Asymmetry Should Be Considered the Norm, Not the Exception: Neuromuscular Asymmetries in Knee Flexors and Extensor Assessed Through a Multimetric Approach

Abstract

Background:

The level of interlimb asymmetry varies significantly across outcome measures, resulting in poor agreement in categorizing participants as (a)symmetric. Several researchers have discussed the need for an individual approach to data analysis and the need to perform multiple tests.

Hypothesis:

Limb dominance does not consistently influence muscle function, and the direction of asymmetry may show low consistency across different muscle groups and across the various metrics applied to the same muscles.

Study Design:

Cross-sectional.

Level of Evidence:

Level 4.

Methods:

A total of 71 subjects visited the laboratory once to undergo assessments for knee extensors (KEs) and knee flexors (KFs) separately for the dominant and nondominant limbs. Maximal voluntary force (MVF), rate of force development (RFD) at various time intervals (50, 100, 150 ms) and at its peak (RFD_peak), and RFD-scaling factor (RFD-SF) were obtained. Approximate entropy (ApEn), coefficient of variation, and detrended fluctuation analysis (DFAα) of force signal were also calculated during sustained submaximal contractions. Analyses of variance or Wilcoxon signed-rank test assessed differences in each metric between muscle groups and limbs. Kappa coefficients (κ) were calculated to determine the level of agreement for the direction of asymmetry.

Results:

Except for RFD₁₅₀ (d = 0.305) and RFD-SF (r_rb = 0.280) for KF, no differences between dominant and nondominant limb were observed. All κ values were null, slight or fair (all κ ≤ 0.23).

Conclusion:

The dominant limb is not necessarily stronger than the nondominant limb. The direction of asymmetry varies depending on the metric and muscle considered. Our findings corroborate that asymmetries are ubiquitous in physical performance with neuromuscular asymmetries observed in 75% of subjects.

Clinical Relevance:

These findings highlight the necessity of comprehensive assessment across a multimetrics approach, identifying imbalances and tailoring personalized programs to address performance asymmetries effectively.

Asymmetries are a constantly debated topic with often conflicting results. ⁶ When considering asymmetries in performance, unilateral tests are generally used to identify the absolute differences between limbs/muscle groups.^8,34 Then, through various formulas proposed in the literature, the percentage difference is often calculated to quantify the asymmetry level.^9,10,42

As noted by different studies, the average level of interlimb asymmetry vary significantly across different outcome measures,^{4,24,28,32,33} resulting in generally poor agreement in categorizing participants as symmetric or asymmetric. This is further supported by several researchers,^7,35,37 who discuss the need for an individual approach to data analysis and the need to perform multiple different tests. In this regard, maximal strength tests measured through isometric or isokinetic tests are the most commonly used. The rate of force development (RFD) is even more crucial for sports settings, especially when the time available to perform an action/gesture is limited and the contraction times are included between 50 ms and 200 ms.^17,21

Being able to produce movements with a high level of quickness is of great interest also in many areas of daily life, including healthy ageing and fall prevention, or in more common actions, such as the grasping reflex to quickly catch an object that slips from your hands, or the ability to avoid an unexpected obstacle. While RFD can provide valuable insights, it might not fully capture the specific requirements of those scenarios. For the reasons mentioned above and for greater ecological validity, the RFD-scaling factor (RFD-SF) offers an interesting approach as it quantifies the quickness of producing force in rapid contractions targeting submaximal force levels.^2,47 RFD-SF protocol consists of a series of ballistic contractions performed with different submaximal amplitudes. The RFD-SF specifically represents the slope of the linear relationship between the peak torque and the peak RFD obtained in each contraction. This approach was used previously by Boccia et al ¹³ on knee extensors (KEs) and knee flexors (KFs) in soccer players, demonstrating that the RFD-SF could identify an even greater number of subjects with interlimb asymmetry compared with the widely used isokinetic test method. The same research group investigated upper limb asymmetries in a sample of 55 people by examining the differences between elbow flexors and extensors. The resulting agreement was very low, ¹⁴ corroborating that asymmetries are muscle-specific and rarely favor the same side across different muscle performance metrics.

