Physical function factors associated with the modified star excursion balance test in healthy university students Authors: Shomaru Ito, Rio Suzuki, Kanta Hiramatsu, Tatsuya Igawa, Ryunosuke Urata, Hiroto Takahashi, Riyaka Ito, Kosuke Suzuki, Yusuke Maeda Published: 2026 ABSTRACT Abstract [Purpose] To investigate the associations between dynamic balance performance assessed by the modified star excursion balance test (mSEBT) and multiple physical function factors in healthy university students. [Participants and Methods] Thirty-five students from a single university participated in this cross-sectional study. Dynamic balance was evaluated using the mSEBT in the anterior, posterolateral, and posteromedial directions. Physical function variables included ankle range of motion, knee extension strength, ankle plantarflexion strength, back muscle strength, skeletal muscle index, trunk lean body fat-free mass, lower limb lean body fat-free mass (leg FFM), and indices of muscle quality. Spearman’s rank correlation coefficients were calculated, followed by forced-entry multiple linear regression analysis for each reach direction. [Results] Bivariate correlation analysis revealed significant associations between mSEBT performance and multiple physical function variables, including ankle plantarflexion strength and lower leg FFM, across all reach directions. In contrast, multiple regression analyses did not identify any significant independent predictors, although some models demonstrated modest explanatory power. [Conclusion] mSEBT performance in healthy university students may be associated with multiple physical function factors rather than a single determinant. Therefore, dynamic balance performance may reflect the combined influence of several physical functional components.Key words: Modified Star Excursion Balance Test, University students, Dynamic balance INTRODUCTION University students experience injuries during organized sports activities, leisure-time physical activity, commuting, and daily life at home or in dormitories, regardless of their competitive level1, 2). Such injuries may lead to a decline in physical and psychological health, academic difficulties, restrictions on social participation, and increased medical costs, resulting in substantial individual and societal burdens3). Therefore, injury prevention is an important public health concern among university students. Injury has been associated with intrinsic factors, such as muscle strength, flexibility, and coordination, in addition to environmental and external factors4). Notably, dynamic balance has been identified as a major intrinsic risk factor for injury5). Dynamic balance is fundamental for maintaining postural stability during daily activities and sports movements6). Reduced dynamic balance has been associated with lower limb injuries and decreased athletic performance5, 7). Bliekendaal et al. reported that university students with below-average performance on the Star Excursion Balance Test (SEBT) had an approximately 7-fold higher incidence of ankle injuries, suggesting that impaired dynamic balance is an important risk factor for this population8). The modified Star Excursion Balance Test (mSEBT), developed as a simplified version of the SEBT, does not require specialized equipment and can be completed within a short time. Its reliability and reproducibility are comparable to those of the original SEBT9), and it has been shown to be applicable as a dynamic balance assessment tool for healthy young adults, including university students10). Previous studies have shown that mSEBT performance is associated with lower limb muscle strength and ankle joint range of motion11). However, muscle strength alone does not sufficiently explain mSEBT performance12), suggesting that multiple physical function factors, including joint mobility and trunk muscle strength, may also contribute. In addition, performance in other dynamic balance assessments has been reported to correlate with the lower limb phase angle13), which is an indicator of muscle quality14). These findings imply that mSEBT performance may also be influenced by body composition factors such as lower limb muscle mass and muscle quality. Accordingly, this study included lower limb muscle mass and indices of muscle quality15) in the analysis in addition to joint range of motion and muscle strength. Therefore, this study aimed to investigate the association between dynamic balance performance assessed using the mSEBT and multiple physical function factors in healthy university students. We hypothesized that mSEBT performance would be associated with several physical function factors rather than determined by a single factor. PARTICIPANTS