This study was conducted at the Sports Rehabilitation Medicine Center of Beijing Sport University, where 60 Chinese nationals were recruited through various methods, including poster and online recruitment. Ultimately, 48 participants completed the entire testing process. Age groups were divided at 15-year intervals (group A: 18–30 years old, group B: 31–45 years old, group C: 46–60 years old, and group D: 61–75 years old) based on the critical postural control decline threshold at age 45 and the WHO-defined elderly boundary at age 60 (22, 23).

To mitigate practice effects documented in CDP protocols (24), all participants completed two familiarization trials on the Bertec^® system prior to data collection. Formal test data were exclusively obtained during the third session. At the initial visit, after detailed explanation of experimental procedures and risks, written informed consent was obtained. Subsequently, the first familiarization trial commenced, the process including SOT and Rhythmic Weight Shift Test (RWS) (Figure 1).

Participants stood barefoot on the force platform in a standardized position. A safety harness—suspended from the dome mount—was secured at the waist to prevent falls without restricting movement (Figure 2). An experimenter stood behind the participant to intervene if necessary. In actual testing, ambient light in the lab is controlled to ≤ 5 lux using blackout curtains and switched ceiling lights to minimize visual distractions. All instructions were delivered in standardized Mandarin by the same experimenter:

There were two postural control tests involved in this study, Sensory Organization Test (SOT) and Rhythmic Weight Shift (RWS). The SOT composite score (CS) was selected as the primary outcome given its recognized validity in tests of postural control. The remaining indicators serve as secondary outcomes.

Statistical analyses were performed using jamovi (version 2.3.28). The normality of continuous demographic variables (age, height, weight, BMI) was assessed with the Shapiro-Wilk test, and homogeneity of variances was evaluated using Levene's test. The association between age and postural control parameters was first analyzed with Spearman's rank correlation. For group comparisons, either standard one-way ANOVA (Fisher's ANOVA) followed by Tukey's HSD post hoc test, or Welch's ANOVA followed by the Games-Howell post hoc test, was employed based on whether the assumptions of normality and homogeneity of variances were met. Categorical variables (e.g., sex) were analyzed with Pearson's chi-square test. Data from the SOT and RWS are presented as mean ± standard deviation. Additionally, one-sample t-tests (for normally distributed data) or Wilcoxon signed-rank tests (for non-normal data) were used to compare our SOT composite scores against existing Danish normative values (20).

Sixty participants were initially enrolled through community posters and offline recruitment channels. After screening, 48 completed all study protocols. Exclusions comprised:

Among the final cohort (n = 48), eight participants had corrected myopia (Corrected visual acuity ≥ 1.0 LogMAR with glasses), and eight reported stable chronic conditions:

No medication-related adverse effects (e.g., dizziness or hypotension) were observed during testing, suggesting negligible pharmacological interference with postural control measures.

The demographic information is summarized in Table 2. In addition to age (p < 0.01), these four groups also differed significantly in terms of height (p = 0.009), and a post hoc test revealed that subjects in Group A were significantly taller than those in Groups C and D. Postural control data in this study have been standardized according to height, so comparisons between groups are valid and meaningful.

This study identified significant age-related declines in postural control and sensory integration across age groups. Spearman's rank correlation analysis demonstrated strong negative associations between age and CS (rs = −0.532, p < 0.001), VIS, (rs = −0.530, p < 0.001)and VEST, (rs = −0.430, p = 0.002) (Table 3). Notably, more challenging SOT conditions (SOT4–6) exhibited steeper age-dependent deterioration (rs = −0.589 to −0.478, p < 0.001), whereas no significant age effects were observed for simple postural control tasks (SOT1–3) and SOM.