Performance asymmetries could also be seen as asymmetries in motor control, specifically in force unsteadiness.^20,30 Unsteadiness has been defined as the involuntary fluctuations in the motor output within a trial, such as those observed in the movement trajectory. ⁵² In healthy adults, attempts to maintain posture or follow moving targets involve a motor unit activity that is modulated at a low frequency (0.4 Hz, also referred to as common drive). ²⁵ This low-frequency oscillation—an intrinsic feature of muscle contractions—is likely the result of the motor scheme formulated to execute a smooth contraction.

Force fluctuations are influenced predominantly by low-frequency (motor control variability) or high-frequency (influenced by physiological tremor) oscillations.^22,41 To this end, motor control metrics provided valuable information for obtaining a comprehensive assessment of neuromuscular interlimb asymmetries by analyzing the complexity and variability of the force signal around a target level that must be sustained for several seconds. ⁴⁶ In this approach, the most common metrics considered are detrended fluctuation analysis (DFAα) and approximate entropy (ApEn).^43,45 The first is a method designed specifically to deal with the nonstationary nature of signals, while ApEn has been developed to categorize complex systems into different scenarios that include both deterministic chaotic and stochastic processes. The coefficient of variation (CoV - a scale ranging from 0% to 100%) is commonly used to assess steadiness, with lower values representing more force control and higher values less force control. ²² These methods assess the randomness or regularity of the output of a system and its fractal scaling over different timescales. In this regard, Oshita and Yano ³⁸ have examined asymmetry in force fluctuations using a mechanomyogram in the KE muscles. They found significant differences only at moderate intensities (30% MVF), suggesting that these differences could be attributed to variations in motor unit activation frequencies within the involved muscles.

The present study is of critical importance in addressing a significant gap in our understanding of interlimb asymmetries both in motor control and muscular functions, particularly through the lens of force fluctuations and RFD in all its declinations. By incorporating the RFD-SF protocol, our objective is to provide a more profound insight into the dynamics of force production during rapid contractions, thereby offering a more comprehensive understanding of neuromuscular function that extends beyond mere strength measures. By investigating the relationship between motor control and functional outcomes, our research can potentially inform the development of more effective interventions. This comprehensive analysis of motor and muscular function asymmetries aims to enhance the accuracy of assessments and strategies employed by experts in movement science to address the topic of asymmetries from various perspectives. To the best of our knowledge, no study has investigated the asymmetries of the lower limbs between homologous and heterologous muscles by analyzing both strength (observed across a spectrum of expressions) and motor control performance metrics.

In the present study, we examined KEs and KFs in a heterogeneous group of healthy, active adults with several objectives: (1) to quantify the prevalence of asymmetries; (2) to determine whether dominance influences muscle function; (3) to examine whether asymmetries in muscle function are specific to certain muscles or whether one side consistently outperforms the other regardless of muscle group; and (4) to assess whether the direction of asymmetry within each muscle group is consistent across different muscular-strength and motor control performance metrics.

We hypothesize that dominance does not affect muscle function. Therefore, the consistency of the direction of asymmetry between different muscles may be low, and, finally, the consistency of the direction of asymmetry between metrics used for the same muscles may also be low.

Methods

Participants

A convenience sample of 71 healthy students of the Faculty of Sports Science (41 men and 30 women; age, 23 ± 2 years; height, 171.3 ± 14.9 cm; weight, 69.4 ± 17.1 kg; years of sports practice, 8 ± 6 years) were recruited to participate in this study. Although the power analysis indicated that a minimum of 44 participants was required (2-sided paired-sample t test, effect size = 0.5, α = 0.05, power = 0.90), a total of 71 participants were recruited to enhance the robustness of the results and to safeguard against potential data loss or variability.

Inclusion criteria were being >18 years old and physically active. Exclusion criteria were any previous history of neuromuscular disorders or lower limb injury in the last 6 months.

All the participants were informed about the testing procedure and provided written informed consent. This study was approved by the local Ethical Advisory Committee (University of Verona - approval no: 12.R1/2023) and conformed with the Declaration of Helsinki.

