients (compared to age-matched controls) after closing the eyes. Although not statistically significant, there was also a very small trend in their study [13] suggesting that the patients also showed a slight transient increase in postural sway after their vision was re-established.

This study does not come without its limitations. Because we investigated postural tasks of various difficulties, we included only older adults that were fit and able to maintain balance independently. Further studies should investigate the transient behavior of postural sway in older adults with different levels of physical fitness. It would also be interesting to assess transient behavior across even longer trials. However, as previous studies have indicated that a quasi-steady state is reached 5–15 s after the change in postural tasks or sensory conditions [7,17], we chose to minimize the trial’s duration to avoid the effect of fatigue. The sample size of the study was relatively small considering that there was substantial between-participant variability in both age groups, which reduced the statistical power of our analyses. Finally, another important limitation of the study is the fact that the trials were initiated approximately 1 s after the participants assumed the single-leg stance. As mentioned earlier, tasks involving postural change are difficult to assess, as it is impossible to accurately distinguish between the effects of the postural change itself (i.e., the CoP moving under the stance leg during transition) and the later transient behavior. Thus, the effect of CoP movement arising from the change in body position (parallel to single-leg) is not entirely known. In parallel stance tasks, the transition between EC and EO was cued by the examiner, who also started the data acquisition at the same time. We believe that any small inaccuracies in this procedure did not significantly affect our results.

5. Conclusions

This study shows the potential for the application of measures of the transient behavior of postural sway after the removal or restoration of visual information. We found a trend for a larger transient response in older adults compared to young adults when visual information was restored during the parallel quiet stance. The measures of the transient behavior of postural sway could add additional information regarding the sensory reweighing process when changes in sensory conditions are introduced. Future studies should explore the potential for measures of the transient behavior of postural sway for the assessment of the risk of falling in older adults.