ANOVA revealed significant between-group differences in SOT (Table 4), particularly for dynamic balance tasks (SOT2-SOT6) and CS (SOT2: p = 0.012; SOT3: p = 0.017; SOT4: p < 0.001; SOT5: p = 0.009; SOT6: p = 0.002; CS: p = 0.001). As detailed in Figure 3, post-hoc analyses indicated that younger groups (A and B) consistently demonstrated superior postural stability compared to middle-aged and older groups (C and D). Specifically: In SOT2, Group C scored significantly lower than both Group A (p = 0.030) and Group B (p = 0.018). In SOT4, Groups C and D exhibited marked reductions relative to Groups A (both p < 0.001) and B (p = 0.005 and p = 0.008, respectively). In SOT5, Group A outperformed both Group C (p = 0.029) and Group D (p = 0.030). The Composite Score (CS) showed a similar pattern, with Group A significantly higher than both Group C (p = 0.010) and Group D (p = 0.012), and Group B outperforming Group C (p = 0.010) and Group D (p = 0.011). For SOT6, which was analyzed using the Games-Howell test due to heteroscedasticity, Groups A and B maintained a significant advantage over Groups C (p = 0.025 and p = 0.009) and D (p = 0.022).

Significant differences were also observed in sensory ratio scores, as summarized in Figure 4. The Visual Score (VIS) showed Group A scoring higher than Group D (p = 0.001), and Group B outperforming both Group C (p = 0.043) and Group D (p = 0.027). The Vestibular Score (VEST) was higher in Group A than in Group D (p = 0.044), and the Preference Score (PREF) was elevated in Group B compared to Group D (p = 0.037).

RWS results indicated poorer directional control in both the mediolateral (p = 0.002) and anteroposterior (p < 0.001) directions in Groups C and D compared to the younger groups. The variance is primarily observed in the DCL (Figure 5), both in the mediolateral and anterior-posterior directions. In the mediolateral test, significant Intergroup differences were observed in DCLS (P = 0.020), DCLF (P < 0.001), DCLC (P = 0.002), and MVLF (P = 0.032). Tukey's HSD post hoc test revealed significant differences in DCLC between Group A and Group C (p = 0.044), Group A and Group D (p = 0.004), Group B and Group D (p = 0.032). In the anteroposterior test, significant intergroup differences were identified in DCLS (P = 0.034), DCLM (P = 0.001), DCLF (P < 0.001), and DCLC (P < 0.001). However, no intergroup differences in MVL were observed. Following this, Tukey's HSD post hoc test was conducted. For DCLC, significant intergroup differences were noted between Group A and Group C (P < 0.001), Group A and Group D (P = 0.002).

To explore potential cross-cultural variations, an exploratory comparative analysis was performed between the normative data from our Chinese cohort and existing data from a healthy Danish population (20). When comparing broadly age-matched groups (e.g., our Group A: 23.31 ± 2.18 years vs. the 20-29-year-old Danish cohort), notable variations were observed in several SOT parameters (Table 5). The most pronounced variations were identified among individuals in their 20s and 30s. These preliminary findings suggest that postural control characteristics may differ between these populations, highlighting the need for further investigation.

A significant negative correlation was observed between advancing age and the composite score (CS) (rs = −0.532). ANOVA results revealed significant differences among age groups for SOT condition 2, 4, 5, and 6, as well as for CS. Post hoc analyses indicated that ES were significantly higher in Groups A and B compared to Groups C and D, particularly in SOT condition 4, 5, and 6. Significant differences were observed between younger adults (Groups A and B) and middle-aged/older adults (Groups C and D). For instance, CS decreased from 85.2 in Group B to 79.1 in Group C (p = 0.01), suggesting a potential acceleration in postural control decline during the transition to middle age.

These findings are consistent with the recognized trajectory of accelerated postural control decline after age 45, driven by convergent degeneration in neuromuscular, articular, and sensory systems (26). Specifically, selective atrophy of fast-twitch muscle fibers—which deteriorate at ~3 × the rate of slow-twitch fibers after age 50—compromises rapid balance responses (22). In parallel, estrogen deficiency during menopause accelerates articular cartilage thinning and impairs joint proprioception (27). This decline represents a transition from adaptive compensation (where the CNS recruits ancillary neural resources before age 45) to progressive decompensation. It is important to note that our study was not powered to conduct sex-stratified analyses. However, the observed decline in middle age (Group C) warrants consideration of sex-specific factors in future research. For instance, in women, the menopausal transition involves endocrine alterations that can undermine musculoskeletal integrity (28, 29), potentially compounding age-related risks (30).

