Association between Anthropometric Variables, Sex, and Visual Biofeedback in Dynamic Postural Control Assessed on a Computerized Wobble Board

: Anthropometrics and sex inﬂuence balance performances, and visual information can change anthropometrics’ relation and the postural sway. Therefore, the aim of the present study was to evaluate the effect of anthropometric characteristics, sex, and visual biofeedback and/or their interaction on a computerized wobble board. Twenty-seven (14 females, 13 males) young adults performed three 30-s double leg stance trials on a wobble board during two conditions: with visual and without visual biofeedback. Visual biofeedback improved ( p = 0.010) balance on a wobble board with respect to the condition without visual biofeedback. Regardless of sex, no differences between conditions were found ( p = 0.088). When investigating the effect of anthropometrics variables, sex, and their interactions on conditions, a signiﬁcant main effect of the lower limb/height ratio, sex, and their interaction on the condition without visual biofeedback was found ( p = 0.0008; R 2 = 0.57). For the visual biofeedback condition, signiﬁcant effects for sex and body mass ( p = 0.0012; R 2 = 0.43) and sex and whole-body moment of inertia ( p = 0.0030; R 2 = 0.39) were found. Results from the present study showed (1) visual biofeedback improved wobble board balance performance; (2) a signiﬁcant main effect of lower limb/height ratio, sex, and their interaction on the wobble board performances without visual biofeedback emerged; (3) signiﬁcant effects were found for sex and body mass and sex and moment of inertia in the visual biofeedback condition. Findings from the present study could have an impact on training and evaluations protocols, especially when several populations such as children, older adults and people with balance disorders are involved.


Introduction
Balance is defined as the ability to maintain or make adjustment in order to keep the body's center of mass (CoM) over the base of support (BoS) through an integrative use of somatosensory, visual, and vestibular systems [1,2]. According to its definition, balance can be divided into static balance, which is the ability to keep the BoS through minimum motion, and dynamic balance, as the ability to maintain balance while the CoM is projected outside the BoS [3].
Several factors, such as anthropometrics, sex, and feedbacks play a key role in postural control. Studies [4][5][6][7] focusing on the anthropometric variables have shown that body mass [4,8], height [6,9] and lower limb length [9] are directly related to postural control over the lifespan. In fact, body mass negatively influences the postural control of adolescents [10], young adults [9,11], and the elderly [12]. Similarly, body height and lower limb length have been identified as the most influencing anthropometric variables in young adults [9]. On the contrary, although the influence of sex on postural control is well known, findings are not conclusive, with some studies showing better balance in women [13], and others in men [14], or showing no differences between sex [15].
Postural control can also be influenced by cognitive processes such as the attentional focus, which can be driven by feedback such as visual biofeedback (VBF) [16,17] and acoustic biofeedback [18], with VBF more effective and accurate when compared to other sensory modalities. VBF can be presented by providing subjects with additional artificial visual information about body movement designed to augment the natural information and facilitate the adoption of appropriate strategies to keep postural control as steady as possible [17,19]. Therefore, when designing a VBF system, the methodology applied plays a crucial role. In the literature, some studies used a direct visualization of real-time location of the subject's center of pressure (CoP) [20][21][22], while others displayed the subject's relative lateral (left vs. right) weight distribution [23][24][25]. In some cases, a numerical representation of the percentage of weight distribution between the left and right feet has also been used [20,26]. Regardless of the modality, these studies showed that VBF has a positive influence on balance and that even small changes in feedbacks can make a difference in balance performances, probably due to the enhanced neuromuscular control [16].
In clinical and field settings, the assessment of balance can provide accurate and sophisticated information regarding the efficiency of the neuromuscular system. Several postural control assessments have been used to evaluate dynamic balance. Force plates [27] and functional tests [28,29] are commonly used to evaluate balance performances due to their accuracy, validity, and reliability. However, because of the multifaceted features of the postural control tests, by clearly reflecting the complexity of this ability, new simple approaches [30,31] are needed to accurately evaluate dynamic postural control. In particular, during large-scale evaluations, practical, inexpensive, administrable, and accurate tools are preferable. In this context, computerized wobble boards (WBs) have been suggested as reliable and simple tool to evaluate dynamic balance in healthy young subjects in field and laboratory settings [31][32][33]. Unstable platforms, such as WBs, are the most used tools to train human balance, showing their effectiveness in improving postural control among different populations [34].
Although the influence of anthropometrics, sex, and VBF on human balance is well documented, focusing on the effect of the above-mentioned factors and/or their interaction on computerized WB could provide useful information to adapt and tailor individualized balance training protocols. Therefore, the aim of this study is to evaluate the effect of anthropometric characteristics, sex, and VBF during dynamic balance performance assessed on a computerized WB in young healthy adults.