Setup

Participants were positioned on a custom-made isometric knee extension device.^15,21 The load cell was placed 2 cm above the malleolus. Using a ratchet strap, the ankle was then immobilized on the support connected to the load cell (Model TS-AMP, 500 kg; AEP transducers). The knee angle was set at 90°. Next, the thigh was immobilized with another ratchet strap to prevent it from lifting off the seat. Finally, the participant was secured to the back of the chair with 2 straps to prevent countermovement. The force signals were sampled at 2048 Hz and converted to digital data with a 16-bit A/D converter (QUATTROCENTO, OT Bioelettronica).

Protocol

A multimetric approach on both limbs and across different muscle groups was used to test the experimental hypotheses.

Participants participated in a single experimental session in which their muscle function was assessed for knee flexion and extension separately for the dominant and nondominant limbs, tested separately in a randomized, counterbalanced order (see Figure 1).

Figure 1.

Schematic representation of experimental procedure timeline. The same procedure was repeated for each muscle group in random order. MVC, maximal voluntary contraction; MVF, maximal voluntary force.

A 5-minute rest period was observed between the testing of each muscle group. For each muscle group and limb, the protocol included: (1) a warm-up with sustained submaximal isometric contractions at increasing intensities; (2) familiarization with ballistic contractions; (3) two MVCs; (4) the RFD-SF protocol^2,47; (5) two 8-second contractions at 50% of maximal voluntary force (MVF) to quantify the complexity and variability of the force signal. ¹⁸

To measure MVF, participants performed two 5-second maximal voluntary contractions (MVCs) with a 2-minute rest between contractions. The RFD-SF protocol began 2 minutes after the last MVC. Participants were instructed to perform 12 ballistic isometric contractions at 20%, 40%, 60%, and 80% of their MVF, with 5 seconds between each contraction. The instruction was to perform fast contractions with peak forces within approximately ±10% of the target force (displayed on a computer screen). If a contraction was not performed correctly, it was repeated. Participants were instructed to generate each pulse as quickly as possible and to relax immediately. Each contraction was interspersed using acoustic cues.

Two 8-second contractions at 50% of MVF were performed 2 minutes after the RFD-SF protocol, interspersed by 1 minute of rest in between.

Data Analysis

The data were analyzed with a customized script (MatLab Version 2023b, TheMathWorks Inc). The force signal was low-pass filtered (fourth order, zero-lag, Butterworth, cut-off frequency 200 Hz). ⁴⁹ MVF was computed as the 200-ms epoch with the highest value of the force signal. Peak RFD (RFD_peak) was also calculated as the highest first derivative of the force signal, adopting a moving window of 20 ms. The onset was selected automatically as the last time the force crossed the zero. RFD at 50 ms, 100 ms, and 150 ms (RFD₅₀, RFD₁₀₀, RFD₁₅₀) from the onset of each contraction were also calculated. Contractions that showed any countermovement (ie, a negative shift >0.5 N in the preceding 200 ms) were excluded. ²⁰ The best three contractions, based on RFD_peak values, were averaged to obtain the RFD metrics RFD_peak, RFD₅₀, RFD₁₀₀, and RFD₁₅₀.

To calculate RFD-SF, the force signal was first preprocessed using an overlapping moving window of 0.1 seconds.^2,16 Then, we calculated the slope of the linear regression between peak force and peak RFD obtained in each contraction (Figure 2). RFD-SF represents how RFD scales with force in a range of submaximal contractions and, thus, quantifies the quickness across a span of intensities. Outliers were detected and removed using the Cook distance methodology to improve the fit of the linear regression.

Figure 2.

Force traces recorded for a representative subject for (a) right limb extension, (b) right limb flexion, (c) left limb flexion, and (d) left limb extension during submaximal ballistic contractions for the RFD-SF protocol. The traces are superimposed with the same onset to highlight the slope differences between different contraction levels. Scatter plots represent the peak force and RFD_peak; the slope of the linear regression represents RFD-SF. MVF, maximal voluntary force; RFD, rate of force development; RFD_peak, peak RFD; RFD-SF, RFD-scaling factor.

