📄 sports medicine

Bench-stepping training improves stair-walking dynamics in older women: evidence from an exploratory nonlinear kinematic analysis

This study demonstrates that a 12-week bench-stepping training program improves stair-walking speed and alters movement complexity in older women, with distinct nonlinear dynamic changes observed between stair ascent and descent that suggest different underlying motor control strategies.

Original authors: Baggen, R. J., van Schooten, K. S., Van Roie, E., Verschueren, S. M., Delecluse, C., Delbaere, K., Lord, S. R., van Dieen, J. H.

Published 2026-07-07

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Baggen, R. J., van Schooten, K. S., Van Roie, E., Verschueren, S. M., Delecluse, C., Delbaere, K., Lord, S. R., van Dieen, J. H.

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

The Big Picture: Climbing vs. Coming Down

Imagine your body is a car driving up and down a steep hill.

Going Up (Ascent): You need to press the gas pedal hard to fight gravity and lift the car up. This requires power and strength.
Going Down (Descent): You need to use the brakes carefully to control the car so it doesn't roll too fast. This requires balance and control.

This study looked at a group of healthy older women to see if a specific exercise program (called "bench-stepping," which is basically stepping up and down on a box like a step-aerobics class) helped them navigate stairs better. While we already knew this exercise made them go faster up the stairs, the researchers wanted to know if it also changed how they moved their bodies to stay balanced.

The Tools: Looking at the "Engine" in Detail

Usually, scientists just measure how fast someone walks or how far they sway. But this study used a special "microscope" for movement called nonlinear analysis.

Think of walking up stairs like a musician playing a drumbeat.

Linear Analysis just counts the beats: "Is the tempo fast? Is the volume loud?"
Nonlinear Analysis listens to the rhythm and predictability. Is the drummer playing a perfect, robotic beat? Or are they improvising with a complex, jazz-like rhythm that adapts to the moment?

The researchers used a sensor on the women's lower backs to listen to the "rhythm" of their body's movement (specifically the acceleration of their center of mass) as they walked up and down a 12-step staircase.

What They Found: Two Different Stories

1. Going Up the Stairs (The "Power" Story)

After 12 weeks of bench-stepping training, the women got faster at climbing stairs.

The Change: Their movements became more predictable.
The Analogy: Before training, their movement rhythm was a bit like a jazz drummer who was slightly unsure of the beat, adding random little variations. After training, they became more like a metronome. Their movements were smoother and more consistent.
Why it matters: Even though they were moving faster (which usually makes things wobbly), their bodies actually became more organized and less "chaotic." This suggests their brains and muscles learned to coordinate better to handle the speed.

2. Going Down the Stairs (The "Control" Story)

Here is where it gets interesting. The training did not make them go faster down the stairs, nor did it change the size of their steps.

The Change: Surprisingly, their movement rhythm became less predictable (more complex).
The Analogy: If going up was like turning a jazz drummer into a metronome, going down was like taking a rigid metronome and letting the drummer start improvising again.
Why it matters: The researchers believe this "messier" rhythm is actually a good thing for going down stairs. It suggests the women's bodies became more adaptable and flexible, ready to react to small slips or changes in balance. It's like a tightrope walker who stops walking stiffly and starts making tiny, fluid adjustments to stay upright.

The "Two Different Tasks" Discovery

The study used a statistical method (Factor Analysis) to see if going up and going down are the same skill.

The Verdict: They are not the same.
The Analogy: Think of it like driving a car. Driving up a hill is a different skill than driving down a hill. You use different muscles and different brain strategies for each. The training helped the women get better at the "uphill" strategy (power and consistency), but it didn't automatically fix the "downhill" strategy (balance and adaptability).

The Main Takeaway

The study found that a simple stepping exercise changed how older women moved, but in different ways depending on the direction:

Going Up: They became faster and more consistent (like a well-oiled machine).
Going Down: They became more adaptable and flexible (like a skilled dancer adjusting to the floor).

The researchers concluded that we shouldn't treat "stair walking" as one single thing. Going up and going down are two different challenges that require different kinds of training and different kinds of body control. The "nonlinear" tools they used (like measuring the rhythm's predictability) were very sensitive and could spot these subtle improvements that standard speed tests missed.

Technical Summary: Bench-stepping training improves stair-walking dynamics in older women: evidence from an exploratory nonlinear kinematic analysis

Problem Statement
Stair walking presents significant balance and coordination challenges for older adults, constituting a leading cause of in-home fall-related injuries requiring emergency care. While previous research established that bench-stepping training improves lower-limb power and stair-climbing speed in healthy older women, it remains unclear whether such training enhances dynamic balance and movement complexity during stair negotiation. Traditional linear analyses (e.g., root mean square, harmonic ratios) often rely on simple stimulus-response assumptions and may lack the sensitivity to detect subtle, training-related improvements in motor control, particularly in fit cohorts where a "ceiling effect" might mask gains. Furthermore, the distinct biomechanical demands of stair ascent versus descent are often treated as interchangeable, despite evidence suggesting different underlying control strategies.