The most pronounced differences in ES emerged in SOT condition 4–6 compared to 1–3. SOT 4–6 employs a sway-referenced platform that disrupts proprioceptive input, thereby increasing reliance on visual and vestibular systems for stability maintenance. Consequently, older adults exhibit marked postural control deficits under proprioceptive perturbation—a critical sensory modality for detecting joint/limb kinematics and body orientation in space (31, 32). The impact of aging on proprioception has become a focal point in developmental and geriatric research (33). As established by Yang et al., proprioception serving as the foundation for motor control, demonstrates joint-specific developmental trajectories rather than generalized maturation (34). Proprioceptive acuity essential for postural control, particularly at the ankle, reaches functional maturity by age 3–4 years (35), and peaks in adulthood. As the primary weight-bearing interface during stance, the ankle complex provides the final neuromuscular adjustment for upright stability, rendering ankle proprioception clinically pivotal. Critically, evidence indicates progressive decline in ankle proprioceptive sensitivity after age 50 (34). Thus, age-related deterioration of proprioceptive function, especially at the ankle, is proposed as a key contributor to the reduced balance performance and composite scores observed in SOT conditions 4–6.

Sensory Scores (SS) reflect the capacity to utilize individual sensory systems for postural stability. Significant negative associations were observed between advancing age and both VIS (rs = −0.530, p < 0.001) and VEST (rs = −0.430, p = 0.002). ANOVA results indicated significant between-group differences in all scores except the SOM.

Vision is fundamental to postural control. Stable visual input from fixed environments reduces postural sway, whereas dynamic visual scenes may induce instability. When proprioception is compromised—such as during sway-referenced platform conditions or pathological states—visual dependence intensifies. This effect is often amplified in older adults, who may exhibit greater susceptibility to visual perturbations than younger individuals (36, 37). Consistent with this, our data showed lower VIS scores in middle-aged and older participants, suggesting a decline in the efficacy of utilizing visual information for balance with age. Crucially, visual reliance should not be conflated with VIS. Multiple studies report a paradoxical age-related phenomenon: a diminished ability to use visual input efficiently for balance may coincide with an increased weighting of visual information in sensorimotor processing (38). Ana Faraldo-García et al. documented a U-shaped trajectory, in which the contribution of vision to postural control declines to its lowest point between ages 40–49 before gradually increasing thereafter (39). Collectively, these findings suggest that older adults may exhibit heightened dependence on visual cues, underscoring the critical—and potentially increasingly dominant—role of vision in aging postural control systems.

Vestibular function typically begins to decline during middle age, though this deterioration often remains subclinical due to robust neural compensation. Comprehensive assessment usually requires specialized protocols—such as rotational chair testing—that quantify vestibular-ocular reflex (VOR) integrity by measuring extraocular muscle responses (40). Our SOT data suggest a progressive attenuation of vestibular contributions to postural stability beginning in midlife. Compromised vestibular processing may manifest as delayed transition movements (e.g., increased sit-to-stand latency) (41), highlighting the clinical value of functional assessments like the Timed Up and Go (TUG) test. Critically, while vestibular pathology can cause vertigo, age-related postural control decline rarely stems exclusively from vestibular dysfunction (40). Instead, it likely involves a multisystem failure including degraded sensory integration, impaired central processing, and compromised motor execution. This multifactorial etiology may explain why isolated vestibular interventions often yield limited functional improvement in geriatric populations.

Visual Preference (PREF) scores are derived from the formula (Condition 3 + Condition 6)/(Condition 2 + Condition 5), which compares balance performance when visual input is altered vs. absent. This score reflects the ability to suppress misleading visual information. It is widely recognized that as individuals age, especially middle-aged and older adults, they tend to rely more heavily on visual input to maintain balance (36). This increased reliance can make them more susceptible to being misled by inaccurate visual information, thereby potentially increasing fall risk. Moreover, the ability to resolve such sensory conflict is closely linked to cognitive function. Age-related neurological changes, such as brain atrophy and white matter lesions, can impair cognitive abilities, particularly those associated with the prefrontal lobes (42). Given that cognitive performance typically peaks around age 35 (43), the observed differences in PREF scores between Groups B and D in our cohort may be related to these age-related cognitive changes.