All measures of variability and complexity were calculated by visually selecting the steadiest part of the second 8-second contraction at 50% of MVF. ²⁰ The amplitude of variability in the torque output for each contraction was measured using the CoV, which quantifies the magnitude of fluctuations normalized to the mean force output. ApEn was used to estimate the complexity of the force output, ⁴⁵ calculating the negative natural logarithm of the conditional probability that a template of length m is repeated in the timeseries. According to previous literature,^20,44 m (conditional probability) was set at 2, and r (tolerance) was set at 0.1 of the standard deviation of the isometric force during the steady state of the contractions. DFAα was utilized to quantify the correlations within the timeseries and identify long-term dependencies or trends in the data. The fluctuation exponent obtained from DFAα reflects the presence of long-term correlations in the signal. In detail, the DFAα exponent theoretically ranges approximately from ~0.5 to ~1.5 and distinguishes between outputs that exhibit randomness (eg, white noise, α = 0.5), statistically self-similar fluctuations (eg, pink or 1/f noise, α = 1.0), or Brownian characteristics (eg, with long-term memory, α = 1.5).

Statistical Analysis

Descriptive values of dependent variables are reported as mean and standard deviations. The index of asymmetry was calculated using the bilateral strength asymmetry (BSA) index, ²⁹ using the highest value collected between the two limbs for each metric considered to identify the stronger limb, and vice versa:

$$\mathrm{BSA}(\%)=\frac{\mathrm{Stronger}\mathrm{limb}-\mathrm{weaker}\mathrm{limb}}{\mathrm{Stronger}\mathrm{limb}}\times 100$$

To respond to the first experimental inquiry, we calculated the prevalence of asymmetries by calculating how many participants presented BSA levels of 0% to 5%, 5% to 10%, 10 % to 15%, 15% to 20%, 20% to 25%, and >25% (Figure 3). In response to the second and third experimental inquiries, first the Shapiro-Wilk test was used to assess the normality of the data distribution for each variable. A significance level of P ≤ 0.05 was considered indicative of a deviation from normality. In the case of non-normally distributed data, a base-10 logarithmic transformation was applied. If normality was achieved after transformation, a repeated measures analysis of variance (ANOVA) was conducted to assess each metric’s differences between muscle groups and sides; otherwise, the Wilcoxon signed-rank test was employed to assess each metric’s differences between sides for each muscle. Post hoc comparisons were corrected using the Bonferroni method. For parametric data, Cohen’s d was used to assess the effect size of the differences between groups identified in the post hoc analysis. For nonparametric comparisons, effect sizes were estimated using the matched-pairs rank-biserial correlation (r_bs), ³¹ interpreted similarly to correlation coefficients according to standard cutoffs, where 0.2 indicates a small effect, 0.5 is a medium effect, and 0.8 or higher is a large effect. ¹⁹

Figure 3.

Percentual distributions of asymmetry of the most common performance metrics for knee flexors and extensors for a representative sample chosen randomly from the whole pool of subjects (5 men and 5 women). Each column represents a different metric. Positive values show the dominant limb, while negative values show the nondominant calculated using the BSA and assigning a negative sign when the stronger limb was the nondominant one. It is possible to observe that there is never a limb that is clearly more performant than the other—it depends on the metric being considered. BSA, bilateral strength asymmetry; MVF, maximal voluntary force; RFD, rate of force development; RFD_peak, peak RFD; RFD-SF, RFD scaling factor.

To determine the presence of “true” asymmetry, the smallest worthwhile change (SWC, 0.2 × pooled standard deviation) was calculated.^10,14 Participants were considered symmetric if the interlimb difference was less than the SWC. Otherwise, they were considered asymmetric, favoring either the dominant or nondominant side. An independent samples t test was conducted to assess potential differences between male and female participants across BSA, and the effect size was evaluated using Cohen’s d.

To respond to the fourth experimental question, we evaluated the level of agreement between two muscle groups or metrics using the Kappa (κ) coefficient. It accounts for the possibility that some agreement might occur by chance. High κ values indicate that the direction of asymmetry is consistent across different muscle groups or metrics. Thus, κ statistics were used to test the dominant versus nondominant advantage across muscle groups (using the same performance metric) and across metrics (for the same muscle group). All statistical analyses were performed using JASP software (Version 0.19).

