Effects of Dual-Tasking on Center-of-Pressure Dynamics and Spectral Balance Control

Abstract

Cognitive tasks play a pivotal role in posture control among young adults. This study examined how concurrent cognitive tasks alter balance stability and sensory integration during single-leg stance by analyzing center-of-pressure trajectories and wavelet spectra to elucidate the neurobehavioral mechanisms underlying dual-task balance degradation. A cohort of 24 young adults completed both single postural control tasks and dual cognitive–postural tasks on a force plate. COP data and wavelet decomposition energy were computed and analyzed. The results revealed significant differences between the dual-task and single-task groups for Lxy, Ly, Vxy, and Vy (p < 0.05). Energy content analysis showed that the dual-task group had significantly different energy ratios across four frequency bands along the x-axis (p < 0.05). Our findings showed that dual-task conditions impair postural control in young adults, increasing anteroposterior sway and altering mediolateral energy patterns. This suggests a shift toward proprioceptive reliance during cognitive division, revealing cognitive–postural interference. These results support using dual-task assessments for fall risk evaluation and inform interventions for populations requiring cognitive–motor integration.

1. Introduction

Postural control is a complex process that involves the real-time and dynamic collaboration of the sensory system (vision, vestibular sense, proprioception), central nervous system (attention distribution, sensory information integration), and motor system (musculoskeletal response) [1,2]. As an adaptive process, it can be adjusted in real time based on task difficulty and environmental changes. Given that people often engage in concurrent cognitive–motor tasks in daily life, the resulting competition for attention can destabilize posture, making this interference effect a central focus in postural control research.

Traditional posture control models once held that maintaining balance was an automated process that did not require conscious participation [3,4,5]. However, an increasing number of studies have shown that higher-order cognitive functions, especially executive functions and attention, play a core role in posture control. Tramontano et al. [6] investigated running speed under dual-task conditions and observed reduced speeds in both cognitive and balance tasks, indicating mutual interference. Similarly, another study concluded that upright control capabilities diminish under dual-task conditions [7]. Other studies have used neuroimaging techniques to detect increased activation in the prefrontal and sensorimotor cortices, which may reflect the brain’s need to integrate additional sensory information during dual-task scenarios [8,9]. Therefore, the introduction of concurrent cognitive tasks may “divert” attentional resources that would otherwise be allocated to postural control, potentially altering quantitative measures of standing postural stability. These alterations are often quantified by analyzing the dynamic changes in the COP, a key metric in postural control research.

Pressure centers (such as swing path, speed, amplitude and frequency) are the gold standard for evaluating posture stability [10] and mainly capture overall steadiness. More advanced indicators, such as the frequency-domain characteristics under wavelet transform, are considered to reflect the complexity and adaptability of the neuromuscular system. In recent years, researchers have begun to move beyond traditional analysis of COP time-domain metrics—such as root mean square distance and total path length—to focus instead on frequency-domain characteristics [11]. Through spectral analysis, the COP signal can be decomposed into different frequency components. A study suggests that low-frequency components may be associated with conscious cognitive control, whereas high-frequency components more likely reflect automatic reflexive regulation at the spinal level [12]. Thus, “Spectral Balance”—that is, the relative distribution of energy between high and low-frequency bands—offers a novel and powerful perspective for understanding postural control strategies under dual-task interference. Although the impact of dual tasks on postural control has been widely studied in older adults, research on young adults—whose physical functions are at their peak—remains relatively limited. It is generally assumed that their postural control systems possess sufficient redundancy to resist cognitive interference. However, recent evidence suggests that even healthy young individuals may exhibit subtle yet significant changes in postural control as cognitive demands increase, changes that are difficult to capture using traditional time-domain analysis [13].

Therefore, this study employed an integrated time–frequency domain approach to analyze COP trajectories. By comparing the motor control characteristics under single-task (quiet standing) and dual-task (standing while performing cognitive tasks) conditions, it explored the impact of dual-tasking on the dynamic and spectral characteristics of COP in young adults, aiming to reveal the neurobehavioral mechanism of cognitive–postural interference. We hypothesized that dual-tasking would lead to increased postural instability, as shown by longer COP trajectory and higher mean velocity in time-domain analysis, and that cognitive tasks would alter frequency band energy, inducing changes in sensory integration and motor control strategies.

2. Research Methods

2.1. Participants

This study was approved by the Research Ethics Committee of Soochow University (No. ECSU-2019000209). Participant recruitment took place from 27 December 2021 to 1 October 2024, and all procedures were performed in full compliance with both institutional and international ethical standards.