Methodology
This study performed a secondary analysis of data from a 12-week randomized controlled trial involving 45 healthy community-living older women (mean age 69 ± 4 years). Participants were assigned to a STEEP bench-stepping intervention group ( $n=24$ ) or a non-training control group ( $n=21$ ). The intervention consisted of 40 minutes of progressive bench stepping (incremental heights and weighted vests) three times per week.

Stair negotiation was assessed pre- and post-intervention using a trunk-mounted inertial sensor (Dynaport MoveTest®) recording three-dimensional linear accelerations and angular velocities at 100Hz. Participants walked up and down a 12-step staircase at a self-selected safe speed without handrails. Data processing involved Kalman filtering for gravity and heading correction, followed by the removal of start/end phases to isolate steady-state gait.

The study employed two analytical approaches:

Linear Dynamics: Measured time, root mean square (RMS) acceleration magnitude, harmonic ratios (symmetry), mean pelvis angular velocity, and dimensionless jerk.
Nonlinear Dynamics: Quantified movement complexity and regularity using Sample Entropy (SE), Recurrence Quantification Analysis (RQA) metrics (determinism, laminarity, mean/max line length), and Higuchi Fractal Dimensions (HFD).
Statistical Modeling: Linear mixed-effects models (LMMs) assessed group $\times$ time interactions. Additionally, exploratory factor analysis (EFA) and principal component analysis (PCA) were conducted separately for ascent and descent to identify latent constructs of motor behavior.

Key Results

Stair Ascent: The intervention group demonstrated a significant reduction in ascent time ( $p=0.018$ ) accompanied by proportional increases in acceleration magnitudes (RMS) across all directions. Crucially, Sample Entropy (SE) decreased significantly in the mediolateral (ML) and vertical (VT) directions ( $p \leq 0.012$ ), indicating more predictable and less complex movement patterns. No significant changes were observed in linear symmetry or jerk measures.
Stair Descent: No significant changes were found in speed or acceleration magnitudes. However, nonlinear measures revealed distinct adaptations: SE in the VT direction increased ( $p=0.001$ ), suggesting greater unpredictability or adaptability, and the maximum RQA line length (Lmax) in the VT direction increased ( $p=0.008$ ), indicating more persistent recurring patterns.
Factor Analysis: Exploratory factor analysis revealed that stair ascent and descent are characterized by distinct latent dimensions. Ascent was dominated by a single factor encompassing speed, acceleration magnitude, and complexity. Descent required five factors, including specific dimensions for vertical control and speed, confirming that the two tasks rely on different motor control processes.
Training Effects on Factors: Only the primary factor for stair ascent showed a significant training-related increase in factor scores, reflecting a shift toward a faster, more vigorous movement strategy with altered complexity characteristics.

Key Contributions

Sensitivity of Nonlinear Measures: The study demonstrates that nonlinear metrics, particularly Sample Entropy, are more sensitive than conventional linear measures in detecting training-induced changes in stair-walking dynamics. SE successfully identified alterations in movement predictability that linear metrics missed.
Task-Specific Adaptations: The research provides evidence that bench-stepping training induces different dynamic changes for ascent versus descent. Ascent improvements were characterized by increased predictability (lower entropy), while descent changes suggested increased adaptability (higher entropy) without speed gains.
Distinct Motor Constructs: Factor analysis supports the hypothesis that stair ascent and descent are not interchangeable phases but distinct motor tasks with different underlying control strategies and latent factors.

Significance and Claims
The authors conclude that bench-stepping training induces measurable, task-specific changes in stair-walking dynamics in healthy older women. They posit that Sample Entropy serves as a sensitive marker for these changes, particularly regarding vertical accelerations.

The paper modestly claims that these findings highlight the importance of distinguishing between stair ascent and descent in both research and clinical assessment. It suggests that rehabilitation programs should not assume that ascent-focused training generalizes to descent safety, as the two tasks impose different control demands (concentric power vs. eccentric control/balance). The study positions nonlinear kinematic measures as complementary tools for monitoring motor control quality, noting that their clinical utility depends on future work to establish minimal detectable changes and link them to fall outcomes. The authors acknowledge limitations, including the healthy, fit nature of the cohort (potential ceiling effects) and the exploratory nature of the analysis, framing the results as hypothesis-generating rather than confirmatory.