Results

The descriptive statistics, presented as means and standard deviations, of all performance metrics, are reported in Table 1. Logarithmic transformation normalized the distribution of all variables, with the exception of RFD-SF, ApEn, CoV, and DFAα, which remained non-normally distributed. The statistical significance, assessed through 2-way repeated measurements ANOVA and consequent post hoc analysis (side × muscles group), was reached only for RFD₁₅₀ (d = 0.305, small effect). The statistical significance, assessed using the Wilcoxon signed-rank test for nonparametric data, was reached only for RFD-SF (r_rb = 0.280, small effect) in the flexor muscles. No differences were achieved for the other metrics.

Table 2.

Asymmetry direction agreement between heterologous muscle groups and within muscles

Parameter	κ	Asymmetry direction agreement
Agreement between muscle groups (KF vs KE)
MVF	0.23	Fair
RFDpeak	0.159	Slight
RFD-SF	0.011	Slight
CoV	0.051	Slight
ApEn	0.023	Slight
DFAα	0	Null
Agreement within muscles (between metrics)
MVC vs RFDrel	–0.147	Null
MVC vs RFD-SF	0.074	Slight
RFDpeak vs RFD-SF	0.127	Slight
RFDpeak vs MVF	0.147	Slight
RFDpeak vs DFAα	0	Null
RFDpeak vs ApEn	–0.12	Null
MVF vs DFAα	0	Null
MVF vs ApEn	0.114	Slight

Kappa values were interpreted as follows: ≤0, null; 0.01-0.20, slight; 0.21-0.40, fair; 0.41-0.60, moderate; 0.61-0.80, substantial; 0.81-0.99, nearly perfect. High κ values would mean that the direction of asymmetry tends to be the same for different muscle groups or metrics. ApEn, approximate entropy; DFAα, detrended fluctuation analysis; KE, knee extensor; KF, knee flexor; MVC, maximal voluntary contraction; MVF, maximal voluntary force; RFD, rate of force development; RFD_peak, peak RFD; RFD-SF, RFD scaling factor.

Table 1.

Descriptive statistics of all performance metrics

Flexors
	Dominant	Nondominant	Signed BSA, % a	BSA, % b	Participant favoring nondominant/symmetric/dominant, %	P value
MVF, N	225.5 ± 61.2	209.4 ± 53.7	4.9 ± 16.8	14.0 ± 10.5	28/25/46	0.09
RFDpeak, N/s	2943.2 ± 1023.8	2820.4 ± 1149.1	5.9 ± 21.6	18.0 ± 13.2	27/25/48	0.13
RFD50, N/s	1127.4 ± 440.7	1113.6 ± 439.8	2.2 ± 35.8	30.3 ± 18.7	41/11/48	>0.99
RFD100, N/s	1389.3 ± 444	1298 ± 420.1	5.6 ± 21.7	17.9 ± 13.4	27/28/45	0.13
RFD150, N/s	1261.7 ± 342.1	1169.2 ± 349.9	7.0 ± 16.5	14.3 ± 10.8	27/24/49	0.003*
RFD-SF, 1/s	9.8 ± 1.5	9.0 ± 1.5	4.4 ± 18.1	14.3 ± 11.9	34/17/49	0.04*
CoV, %	3.07 ± 1.12	3.52 ± 1.17	-3.2 ± 32.5	27.7 ± 17.1	39/21/39	0.54
ApEn	0.63 ± 0.16	0.61 ± 0.17	3.4 ± 20.4	16.3 ± 12.6	32/27/41	0.25
DFAα	1.52 ± 0.14	1.55 ± 0.15	-2.7 ± 8.1	6.2 ± 5.7	51/23/27	0.09
Extensors
MVF, N	623.9 ± 168.5	609 ± 167.6	2.3 ± 11.5	9.3 ± 7.0	27/35/38	0.55
RFDpeak, N/s	6933.9 ± 2284.4	6471.3 ± 2039.8	6.1 ± 21.5	17.7 ± 13.6	25/23/52	0.14
RFD50, N/s	2734.8 ± 1077.6	2427.1 ± 911.2	8.0 ± 28.9	24.6 ± 16.9	27/18/55	0.18
RFD100, N/s	3437.1 ± 940.8	3278.3 ± 854.4	3.9 ± 13.9	11.7 ± 8.3	25/23/52	0.21
RFD150, N/s	3054.7 ± 771.7	2956.7 ± 750.6	3.1 ± 12.7	10.4 ± 7.8	31/30/39	0.29
RFD-SF, 1/s	9.2 ± 1.4	8.9 ± 1.5	2.1 ± 16.9	12.5 ± 11.4	28/23/49	0.12
CoV, %	2.58 ± 0.84	2.75 ± 1.04	-3 ± 27.3	23.3 ± 14.3	42/18/39	0.48
ApEn	0.60 ± 0.09	0.61 ± 0.13	0.8 ± 20.9	12.0 ± 17.0	30/32/38	0.15
DFAα	1.37 ± 0.18	1.40 ± 0.20	-2.1 ± 9.7	7.7 ± 6.2	44/21/35	0.19