AND METHODS Thirty-five students from a single university (23 male, age: 20.4 ± 1.3 years, height: 167.5 ± 9.0 cm, weight: 63.0 ± 10.3 kg) participated in this study. The inclusion criterion was the absence of orthopedic or neurological disorders affecting motor function. Participants who reported pain during testing were excluded, although none met this criterion. All participants provided written informed consent after receiving verbal and written explanations, in accordance with the Declaration of Helsinki. This study was approved by the Ethics Committee of the International University of Health and Welfare (25-TA-100). Dynamic balance was assessed using the mSEBT. Participants performed a single-leg stance while reaching maximally in the anterior, posterolateral, and posteromedial directions16). Participants were instructed to reach as far as possible in each direction. The hands were placed on the waist, and the stance heel remained in contact with the floor. A total of nine consecutive trials were performed for each direction, and the mean value of the final three trials was used for analysis. Additional trials were performed and recorded when the stance heel was lifted or when the weight was placed on the reaching foot. The reach distance was normalized to the lower limb length (from the anterior superior iliac spine to the medial malleolus). The starting lower limb (left or right) was randomly assigned. When measurements were initiated from the right lower limb, the order of directions was anterior, posterolateral, and posteromedial. The same order was applied for the left lower limb. A 1-minute rest period was provided after completing each direction. If participants reported excessive fatigue, additional rest was allowed as needed. Measurements were obtained separately for the left and right lower limbs, and statistical analyses were performed using the mean values of both lower limbs, consistent with previous studies that used averaged values to represent overall dynamic balance performance10). Ankle dorsiflexion and plantarflexion range of motion were measured using a goniometer in 5-degree increments. Knee extension strength (KES) and ankle plantarflexion strength (APS) were measured during maximal voluntary isometric contraction using a handheld dynamometer (ANIMA Corp., Tokyo, Japan)17, 18), and back muscle strength (BMS) was measured using a digital dynamometer (Takei Scientific Instruments Co., Ltd., Niigata, Japan)19). Each strength measure was recorded three times (twice for BMS), and the maximum value was used. Strength values were adjusted using a lever arm and normalized to body weight20). For KES, participants were seated at the edge of a treatment bed with their arms crossed over the chest. The hip and knee joints were positioned at 90° flexion. The dynamometer was placed on the anterior aspect of the distal lower leg, and a stabilization belt was applied at the same position and fixed to the bed frame. The lever-arm length was defined as the distance from the lateral knee joint line to the center of the dynamometer application point, and torque was calculated by multiplying the measured force by this distance17). For APS, participants were positioned in long sitting with the knee fully extended. The dynamometer was placed on the plantar surface of the forefoot. The ankle was stabilized using a belt wrapped around the distal lower leg and secured to the trunk. The lever-arm length was defined as the distance from the lateral malleolus to the center of the dynamometer application point, and torque was calculated by multiplying the measured force by this distance,18). For BMS, participants stood with the knees fully extended and the trunk inclined approximately 30° forward while maintaining a straight back. The chain length was adjusted to maintain this posture prior to measurement19). Participants exerted maximal isometric force for approximately 3 seconds. The highest value of two trials was used for analysis. Body composition was assessed using a bioelectrical impedance analysis device (InBody 270; InBody Co., Ltd., Seoul, Republic of Korea). Skeletal muscle index (SMI), trunk fat-free mass (trunk FFM), and lower limb fat-free mass (leg FFM) were included as indicators of muscle mass. Trunk and leg FFM were normalized to body weight. In addition, to evaluate muscle quality, the ratios of knee extension strength to leg FFM (KES/FFM), ankle plantarflexion strength to leg FFM (APS/FFM), and back muscle strength to trunk FFM (BMS/FFM) were calculated15). For the mSEBT performance, ankle range of motion, KES, APS, and leg FFM were measured bilaterally, and the mean of the left and right sides was used for analysis. Statistical analyses were