The sample was recruited using a convenience sampling method from the local university population. Participants were selected as healthy young adults (mean age: 22.6 ± 2.3 years, mean height: 174.9 ± 8.4 cm, mean body weight: 70.1 ± 10.9 kg) based on strict inclusion criteria. All participants exhibited normal cognitive function, with no history of neurological or muscular impairments. Additionally, none had sustained lower limb joint injuries within the six months preceding the study, and all refrained from vigorous physical activity for at least 24 h prior to the experiment. The sample size was calculated using G*power software (Version 3.1). With reference to a prior study [14], a sample size of 12 was determined (effect size = 0.25, α = 0.05, power = 0.95); however, 24 participants were enrolled to compensate for potential attrition.

2.2. Experimental Setup

The experiment utilized a three-dimensional force platform (Model 9281A, KISTLER, Winterthur, Switzerland), measuring 60 cm × 40 cm. The COP signal was derived from the ground reaction force (GRF) data collected by the force plate at a sampling frequency of 1000 Hz. The data were exported through a Datalog data box.

2.3. Experimental Procedure

The experiment comprised two conditions: single task (ST) and dual task (DT). In both conditions, participants were required to maintain a quiet, single-leg stance on a force platform for 60 s (Figure 1), with arms crossed over the chest and gaze fixed on a visual target. The distance between the participant and the target was 200 cm (for single task and dual task). The order of the conditions (single task or dual task) was randomized across participants using a blocked procedure. All participants performed two trials for each condition, with the order randomized and a rest period of at least one minute provided between trials.

2.3.1. Single Task

In the single-task condition, participants stood barefoot on their dominant leg on the force plate, with arms crossed over the chest and gaze fixed on a black dot. They were instructed to minimize body sway and maintain a stable upright posture throughout the 60-s trial.

2.3.2. Dual-Task

In the dual task condition, participants performed the same postural task while simultaneously engaging in a cognitive arithmetic task. Before the trial, they were instructed to prioritize the cognitive task. Based on previous research, the cognitive task chosen was an arithmetic task [15,16,17]. Participants mentally subtracted 7 from a randomly chosen number between 300 and 900. Errors were not corrected during the task.

2.4. Data Processing

The raw COP signal was resampled to 100 Hz. Research has shown that 95% of the COP signal was concentrated below 6.25 Hz [18]. The original COP signal was filtered using a 2nd-order Butterworth low-pass filter with a cutoff frequency of 12.5 Hz. Data analysis was performed using MATLAB (Version R2021a; The MathWorks, Inc., Natick, MA, USA). Time-domain indicators included total trajectory length Lxy, x-axis trajectory length Lx, y-axis trajectory length Ly, unit area trajectory length Lxy/Area, envelope area, average velocity Vxy, x-axis average velocity Vx, and y-axis average velocity Vy.

Frequency-domain processing was performed using a 9-level Symlet-8 discrete wavelet transform. As shown in Figure 2, the original signal was decomposed layer by layer during the signal reconstruction process. Each decomposition divided the signal into low-frequency and high-frequency parts. Immediately after this, the low-frequency signal obtained from the decomposition was again decomposed to obtain two parts. This process was repeated for nine iterations.

As shown in Figure 3, the COP data were decomposed into four frequency bands: moderate frequency (1.56–6.25 Hz), low frequency (0.39–1.56 Hz), very-low frequency (0.10–0.39 Hz), and ultra-low frequency (<0.10 Hz) [19]. The energy content of each frequency band to the total COP signal was calculated.

The temporal energy for each frequency band was calculated using the following formula:

$$E(i)={\displaystyle \frac{1}{N}}\sum _{k=0}^{N-1}{\left|X\left[k\right]\right|}^2$$

$E_{\left(i\right)}$ is the total energy of frequency band (i) of the COP signal. N is the total amount of data in frequency band a. X is the temporal data set, and k denotes the k-th data in X set.

The ratio of energy to total energy for each frequency band is calculated with the following formula:

$$E\mathrm{\%}(a)={\displaystyle \frac{E\left(a\right)}{E_T}}\times 100\mathrm{\%}$$

$E\%\left(a\right)$ denotes the percentage of the COP signal a-band energy in the total energy. $E\left(a\right)$ is the total COP signal a-band energy, and $E_T$ is the total COP signal energy.

2.5. Statistical Analysis

All data were analyzed using IBM SPSS Statistics for Windows (Version 21.0) and presented as means ± standard deviations (M ± SD). Initially, normality was assessed using the Shapiro–Wilk test for both single-task and dual-task conditions. When the data satisfied the normal distribution assumption, paired sample t-tests were used for significance analysis, and the Bonferroni correction was applied to account for multiple comparisons across the dependent variables. The statistical significance level was set at p < 0.05, with p < 0.001 indicating a highly significant difference.

3. Results

3.1. COP Indicators

The results indicated that the data from both conditions were normally distributed (p > 0.05), supporting the use of parametric statistical methods for subsequent analyses.