Values reported as mean ± SD. ApEn, approximate entropy; BSA, bilateral strength asymmetry; CoV, coefficient of variation; DFAα, detrended fluctuation analysis; MVF, maximal voluntary force; RFD, rate of force development; RFD_50,100,150, RFD assessed from the onset to 50 ms, 100 ms, and 150 ms; RFD_peak, peak RFD; RFD-SF, RFD scaling factor; SWC, smallest worthwhile change.

^aSigned BSA was calculated as follows: ([(Stronger Leg - Weaker Leg)/Stronger Leg] × 100) × IF [dominant > nondominant, 1, –1]. Negative values indicate favor of the nondominant limb and vice versa. The percentage of participants symmetric or favoring nondominant, dominant limbs is calculated based on the SWC.

^bBSA was calculated as follows: ([Stronger Leg – Weaker Leg]/Stronger Leg) × 100.

Figure 3 shows the distribution of BSA across performance metrics for 10 representative subjects (5 women). It is evident that the distribution does not consistently favor either the dominant (positive values) or nondominant (negative values) side; rather, it varies depending on the specific metric considered. Most subjects display some metrics that favor one side and some others that favor the other side.

Figure 4 shows the prevalence of absolute BSA for each metric, stratified in 5% levels. Considering the KE, the most common BSA indices for MVF, RFD-SF, RFD₁₀₀, RFD₁₅₀, ApEn, and DFAα were 0% to 5% or 5% to 10%. For RFD₅₀ and CoV, the most common BSA was >25%. Considering the KF, the most common BSA indices for MVF, RFD₁₅₀, and DFAα were 0% to 5% or 5% to 10%. For RFD-SF was 10% to 15%, and for RFD_peak, RFD₅₀, RFD₁₀₀, CoV and ApEn, the most common BSA was >25%.

Figure 4.

The percentage of subjects with different levels of asymmetry was assessed through BSA across all performance and motor control metrics for both KEs and KFs. This figure includes all subjects, regardless of whether their level of asymmetry is to be considered real or not based on the SWC. ApEn, approximate entropy; BSA, bilateral strength asymmetry; CoV, coefficient of variation; DFAα, detrended fluctuation analysis; KE, knee extensor; KF, knee flexor; MVF, maximal voluntary force; RFD, rate of force development; RFD_50,100,150, RFD assessed from the onset to 50 ms, 100 ms, and 150 ms; RFD_peak, peak RFD; RFD-SF, RFD scaling factor; SWC, smallest worthwhile change.

No significant differences were found between men and women in any BSA metrics (all P ≥ 0.12; see Figure 5 for density distributions).

Figure 5.

