Predictors of Step Length from Surface Electromyography and Body Impedance Analysis Parameters

Jin-Woo Park, Seol-Hee Baek, Joo Hye Sung, Byung-Jo Kim2022 · DOI

TL;DR

This study identifies key modifiable factors—specifically, knee flexor muscle strength (measured via sEMG) and body fat mass—that significantly predict an individual's step length. These findings suggest that targeted interventions like strengthening knee flexors and reducing body fat could be effective strategies for improving gait performance and potentially mitigating age-related declines.

What Problem Does This Paper Address?

Walking performance is a crucial indicator of overall health and functional status, with declines often linked to aging and neurodegenerative diseases. Step length, a fundamental component of gait, is particularly important as a shortened step length in the elderly can signal an increased risk of falls and functional loss. While many factors influence gait, the specific modifiable factors related to musculoskeletal function and body composition that predict step length have not been thoroughly investigated. This paper aims to fill this gap by exploring how surface electromyography (sEMG) parameters from leg muscles, body impedance analysis (BIA) metrics, and demographic data can predict individual step length.

Methods

The researchers conducted an exploratory, cross-sectional study involving healthy individuals aged 40-80. They employed a multi-faceted approach to collect data:

  • Step Length Measurement: The GAITRite® system, an electronic walkway with embedded sensors, was used to precisely measure step length. Participants walked at their comfortable pace, and the average step length was calculated from multiple steps.
  • sEMG Measurement: Surface electromyography (sEMG) was used to assess the electrical activity of key leg muscles. Specifically, electrodes were placed on the rectus femoris (knee extensor) and biceps femoris (knee flexor) muscles. Participants performed maximal voluntary knee flexion and extension against resistance, and the root mean square (RMS) value of the sEMG signal during these efforts was recorded. This RMS value serves as an indicator of muscle activation intensity.
  • Body Impedance Analysis (BIA): A multifrequency bioelectrical impedance analyzer (Inbody770®) was used to measure body composition, focusing on appendicular skeletal muscle mass (ASM) and body fat mass (BFM). These values were adjusted for height to account for body size.
  • Statistical Analysis: A series of regression analyses were performed. Simple linear regressions were used to identify initial correlations between individual variables (age, height, BMI, ASM, BFM, sEMG RMS values) and step length. Subsequently, multivariable regression models with stepwise selection were built to determine the most significant predictors of step length. Three models were developed: one including all variables, one excluding sEMG, and one excluding both sEMG and BIA, to understand the relative contributions of each data type.

Datasets

The study utilized data collected from 179 healthy participants (82 men, 97 women) aged 40-80 years. This dataset, comprising demographic information, sEMG measurements, BIA metrics, and GAITRite® step length data, was collected prospectively for this study. No pre-existing public datasets were used for training or evaluation.

Key Insights and Ideas

  • Modifiable Factors are Key: The most significant insight is the identification of modifiable factors—knee flexor strength (sEMG RMS) and body fat mass (BFM)—as important predictors of step length. This shifts the focus from purely inherent characteristics like age and height to actionable interventions.
  • Knee Flexors Matter More Than Extensors for Step Length: The study found that the RMS value of the biceps femoris (knee flexor) during maximal flexion was a stronger predictor of step length than the rectus femoris (knee extensor) during extension. This highlights the critical role of knee flexion strength in generating forward propulsion and controlling the gait cycle.
  • Body Fat as a Gait Limiter: A higher body fat mass was associated with shorter step lengths. This suggests that excess adipose tissue may impede efficient movement and contribute to reduced gait performance.
  • Age and Height Remain Important, But Not the Whole Story: While age and height were consistently significant predictors across models, their predictive power was enhanced when combined with sEMG and BIA data, underscoring that step length is a complex interplay of factors.

Key Results

The study developed three regression models to predict average step length:

  • Model 1 (Best-fit): Included RMS value at maximal knee flexion and height. This model explained the most variance in step length, with an R² of 0.244 (adjusted R² = 0.235, p < 0.001).
  • Model 2 (Excluding sEMG): Included age, height, and BFM. This model showed a predictive power of R² = 0.198 (adjusted R² = 0.185, p < 0.001).
  • Model 3 (Excluding sEMG and BIA): Included only age and height. This model had the lowest predictive power, with an R² of 0.158 (adjusted R² = 0.148, p < 0.001).

These results quantitatively demonstrate that knee flexor strength and body fat mass, in addition to age and height, significantly contribute to predicting step length.

Limitations and Future Work

The authors acknowledge several limitations:

  • Small R² Values: The R² values for the models were relatively small, indicating that while these factors are significant predictors, they do not fully capture all the variability in step length. The models were not designed to measure "accurate step length" for individuals but rather to identify key predictors.
  • Limited sEMG Scope: sEMG data was only collected from thigh muscles (rectus femoris and biceps femoris). Future work could include muscles from the lower leg (e.g., tibialis anterior, gastrocnemius) for a more comprehensive analysis.
  • Population Specificity: The study was conducted on a Korean population. While the strong correlations suggest generalizability, ethnic differences in body proportions might influence the direct applicability of these specific models to other populations.
  • Measurement Methods: While BIA is cost-effective and readily available, dual-energy X-ray absorptiometry (DXA) is considered a gold standard for body composition. Future studies could compare these methods.

The authors suggest that further research could explore joint angle measurements during gait and investigate the impact of other lower limb muscles. They also propose that strengthening knee flexor function and reducing body fat mass could be beneficial interventions for improving step length, particularly in older adults.

Summary updated

BibTeX
@misc{park2022predictors,
  author = {Jin-Woo Park and Seol-Hee Baek and Joo Hye Sung and Byung-Jo Kim},
  title = {{Predictors of Step Length from Surface Electromyography and Body Impedance Analysis Parameters}},
  year = {2022},
  doi = {10.3390/s22155686},
}
Figure