Subjects and Procedures
The Institutional Review Board of the Department of Human Sciences, Society, and Health of the University of Cassino and Lazio Meridionale approved this study (approval No.: 14357; date: 18 June 2019), designed to evaluate the effects of anthropometric characteristics, sex, and VBF on balance performances in healthy young adults. Prior the evaluation, twenty-seven subjects (14 females, 13 males) were voluntarily selected among the students' population. They were fully informed about the procedures and the aim of the study, and subsequently their informed consent was provided. Subjects were excluded if they reported any pre-existing condition such as neurological condition, musculoskeletal injury of the back or lower extremities, or any other disorder that could influence their balance ability.
Before starting the testing session, subjects' characteristics were assessed. Body mass and height were measured by means of a scale with an integrated stadiometer with a precision of 0.1 kg and 0.1 cm (Seca, model 709, Vogel & Halke, Hamburg, Germany), and body mass index was calculated. In addition, whole-body moment of inertia (MI), an inertial quantity measurement of the human body essential for quantitative analysis of human motion, was collected. Subjects' MI on the frontal axis through the center of mass was computed using the following Equation (1) [35]: where HT (m) represents height and M (kg) body mass. Lower limb length was also measured from the anterior superior iliac spine to the most distal part of the medial malleolus by using a tape measure while the subject laid in supine position. Subsequently, the relative lower limb length, which defines the proportion of total stature that is comprised by the lower limbs, was also computed.

Wobble Board Test
Computerized Balance Board WSP (Well Sport Project, G.S.J. Services S.r.l., Rome, Italy) is a proprioceptive platform; we used the WB model, incorporating a triaxial accelerometer (Phidget Spatial 0/0/3 Basic 1041, Phidgets Inc. 2016, Calgary, AB, Canada). The platform is composed by a circular surface (diameter 40 cm, height 2 cm) and placed on a plastic material semispherical support (diameter 12 cm, height 6 cm). The WB model is connected via USB cable, with a sampling frequency of 200 Hz. The WB tilt (maximal tilt angle = 20 • ) angle data is then transmitted to a customized software displaying real time balance performance on a monitor (resolution = 1920 × 1080) through a motion marker (MM, diameter = 6 mm). The software user interface showed the MM, represented by a yellow circle, a Target Zone (TZ, diameter = 6.5 cm) displayed by a red circle which represented the stability area (0 • tilt angle), and a countdown of the trial. The boundaries of the TZ and the MM were the same for all subjects during experimental sessions.
After a detailed explanation of the testing procedures, with a short demonstration and verbal support, subjects were asked to stand barefoot on the WB, which was placed directly on the floor, with a comfortable double leg stance, keeping their hands on their hips and instructed to keep the board flat (0 • tilt) and as still as possible for as long as possible within a recording period of 30 s.
The test session consisted of a 3-min free practice on the WB followed by three 30-s double leg stance trials and 1-min sitting recovery in between. Subjects were asked to perform two conditions: (1) with VBF, where subjects were asked to keep visual focus on the MM showed on the display and try to keep it inside the TZ as long as they could within 30 s; and (2) without Visual Biofeedback (NVBF), looking at a fixed point on a black board, where subjects were asked to keep visual focus on a marking in front of them and instructed to maintain the WB as flat and still as possible for as long as possible within the recording period of 30 s. During the VBF and NVBF condition, the screen or the black board were positioned 2 m far in front of the WB while standing on it. The order of conditions was randomly assigned.