performed using the IBM SPSS Statistics version 29 (IBM Corp., Armonk, NY, USA). Spearman’s rank correlation coefficients were calculated to examine the association between mSEBT performance in each direction, physical function, and body composition variables. Variables showing significant correlations were entered into forced-entry multiple linear regression models as independent variables with mSEBT performance as the dependent variable. This approach was used because lower limb muscle strength alone does not sufficiently explain mSEBT performance, and multiple physical function factors may contribute to combination12). Prior to the multiple regression analysis, multicollinearity among the independent variables was assessed because it may cause unstable coefficient estimates and misleading statistical interpretations21). The condition index, variance proportion, and variance inflation factor (VIF) were calculated using an initial model that included all candidate independent variables identified in the correlation analysis. Multicollinearity was considered present when high VIF values or high condition indices accompanied by variance concentrations in the same component were observed. Multicollinearity was considered present when the variance inflation factor (VIF) exceeded 5 or when the condition index exceeded 30 with variance concentrations in the same component21). When multicollinearity was detected, some variables were excluded, and the model was reconstructed. This process was repeated until acceptable multicollinearity levels were achieved, and the resulting model was adopted as the final model. Assumptions of linear regression were evaluated for the final models, including residual normality, homoscedasticity, influential observations, and multicollinearity. The level of statistical significance was set at p<0.05. RESULTS The mSEBT performance in each direction, physical function data, and body composition data of the participants are shown in Table 1. Table 1. mSEBT performance and physical function data (n=35)Anterior (%)77.2 ± 10.5Posterolateral (%)102.8 ± 14.0Posteromedial (%)95.7 ± 15.0Age (years)20.4 ± 1.3Height (cm)167.5 ± 9.0Weight (kg)63.0 ± 10.3BMI (kg/m²)22.4 ± 2.5Sex (male)23ADF ROM (degree)21 ± 6APF ROM (degree)45 ± 7APS (N*m/kg)0.7 ± 0.2KES (N*m/kg)1.7 ± 0.4BMS (N/kg)17.8 ± 3.1Trunk FFM 0.34 ± 0.03Leg FFM 0.12 ± 0.01SMI (kg/m²)7.2 ± 1.0APS/FFM (N*m/kg)13.2 ± 2.9KES/FFM (N*m/kg)5.1 ± 1.4BMS/FFM (N/kg)38.8 ± 7.5Mean ± standard deviation. leg FFM, and trunk FFM are expressed relative to body weight. KES/FFM, APS/FFM, and BMS/FFM represent ratios of muscle strength to fat-free mass. mSEBT: modified Star Excursion Balance Test; ADF/APF ROM: ankle dorsiflexion/ankle plantarflexion range of motion; KES: knee extensor strength; APS: ankle plantarflexion strength; BMS: back muscle strength; FFM: fat-free mass; SMI: skeletal muscle index. mSEBT performance and physical function data (n=35) Mean ± standard deviation. leg FFM, and trunk FFM are expressed relative to body weight. KES/FFM, APS/FFM, and BMS/FFM represent ratios of muscle strength to fat-free mass. mSEBT: modified Star Excursion Balance Test; ADF/APF ROM: ankle dorsiflexion/ankle plantarflexion range of motion; KES: knee extensor strength; APS: ankle plantarflexion strength; BMS: back muscle strength; FFM: fat-free mass; SMI: skeletal muscle index. The bivariate correlations between mSEBT performance and physical function factors are shown in Table 2. In the anterior direction, significant positive correlations were observed with APS (r=0.37) and leg FFM (r=0.37). In the posterolateral direction, significant positive correlations were found with APS (r=0.49), KES (r=0.48), BMS (r=0.39), trunk FFM (r=0.40), leg FFM (r=0.41), and SMI (r=0.42). In the posteromedial direction, significant positive correlations were observed with APS (r=0.51), KES (r=0.40), BMS (r=0.39), trunk FFM (r=0.34), leg FFM (r=0.35), SMI (r=0.49), and BMS/FFM (r=0.36). Table 2. Correlations between mSEBT performance and physical function parameters (n=35)AnteriorPosterolateralPosteromedialADF ROM0.01 (–0.34 to 0.35)–0.00 (–0.34 to 0.34)0.06 (–0.29 to 0.40)APF ROM0.18 (–0.17 to 0.50)–0.07 (–0.41 to 0.28)–0.04 (–0.38 to 0.31)APS0.37* (0.03 to 0.63)0.49* (0.18 to 0.71)0.51* (0.20 to 0.72)KES0.30 (–0.05 to 0.58)0.48* (0.16 to 0.70)0.40* (0.07 to 0.66)BMS0.21 (–0.14 to 0.52)0.39* (0.05 to 0.64)0.39* (0.05 to 0.64)Trunk FFM0.23 (–0.12 to 0.53)0.40* (0.07 to 0.65)0.34* (–0.01 to 0.61)Leg FFM0.37* (0.04 to 0.63)0.41* (0.08 to 0.66)0.35* (0.00 to 0.62)SMI0.28 (–0.07 to 0.57)0.42* (0.09 to 0.67)0.49** (0.18 to 0.71)APS/FFM0.06 (–0.29 to 0.40)0.18 (–0.17 to 0.49)0.13 (–0.22 to 0.46)KES/FFM0.00 (–0.34 to 0.35)0.17 (–0.18 to 