As shown in

Table 1. Comparison of envelope area indicators in single-task and dual-task conditions.

Indicators	Group	M ± SD	t	p
Area (mm2)	Single Task	472.21 ± 145.22	1.592	0.128
	Dual Task	398.24 ± 130.20
Lxy/Area (mm−1)	Single Task	9.65 ± 3.32	−2.662	0.015 *
	Dual Task	12.31 ± 3.61

Note: “*” means p < 0.05.

, compared with the single-task condition, the dual-task condition showed significant differences in Lxy/Area (p < 0.05). The dual-task condition demonstrated a larger Lxy/Area, with a moderate effect size (p = 0.015, d = 0.595). This finding highlighted a specific alteration in dual-task performance. However, these area indicators showed no significant differences between conditions (p > 0.05).

As Figure 4 displays, compared with the single-task condition, the dual-task condition showed significant differences in Lxy, Vxy, Ly and Vy (p < 0.05). Specifically, the dual-task condition demonstrated a larger Lxy and Ly (p = 0.007, d = 0.674; p = 0.012, d = 1.024), respectively. The effect size for Ly was large, indicating a substantial increase in the COP trajectory length, while the increase in the area covered (Lxy) was of medium magnitude. Furthermore, the Vxy of the dual-task condition was higher than that of the single-task condition (p < 0.001, d = 0.625; p < 0.001, d = 0.943), also showing medium to large effects. The exceptionally large effect for Vxy strongly suggests a meaningful increase in overall sway velocity.

3.2. Energy Content

The energy content of different frequency bands in both the x and y directions is presented in Figure 5. In the x-axis direction, significant differences were observed in the four frequency bands between the single-task and dual-task conditions (p < 0.05). Under dual-task conditions, the energy content in the moderate-, low-, and very-low-frequency bands is relatively low (p = 0.005, d = 0.711; p = 0.002, d = 0.785; p = 0.005, d = 0.715), reflecting medium effect sizes, whereas the ultra-low-frequency band showed a significant increase in energy (p = 0.001, d = 0.838), indicating a large effect. This pattern suggests a clear shift in energy distribution toward the lowest frequency during dual-task performance in the x-direction. In contrast, no significant differences in energy content were observed along the y-axis across any of the frequency bands (p > 0.05).

4. Discussion

This study investigated the effects of dual-task processing on the dynamic and spectral characteristics of the COP in adolescents. The results demonstrated that dual-task conditions induce significant changes in COP dynamics. Analysis revealed significant differences between the single-task and dual-task groups across multiple parameters, including the total trajectory length (Lxy), trajectory length per unit area (Lxy/area), average velocity (Vxy), y-axis trajectory length (Ly), and y-axis average velocity (Vy). The increase in total trajectory length suggests a reduction in postural stability during dual-task conditions; the alterations in trajectory length per unit area may reflect increased complexity or heightened interference in postural control under cognitive load. The y-axis trajectory length and average velocity provide detailed information on postural control in the y-axis direction, highlighting changes in postural fluctuations and adjustment speed along this axis. Overall, these changes suggest that dual-task conditions have a significant impact on postural control, potentially indicating a significant reduction in postural control capability. This result is consistent with previous studies. For instance, Ren Jie et al. [20] investigated the effects of cognitive tasks on visual control during standing posture and found that such tasks inhibit visual function, underscoring the shared central processing system for cognitive and balance functions. This hints at a cognitive–motor conflict in the brain when performing concurrent cognitive and balance tasks, which indirectly explains the reduced visual contribution and diminished postural control ability observed in our study. Furthermore, another study using functional near-infrared spectroscopy (fNIRS) reported significant activation in the prefrontal cortex during dual-task performance [21], indicating the crucial role of this region in integrating cognitive and postural control, irrespective of the theoretical framework applied. From a neural mechanism perspective, the prefrontal cortex may be related to attention allocation and the integration of visual and proprioceptive information [22,23].

However, some studies report that cognitive tasks may have no significant effect or even a facilitatory influence on postural control [24], which may be attributable to differences in task complexity. Increased cognitive demand has been shown to alter balance performance, with varying effects depending on task difficulty [25]. Increased task complexity can reshape attentional allocation, and more difficult tasks may consume greater cognitive resources, thereby interfering with postural performance [25]. In healthy young adults, increased cognitive load induces subtle postural changes that are discernible yet elusive to traditional time-domain analysis. Fortunately, wavelet analysis offers a promising approach to enhance our understanding of postural control characteristics under dual-task conditions [11,12]. This method decomposes COP signals into distinct frequency components, with each associated with a specific postural control subsystem [13]. According to previous studies, different frequency bands in COP data correspond to various physiological mechanisms: muscles and proprioception, the cerebellum, the vestibular system, and vision are linked to specific frequency ranges, from high to low [13,26]. Consequently, we leveraged wavelet analysis to examine how dual-tasking alters the dynamic and spectral properties of COP in adolescents. This approach provided a novel perspective for uncovering the specific neurophysiological mechanisms of postural control in the developing brain under cognitive challenge.