The RWS task represents a simple fixed-speed visual tracking paradigm. Significant between-group differences were observed in DCL, particularly during fast-tracking conditions. For instance, in left-right fast oscillation (DCLF), Group A demonstrated better performance than Group D (92.8 ± 2.32 vs. 84.5 ± 5.83, p < 0.001), which is consistent with the expected superior motor control in younger adults.

The observed age-related decline in DCL, which reflects motor control precision, may be attributed to several interconnected mechanisms reported in the literature: (1) Peripheral degradation, including altered nerve conduction velocity and proprioceptive shifts; (2) Central atrophy, involving gray and white matter loss in motor cortices and associated network reorganization (44); (3) Neurochemical decline, such as dopaminergic depletion that can impair motor skills; (4) Reduced neural compensation efficiency, where over-recruitment of prefrontal-basal ganglia circuits may create a supply-demand imbalance vulnerable to age-related decline (45). These mechanisms collectively could help explain the reduced DCL performance in middle-aged and older adults during rapid oscillations, where precise control is paramount.

While DCL showed consistent age-related declines in both coronal and sagittal planes, the magnitude of decline differed biomechanically. Coronal plane control primarily relies on hip abductor/adductor synergy, whereas sagittal control is dominated by ankle strategies. The preferential atrophy of fast-twitch fibers in proximal musculature may accelerate coronal plane dysfunction (46). This is compounded by evidence suggesting ~30% greater hip joint shear forces during coronal vs. sagittal motion (47), which could accelerate articular cartilage wear and attenuate proprioceptive feedback. This biomechanical context aligns with the exclusive negative correlation between aging and coronal plane velocity found in our analyses and may underlie the stronger age-correlated decline in MVL observed in the coronal plane.

Compared to published Danish normative data (20), Groups A and B in our Chinese cohort demonstrated significantly higher SOT scores. This observed difference may be attributed to several factors. Methodologically, our protocol included ≥2 full practice trials prior to formal testing, which may have enhanced task familiarity and introduced procedural learning effects. Furthermore, a substantial proportion of our younger participants (36% in Group A, 29% in Group B) reported engaging in structured exercise regimens (>3 sessions/week), suggesting generally higher physical activity levels that could contribute to superior performance. It is also plausible that culturally influenced lifestyle factors play a role. For instance, whereas cycling culture in Denmark may promote sagittal-plane stability (48), the prevalent use of public transportation (e.g., buses and subways) in China could provide frequent, incidental exposure to multi-directional perturbations, potentially serving as a form of daily balance training in dynamic environments. In summary, the higher scores observed in our cohort are likely driven significantly by protocol-specific training effects and potentially higher baseline physical activity, rather than representing an inherent physiological advantage. These findings underscore the importance of standardizing acclimatization procedures and accounting for lifestyle factors in cross-cultural posturography research.

The clinical implications of our findings are two-fold. First, the established normative data provide a crucial reference for clinicians to objectively assess sensory integration function in Chinese adults across different ages. By comparing a patient's performance against these benchmarks, particularly the sensory ratios that pinpoint reliance on visual, vestibular, or somatosensory inputs, practitioners can identify specific deficits beyond what traditional balance tests reveal. Second, the observed differences between our study and Western data underscore the necessity of using population-specific norms for accurate risk assessment. Employing Western criteria could potentially lead to either over- or under-estimation of fall risk in Chinese individuals. Therefore, this work provides the foundational evidence needed to develop targeted, evidence-based fall prevention strategies that are sensitive to the specific sensorimotor profile of the Chinese population.

This study has several limitations. The modest sample size (N = 48) and the upper age limit of 72 years limit the generalizability of our findings, particularly regarding the oldest-old population. The non-random recruitment method may also introduce selection bias. Furthermore, detailed physical activity profiles and stricter health screening were not implemented; including these in future studies would help control for confounding variables. Finally, the acclimatization protocol (two practice trials) may have been insufficient for older adults, potentially attenuating the observed age-related differences. Therefore, our results should be interpreted as exploratory and might underestimate the true magnitude of postural control decline in the broader Chinese population.