Statistical Analysis
Normal distribution was verified by the Shapiro-Wilk test, and means and standard deviations (SD) were calculated for all variables. Data were analyzed using STATA 15 (StataCorp LP, College Station, TX, USA). Statistical significance was set at p < 0.05. Firstly, a repeated measures mixed model was applied to evaluate the possible differences in balance performance between the VBF and NVBF conditions in relation to sex. Participants were considered as random effect, whereas conditions (VBF and NVBF) and sex were treated as fixed effect. The models were fitted using the residual maximum likelihood to account for the small sample. Subsequently, the trend over trials for each condition in relation to sex was checked by using orthogonal polynomial contrasts. Finally, the main effect of sex and its interactions with selected anthropometric characteristics for each condition was investigated using mixed linear regression models. Bryk/Raudenbush R-squared (R 2 ) values were calculated for each model.

Results
Subject's characteristics are represented in Table 1. The repeated measures mixed model analysis showed a significant main effect for the VBF condition (p = 0.010; 95% Confidence Intervals (CI) = 0.74−-5.45) (Figure 1). No significant differences were found between sexes (p = 0.088; 95% CI = −5.62-0.39). As there were no sex differences between conditions, the trend over trials for each condition was checked by aggregating both sexes. The orthogonal polynomial contrasts analysis showed a significant linear trend (p ≤ 0.0001; 95% CI = 1.16-3.13) only for the VBF condition with an estimated linear slope 2.15 s (Figure 2).
When investigating the effect of selected anthropometrics, sex, and their interactions on both conditions, the mixed linear regression analysis showed a significant main effect of the lower limb/height ratio (HTR), sex, and their interaction on the NVBF performance (p = 0.0008; R 2 = 0.57) ( Table 2). For VBF, two models were developed, and significant effects were found for sex and body mass (Table 3; p = 0.0012; R 2 = 0.43) and sex and MI (Table 4; p = 0.0030; R 2 = 0.39). Interactions and main effects are graphically represented in