Density distributions of asymmetry values were calculated using the BSA, divided by sex. The density curves represent the distribution of asymmetry values for male (M) and female (F) participants. The y-axis indicates the relative density, while the x-axis shows the range of asymmetry values. The curves highlight differences in asymmetry distributions between the 2 sexes, showing central tendencies and variations in bilateral performance symmetry. The graphs are presented for both extensor and flexor metrics across performance (RFD_peak, RFD-SF, MVF) and motor control (CoV, ApEn, DFAα) measures. ApEn, approximate entropy; BSA, bilateral strength asymmetry; CoV, coefficient of variation; DFAα, detrended fluctuation analysis; MVF, maximal voluntary force; RFD, rate of force development; RFD_50,100,150, RFD assessed from the onset to 50 ms, 100 ms, and 150 ms; RFD_peak, peak RFD; RFD-SF, RFD scaling factor.

The results about the asymmetry direction agreement between muscle groups and metrics were either null, slight, or fair (Table 2, all κ ≤ 0.159).

Discussion

The first aim of the present work was to investigate the effect of dominance on maximal strength, rapid strength, and force steadiness of the KE and KF in a sample of active, healthy young adults. With a few exceptions (RFD₁₅₀ and RFD-SF for KF), the results show that dominance, reported in agreement with a previous study, ⁵⁰ does not influence performance (Table 1). This is further supported by the asymmetry direction agreement calculated through the κ coefficients, which were considered mostly slight or null (Table 2), suggesting that neither side is better performing than the other in terms of depending on muscle group and performance metrics adopted. ²⁰

The findings of the current study suggest that the dominant limb is not necessarily stronger than the nondominant. This apparently paradoxical trend is supported by Kozinc and Šarabon, ³² who found that, except for a few outcome measures, the dominant limb is not guaranteed to be the one with the best performance. In a study of volleyball and basketball players, the authors found that dominant limbs did not predict better jumping performance in ~60% of cases, suggesting that leg dominance cannot be used to predict limb jump performance. ²⁷ In another study on kickboxing athletes, ³⁴ it was found that there was no difference between dominant and nondominant limbs in terms of absolute strength. The only difference was found in the interlimb asymmetries on RFD_peak of the KE in favor of the dominant limb, probably since the dominant limb is used more extensively for kicking during fighting matches and training.

Our findings corroborate that asymmetries are ubiquitous in physical performance.^4,14 Based on the SWC, 75% to 78% of participants can be considered asymmetrical in muscular strength or motor control metrics. Therefore, we found that only 22% to 25% of participants were considered symmetrical in flexors and extensors, respectively (Table 1). This was valid for muscular strength and motor control performance metrics of both muscle groups (Table 1). Our results perfectly agree with and further strengthen the findings of Markström et al. ³⁶ In their work on lower limb asymmetries, the latter authors found that only 24% of healthy subjects passed the symmetry criterion assessed with a battery of tests that included hop for distance, vertical hop, and isometric knee extensions and flexions. The asymmetry should be considered the norm, not the exception. On average, the absolute values of BSA ranged between 14% and 18%, considering both muscular strength and motor control performance metrics (Table 1). As can be seen in Figure 4, most participants showed an asymmetry level >10%, which is a cut-off (insensately) considered safe in many sports and rehabilitation contexts. We provide evidence that most apparently healthy people exceed the 10% asymmetry threshold across various neuromuscular functions; therefore, efforts to reduce asymmetry levels below this threshold should be approached with caution. Furthermore, it is important to emphasize that arbitrary thresholds that many studies have perpetuated over time should be avoided in favor of individualized approaches. ³

Potentially, more than being related to the risk of injury, the presence of asymmetries might be particularly relevant for identifying people who could adopt dysfunctional or ineffective movement patterns.^5,6,12 Indeed, large unilateral asymmetries, as measured by both the countermovement jump and drop jump tests, have been linked to poorer performance in acceleration, sprinting, and changes in direction.^11,26 Findings of a previous study showed that young subelite athletes show greater strength asymmetries between limbs than elite athletes. ⁵ However, the literature suggests that asymmetries do not necessarily negatively impact athletic performance. In some cases, a certain degree of asymmetry could be considered a requirement of some sports that require sprinting in curves as part of the technical gesture (e.g., some athletics disciplines such as relays or open-skill sports). This is also supported by a study on sprint-trained athletes that showed no significant relationships between kinematic and kinetic asymmetry with sprint running performance ²³ ; again, in another study on fencers, ⁵¹ where athletes with greater lower limb strength differences were considered more experienced because they had been practicing the sport longer. Taken together, this information shows that in active people who engage in various activities/disciplines, asymmetries are commonly present and manifest in different ways depending on the test, muscle group, and metric considered.