0.48)–0.04 (–0.38 to 0.31)BMS/FFM0.18 (–0.18 to 0.49)0.31 (–0.04 to 0.59)0.36* (0.02 to 0.63)Values are presented as spearman’s rank correlation coefficients (r) with 95% confidence intervals. r: spearman’s rank correlation coefficients. *p<0.05, **p<0.01. ADF/APF ROM: ankle dorsiflexion/ankle plantarflexion range of motion; KES: knee extensor strength; APS: ankle plantarflexion strength; BMS: back muscle strength; FFM: fat-free mass; SMI: skeletal muscle index. Correlations between mSEBT performance and physical function parameters (n=35) Values are presented as spearman’s rank correlation coefficients (r) with 95% confidence intervals. r: spearman’s rank correlation coefficients. *p<0.05, **p<0.01. ADF/APF ROM: ankle dorsiflexion/ankle plantarflexion range of motion; KES: knee extensor strength; APS: ankle plantarflexion strength; BMS: back muscle strength; FFM: fat-free mass; SMI: skeletal muscle index. The results of the multiple linear regression analyses are presented in Table 3. In the anterior direction, the regression model was statistically significant (F (2, 32)=5.133, p=0.012, R2=0.243, adjusted R2=0.196). No independent variables were identified as significant predictors. Multicollinearity was detected in the initial model for the posterolateral direction because the condition index exceeded 30 and variance concentrations were observed within the same component. Therefore, a modified model including APS, KES, leg FFM, and SMI was constructed. The modified model no longer showed evidence of multicollinearity according to the predefined diagnostic criteria. No independent variables were identified as significant predictors in the final model (F (4, 30)=3.120, p=0.029, R2=0.294, adjusted R2=0.200). Multicollinearity was also detected in the initial model for the posteromedial direction because the condition index exceeded 30 and variance concentrations were observed within the same component. Therefore, a modified model including APS, KES, leg FFM, and SMI was constructed. The modified model no longer showed evidence of multicollinearity according to the predefined diagnostic criteria. The final model was not statistically significant (F (4, 30)=2.507, p=0.063, R2=0.251, adjusted R2=0.151), and no independent variables were identified as significant predictors. Table 3. Multiple regression analysis of mSEBT performance (n=35)B (95% CI)βAnteriorAPS0.02 (0.00 to 0.04)0.33Leg FFM1.80 (–0.40 to 3.99)0.27PosterolateralAPS0.01 (–0.03 to 0.05)0.13KES0.01 (–0.01 to 0.03)0.25Leg FFM2.73 (–0.38 to 5.83)0.31SMI0.00 (–0.06 to 0.07)0.03PosteromedialAPS0.01 (–0.03 to 0.06)0.13KES–0.00 (–0.02 to 0.02)–0.04Leg FFM1.47 (–1.96 to 4.90)0.16SMI0.05 (–0.02 to 0.12)0.35Values of B are presented with 95% confidence intervals. B: partial regression coefficient, β: standardized partial regression coefficient. KES: knee extensor strength, APS: ankle plantarflexion strength; FFM: fat-free mass; SMI: skeletal muscle index. Multiple regression analysis of mSEBT performance (n=35) Values of B are presented with 95% confidence intervals. B: partial regression coefficient, β: standardized partial regression coefficient. KES: knee extensor strength, APS: ankle plantarflexion strength; FFM: fat-free mass; SMI: skeletal muscle index. DISCUSSION This study investigated the association between mSEBT performance and physical function in healthy university students. Bivariate correlation analysis showed that mSEBT performance was associated with multiple physical function factors. However, multiple regression analysis identified only limited independent contributions of these variables, despite some statistically significant models. Consistent with previous studies11, 12), these findings indicate that dynamic balance performance cannot be explained by a single factor, but reflects the combined influence of multiple physical function components. Although no significant independent predictors were identified in the multiple regression analyses, APS and leg FFM showed consistent significant associations with mSEBT performance across all directions in the bivariate analyses. Previous studies using the Y-Balance Test, which has similar movement characteristics to the mSEBT, have reported significant associations between dynamic balance performance and APS22). Therefore, the present findings partially support previous evidence and suggest that APS may contribute to dynamic balance performance. In addition to muscle strength, muscle mass may also contribute to dynamic balance performance. Muscle mass has been reported to be associated not only with muscle strength but also with neuromuscular control mechanisms23). Additionally, dynamic balance performance may depend on coordinated muscle