In this study, we leveraged wavelet analysis to examine how dual-tasking alters the dynamic and spectral properties of COP in adolescents. We observed an increase in low-frequency energy content in the anterior–posterior direction. Analysis of frequency band energy revealed that, compared to the single-task group, the dual-task group exhibited a decrease in ultra-low-frequency energy along the anterior–posterior axis (x-axis), while energy in other frequency bands increased. These findings indicate that the cognitive task may have induced the relocation of central nervous system (CNS) resources, reflecting heightened nervous system activity. Some studies suggest that the reduction in ultra-low-frequency energy implies a diminished role of vision in postural regulation [19,26], while the increase in energy within other frequency bands indicates greater reliance on proprioceptive and vestibular mechanisms [13]. Reduced visual input causes increased body sway, accompanied by active and passive muscle reflexes [27]. When the head moves, vestibular system amplification maintains head stability, and muscle–joint activities transmit proprioceptive signals to the nervous system [28,29]. The above research results indicate that although the influence of the vestibular and proprioception has been enhanced, they cannot fully compensate for visual defects, resulting in an overall decline in stability. In contrast, energy content in the left-right direction (y-axis) showed no significant differences, although the overall trends were similar. This suggested that the processing capacity of the central system was limited and could not meet the demands of postural control in the left-right direction [20]. We speculated that the central nervous system prioritized the allocation of limited cognitive resources to maintain stability in the x-axis. This might have come at the cost of sufficient control in the y-axis, leading to the observed increase in sway metrics. This directional difference in control could be further explained by joint biomechanics: the hip and ankle joints exhibit better flexibility and neuromuscular control in the sagittal plane (flexion–extension), which corresponds to the x-axis. Under cognitive load, increased muscular tension restricted joint motion in this plane, while the weaker control in the frontal plane (corresponding to the x-axis) became more apparent, resulting in poorer medial-lateral stability(y-axis).

Moreover, the increase in moderate-frequency energy suggests greater involvement of muscular and proprioceptive systems in postural control. In general, faster neuromuscular responses are associated with higher frequency signals, while medium-frequency components reflect muscular and proprioceptive activity. In contrast, vestibular function and visual modulation operate more slowly and stably, corresponding to lower frequency bands. Consequently, wavelet-based decomposition of COP signals can distinguish the contributions of various factors—such as muscle proprioception, neural processing, vestibular input, and vision—across different frequency ranges [13,26]. It could be inferred from this that in the cognitive tasks in this study, the muscle tone of the subjects might have increased when standing upright to reduce limb swinging and joint movement. Vestibular and proprioceptive signals convey balance-related information to the CNS, enabling timely motor adjustments. However, due to compromised visual input, this adjustment process becomes slower, leading to elevated muscle activation. Postural control depends on the integration of sensory inputs—including vestibular, proprioceptive, and visual cues—and the regulation of motor output; impairments in this process are often attributable to dysfunction within these sensory or motor components [30,31]. Although this study did not directly evaluate the contribution of each sensory modality to postural control, frequency-domain analysis offers a means to infer underlying control mechanisms and the relative involvement of system.

Limitations and Future Research Directions

This study provided deeper insights into the postural control mechanisms of young adults across different frequency bands during cognitive tasks, offering a scientific basis for fall prevention interventions. However, there are still some limitations. First, the demographic scope of our sample was narrow, limited to young people with an uneven gender distribution, which restricted the generalization of our research results among a broader population with different ages, genders, or health characteristics. Second, the cognitive task used in this study involved mental subtraction, and although this serves the purpose of dual-tasking, it does not fully reflect the diversity of cognitive tasks that individuals face in the real world. Finally, the scope of our analysis was centered on successfully performed tasks. Consequently, the potential interaction between cognitive performance errors and postural control represents an important and interesting direction for subsequent studies.

5. Conclusions

This study demonstrates that dual-task conditions significantly impair postural control in young adults, as evidenced by increased COP trajectory lengths and velocity, indicating greater sway and reduced stability. Wavelet-based spectral analysis further revealed a redistribution of energy across frequency bands along the x-axis, suggesting altered sensorimotor integration strategies with a shift toward increased reliance on proprioceptive control under cognitive load. These findings not only confirm the presence of cognitive–postural interference even in a young healthy population but also highlight the utility of spectral balance metrics in uncovering subtle neural control adaptations. The results underscore the value of incorporating dual-task paradigms and frequency-domain analyses in future assessments of fall risk and in the design of interventions aimed at improving cognitive–motor integration in both clinical and general populations.