The direction of asymmetry varies depending on the metric and muscle considered (Figure 3). Similar graphs have also been proposed in other studies,^14,40 showing similar trends highlighting the differences in asymmetry based on the choice of different muscle groups or tests. These results are consistent with the existing literature. Indeed, Boccia et al ¹⁴ have shown that asymmetries are metric- and muscle-dependent when considering the extensor and flexor muscles of the upper limbs. Moreover, a study by Smajla et al ⁴⁸ on the lower limbs reached the same conclusion, as the κ values they calculated indicate a level of agreement that can be considered random.

Our results showed no significant differences emerged when comparing BSA between the group of men and women. This result contrasts with those of Bailey et al, ¹ who found that female athletes exhibited greater between-limb differences in bilateral jumping compared with their male counterparts. Another study on pole vaulters revealed significant sex differences in interlimb asymmetry, ³⁹ with male athletes exhibiting greater asymmetry in step length and step frequency than their female counterparts. However, we agree with Fort-Vanmeerhaeghe et al, ²⁶ who reported no significant differences in asymmetry values between sexes in vertical and horizontal jumping tests. Figure 5 shows the density of the asymmetry distributions divided by sex. Although some metrics exhibit more heterogeneous distributions for the male group (eg, MVF) and others for the female group (eg, DFAα), the t test showed no significant differences between the two groups.

Some methodological notes should be made. The choice to use the SWC to determine whether subjects should be considered asymmetric or not is justified by the fact that the commonly used 10% to 15% thresholds are arbitrary,^10,42 while the SWC, recommended by previous studies,^9,10 is based on effect size principle and take into account the variability of the population. Also, the choice of formula for calculating the level of asymmetries could lead to different results.^9,10,42 However, among these, Bishop et al ¹⁰ elegantly justify the use of the BSA in cases where the focus is on a unilateral test. However, as clarified by Impellizzeri et al, ²⁹ this formula has the disadvantage of always producing positive results, thereby losing the indication of the direction of asymmetry in favor of the dominant or nondominant limb. To address this issue, the arbitrary use of a negative sign when the nondominant limb is the stronger one is a solution already used in the literature, ²⁹ which allows for an indication of the direction of any potential asymmetry. Therefore, our choice, although justified, is arbitrary, and this may make the results not necessarily comparable with other studies.

It is important to highlight that, as isometric tests, which are far from being ecologically valid measures, were used, the results cannot be extended to dynamic performance. This is because asymmetry assessments are task-dependent. Other factors that should be considered, which may not necessarily make our results applicable to other populations, include the age of the subjects and the fact that they were all active people involved in some form of sports or fitness activities. The same study might produce different outcomes if conducted on populations such as elderly, sedentary, or pathological people. The lack of neuromuscular/physiological assessments that investigate the underlying mechanisms of asymmetries does not allow for the identification of their causes. Similarly, the truly critical thresholds and their correlation with injury risk have yet to be identified. These are currently open questions without definitive answers that researchers should aim to investigate thoroughly.

These results suggest that the muscle asymmetries of the lower limbs for KE and KF depend on the metric and the muscle considered. Furthermore, neither limb is superior in overall performance to the other; it always depends on which parameter is considered. Finally, the BSA did not show any differences between men and women.

From a practical point of view, rather than relying on a single test, which may give an incomplete picture of the subject’s ability, it is advisable to take a comprehensive approach to identifying asymmetries. Indeed, to avoid assuming that if one metric shows asymmetry, others will necessarily follow the same pattern, or that the dominant side is always stronger than the nondominant side, it is essential to evaluate all variables considered relevant to both limbs. Therefore, according to the literature, ²³ testing for asymmetry is essential before initiating any “compensatory” training. This approach ensures that imbalances are identified and addressed accurately, resulting in a more effective and personalized training program that supports balanced interlimb capacity where needed. Therefore, multifactorial/multimetric tests or analyses should be considered to assess asymmetries adequately.