activation and appropriate timing of force production24). Therefore, the association observed between mSEBT performance and leg FFM in this study may be mediated, at least in part, by neuromuscular control capacity. However, because neuromuscular control was not directly evaluated in the present study, this interpretation remains hypothetical. Future studies should incorporate surface electromyographic analyses and balance perturbation tests to better understand neuromuscular responses. In particular, muscles such as the gluteus medius and soleus may play important roles in dynamic balance performance11, 22). The lack of independent predictors in the multivariate analyses indicates that these relationships are not independent and may reflect the combined influence of multiple interrelated physical function factors rather than the effect of a single dominant factor. Furthermore, the relatively low explanatory power of the regression models (adjusted R2 » 0.2) suggests that mSEBT performance is influenced by multiple interacting components beyond those included in the present analyses, and that a substantial proportion of variance remains unexplained. These findings have implications for interpreting dynamic balance assessment in university students who are at risk of injury in daily life, sports activities, and commuting1, 2). Because dynamic balance is an important intrinsic risk factor for injury5), interpreting mSEBT performance based solely on a single physical function indicator may lead to an oversimplified understanding of balance impairment and potentially influence the selection of intervention targets. Accordingly, the use of the mSEBT to assess dynamic balance in university students should be interpreted by considering multiple physical function factors. This study had several limitations. First, its cross-sectional design precludes causal inferences. Second, the relatively small sample size (n=35) may have affected the stability and reliability of the multiple regression models. In particular, regression analyses with multiple candidate independent variables are susceptible to overfitting and unstable coefficient estimates when the sample size is limited. In addition, candidate variables for the regression analyses were selected based on bivariate correlation analyses, which may have increased the risk of selection bias and type I error. Furthermore, multiple correlation analyses were conducted without adjustment for multiple comparisons, and therefore the results should be interpreted with caution. The relatively low explanatory power of the regression models (adjusted R2 » 0.2) suggests that only a limited proportion of variance in mSEBT performance was explained by the variables included in the model. This study was exploratory in nature and conducted with a relatively small sample size. Therefore, the findings should be interpreted with caution. Third, the lack of biomechanical and neuromuscular measurements, such as motion analysis or electromyography, limits the interpretation of the underlying mechanisms of the observed associations. Body composition measurements using bioelectrical impedance analysis may also have been influenced by factors such as hydration status, meal timing, time of day, and other uncontrolled conditions. In addition, bilateral measurements were averaged to represent overall physical function. However, this approach may have obscured potential asymmetries between lower limbs. Potential confounding factors such as sex, height, and sport participation were not controlled, which may have influenced the observed associations and may limit the generalizability of the findings beyond healthy university students, such as clinical populations or elite athletes. Finally, only directional mSEBT scores were analyzed. Although this approach allowed for the identification of direction-specific associations, it may limit direct comparisons with studies reporting composite scores10). Future studies should use longitudinal designs with larger sample sizes, incorporate neuromuscular and movement analyses to clarify the determinants of dynamic balance and improve injury prevention strategies, and employ theory-driven variable selection or penalized regression methods. This study suggests that the mSEBT performance in healthy university students is associated with multiple physical function factors. Therefore, dynamic balance performance may reflect the combined influence of several physical functional components. Funding This work was supported by JSPS KAKENHI Grant Number JP24K20382. Conflict of interest The authors declare no conflicts of interest. Funding Statement This work was supported by JSPS KAKENHI Grant Number JP24K20382. REFERENCES