Discussion
The purpose of this study was to evaluate the effect of anthropometric characteristics, sex, VBF, and their interaction on computerized WB during dynamic balance performance in healthy young adults.
Biofeedback has been used for many years in rehabilitative and preventive training protocols among different populations, such as healthy [19,36] or pathological populations [37][38][39]. However, VBF, due to its immediate, continuous, correct, and accurate information, represents the most effective modality compared to others sensors during dynamic balance performances [40]. To confirm this, evidence showed better balance in judo athletes [41], young karatekas [42], healthy subjects [40,43], young adults [44], and elderly people [45] when performing VBF compared to NVBF. Although previous studies are in line with the present findings, comparisons are difficult because of the different VBF and balance outcome used. However, Cawsey and colleagues [46] showed that an increase in dependence on augmented sensory information for the control of standing posture influences the somatosensory input conditions of the foot and ankle, confirming the significant differences during the VBF condition with respect to NVBF. It is well known that standing on an unstable platform results in changes in sensory biofeedback and subjects increasing their reliance on visual information. Therefore, the postural control could be more efficient in the VBF condition when standing on an unstable platform. Another possible explanation could be related to the VBF methodology applied. In the present study, real-time VBF showing a MM and a TZ portrayed by a red circle was used. For this reason, such VBF characteristics could have improved the WB performance by influencing the subjects' postural strategies and facilitating accuracy and goal directedness of postural dynamic control. Therefore, based on the present results, VBF condition should be taken into consideration during WB balance assessment and neuromuscular training.
Literature also suggests that visual information changes the relationship between anthropometrics and the postural sway [9]. In fact, in a previous study [9] a greater correlation between postural sway and body mass was found when the balance test was performed with eyes opened. Similarly, in the present study, a significant relationship between body mass and balance performance was found in the VFB (in both sexes) condition. On the contrary, in other studies, postural sway increased in NVBF conditions such as balance tests with eyes closed [47,48]. This is probably due to the difference in the NVBF modality.
In fact, although the eyes closed condition might be included in the NVBF category, there might be a difference in terms of difficulty between having a visual cue with eyes open (a mark on a black board) and the eyes closed with no visual orientation.
In addition, in this study, a significant relationship between MI during the VBF condition in males and females was also found. MI is a mechanical parameter of the human body, usually used in studies on balance and posture, in correlation with other parameters, such as body mass and lower limb strength [35]. An explanation about this result might be found in the test execution. In fact, although the test protocol was standardized in terms of execution with clear directions, such as standing barefoot on the WB with a comfortable double leg stance, keeping hands on the hips and the board flat at 0 • tilt, no further indications were provided about the trunk control. It might have happened that during the test execution the subjects leaned forward or backward with their trunks in order to focus on the screen to keep their balance, and as MI estimates the subjects' whole body MI on the frontal axis through the centre of mass, this further centre of mass displacement might have influenced the performance during the VBF condition [7,49].
Commonly, limbs' length, especially upper limb length, has shown a positive correlation in postural sway in both eyes opened and closed condition in females [50]. This positive relationship has been mainly attributed to reaching tests, where longer limbs might favour the subjects during the tests. However, in the present study, a significant negative relationship between HTR and postural control in males and females during the NVBF condition was found. Usually, taller individuals, more evident in males, tend to have longer lower limbs, and this condition is often associated with a greater distance between the centre of mass and BoS, resulting in a higher postural sway [51]. However, in the present study, an interaction between HTR and sex during the NVBF was found. In particular, at higher HTR values, men increased their dynamic balance performance, while women decreased it. Direct comparisons of these results are difficult as no studies have compared the effect of HTR and sex on WB performances. However, Alonso et al. [50] focused their attention exclusively on lower limbs length, and in contrast with this study, showed a moderate path sway in males when eyes were closed. The authors hypothesized that the sex differences found during performances when eyes were opened and closed were due to greater anthropometric variables, lower flexibility, and slower neurophysiologic processing of inferences in men. Nevertheless, focusing on these findings, we cannot confirm these assumptions.
In the present study, when comparing the VBF test trials, learning effect was evident regardless of sex. Few studies [52,53] focused their attention on learning effect. A previous study [52] showed that one trial for each test task created insufficient self-confidence, allowing individuals not to feel confident with the tasks. For increasing sufficient selfconfidence, a good strategy could be to increase the trials' number. In fact, in the present study, individuals performed three dynamic trials on the WB. Similarly, in Wrisley et al. [53], young adults performed three trials for each test session, showing significant learning effect during both the eyes opened and eyes closed conditions. In the present study, during VBF condition, both males and females reported the same learning effect. On the other hand, during NVBF condition, no learning effect was found for both sexes. Evidence [54,55] has shown that with repeated balance activities, performance improves, especially during complex tasks such as standing on unstable platform or when visual information is suppressed. However, due the different methodologies used, comparison was difficult. It also could be interesting to evaluate if the learning effect would be as evident among several populations, such as subjects with visual or proprioceptive deficits.
Several limitations need to be acknowledged for this study. The sample was limited to healthy young adults. Other populations such as elderly people, athletes, or subjects with chronic diseases should be evaluated to explore possible differences. In addition, it could be possible that other VBF strategies might have a different influence on WB performances. Finally, since only a few studies using WB as an assessment tool are available in the literature, the results of this study cannot be compared with other studies.

Conclusions
Findings from this study highlighted that VBF may improve WB balance performance with respect to the NVBF condition in healthy young adults. Regarding the anthropometrics variables, results showed a significant main effect of HTR, sex, and their interaction on the NVBF WB performance. For the VBF condition, instead, significant effects were found for sex and body mass and sex and MI. In addition, results from the present study could have an impact on training and evaluations protocols, especially when several populations such as children, athletes, older adults and people with balance disorders are involved. In fact, specific balance exercises such as the monopodalic stance, the double leg stance, the tandem stance on foam, and walking on unstable surfaces with (to ensure the safety of individuals) or without support during eyes opened and closed conditions could improve neuromuscular control. Lastly, the affordability and transportability of WBs are key factors during filed evaluations, making data collection on balance performances feasible for health scientists and/or coaches looking for inexpensive, portable, reliable, and valid assessment tools. However, further research should assess different populations to evaluate the effect of anthropometric characteristics, sex, VBF and NVBF, and their interaction on WB.