Deviations in whole body angular momentum are largely corrected before foot placement

This study demonstrates that mediolateral foot placement during walking is primarily governed by whole-body linear momentum rather than angular momentum, as deviations in angular momentum are largely corrected before foot placement occurs.

The Gist

The Big Question: How Do We Stay Upright?

Imagine you are walking down a busy street. Suddenly, someone bumps into you from the side, or maybe you trip over a crack in the sidewalk. Your body has to react instantly to keep from falling.

Scientists have long known that our brains use a "mental map" of our body's speed and position (called linear momentum) to decide where to put our next foot. It's like a driver checking their speed and distance from a curb before turning the wheel.

But this study asked a new question: Does our brain also pay attention to how much we are spinning or twisting (called angular momentum) when we stumble?

Think of it this way:

• Linear Momentum is like a car sliding sideways on ice. The driver needs to steer to stop the slide.
• Angular Momentum is like a figure skater who gets pushed and starts to spin uncontrollably. The skater needs to stop the spin before they can steer.

The researchers wanted to know: When we get pushed, do we fix the spin first, or do we just focus on where to put our foot to stop the slide?

The Experiment: The "Push" and the "Twist"

The researchers put 10 healthy adults on a giant treadmill and gave them a gentle safety harness. Then, they used two motors to give the participants a sudden, sharp push. They did two types of pushes:

1. The "Slide" (Translation): They pulled the person's hip to the side. This made the whole body slide sideways, changing their speed and position, but not really making them spin.
2. The "Twist" (Rotation): They pulled the hip one way and the shoulder the other way at the same time. This didn't move the body sideways much, but it made the torso twist and spin like a top.

They measured everything: how fast the people moved, how much they twisted, and exactly where they placed their next foot.

The Surprising Discovery

The researchers expected that when people got "twisted," their brains would use that twisting sensation to decide where to put their foot. They thought, "If I'm spinning, I need to step wide to stop the spin."

But that's not what happened.

Here is the breakdown of what they found:

• The "Spin" was fixed immediately: When the participants got twisted, their bodies reacted instantly while their foot was still on the ground (the "stance phase"). They used their hips and ankles to twist their body back to a straight position almost before they even lifted their foot to take the next step.
• The "Foot Placement" only cared about the "Slide": By the time the participants were in the air, swinging their leg to take the next step, the "spin" was already gone. Their brains were only looking at the "slide" (how fast they were moving sideways) to decide where to put their foot.

The Analogy:
Imagine you are riding a unicycle.

1. Someone pushes you, and you start to wobble and spin (Angular Momentum).
2. Your body instantly shifts your weight and pedals to stop the spin while you are still on the wheel.
3. Once you are stable and not spinning anymore, you look at how fast you are moving forward and sideways to decide where to steer next.

The study found that our brains are like that unicyclist. We fix the spin so quickly that we don't need to worry about it when we actually place our foot down. We only use our foot placement to fix the sliding motion.

Why Does This Matter?

This is a big deal for two reasons:

1. Simpler Brains: It suggests our brains have a very efficient, two-step strategy. First, fix the orientation (stop the spin) using muscles in the legs and hips while standing. Second, fix the position (stop the slide) by stepping. We don't try to do both at once with our foot placement.
2. Better Robots and Prosthetics: If we are building robots or artificial legs that need to walk like humans, we don't need to program them to constantly calculate complex "spinning" data to decide where to step. We can focus on the simpler "sliding" data, which makes the technology easier to build and more stable.

The Bottom Line

When you get bumped while walking, your body is a master of multitasking. It fixes the "twist" instantly while your foot is still on the ground. By the time you swing your leg to take the next step, the twist is history. Your foot placement is almost entirely about correcting the "slide," not the "spin."

In short: We fix the spin with our muscles, and we fix the slide with our feet.

Technical Summary

1. Problem Statement

While the control of linear momentum (specifically the Center of Mass, or CoM) during human walking is well-understood through frameworks like the Extrapolated Center of Mass (XCoM), the role of whole-body angular momentum (WBAM) in mediolateral foot placement control remains less clear.

• The Gap: Previous studies suggest humans prioritize angular momentum control in certain scenarios (e.g., trips), and recent literature indicates foot placement correlates with integrated WBAM. However, it is unknown whether WBAM provides independent predictive value for foot placement adjustments beyond CoM state (position and velocity) when specific mechanical perturbations target angular momentum directly.
• Research Question: Does incorporating whole-body angular momentum into predictive models significantly improve the estimation of mediolateral foot placement following mechanical perturbations, compared to models based solely on CoM linear dynamics?

2. Methodology

Participants and Setup

• Subjects: 10 healthy adults (after excluding 4 due to data loss) with no musculoskeletal injuries.
• Environment: Walking on a large treadmill (250 cm × 450 cm) at two speeds: Slow (2 km/h) and Normal (5 km/h).
• Perturbation System: A custom "Bump'm" system using two motors to apply lateral pulls via strings attached to a safety harness.
  • Translation Perturbation: A pull to the pelvis only (primarily alters linear momentum).
  • Rotation Perturbation: Simultaneous pulls to the pelvis and opposite shoulder (primarily alters angular momentum with minimal net force).
• Perturbation Parameters: Applied at heel strike, magnitude ~120 N, duration 300 ms.

Data Collection & Processing

• Kinematics: Captured via 7 Qualisys cameras (100 Hz) with full-body marker sets.
• Modeling: Data processed in OpenSim (Gait2392 model) to compute joint angles, segment orientations, and CoM trajectories.
• WBAM Calculation: Computed using the Herr & Popovic (2008) method, summing the angular momentum of all body segments relative to the whole-body CoM.
• Outcome Measures: Step width, step time, and Margin of Stability (MoS) for the pre-perturbation step and the first three post-perturbation steps.

Analytical Approach
Two linear regression models were fitted to predict mediolateral foot placement at heel strike based on data collected during the swing phase:

1. Model 1 (CoM only): Predictors = CoM position ($r_{CoM}$) and CoM velocity ($v_{CoM}$).
2. Model 2 (CoM + WBAM): Predictors = $r_{CoM}$, $v_{CoM}$, and Whole-Body Angular Momentum ($WBAM_{AP}$).

• Statistical Comparison: The explained variance ($R^2$) of both models was compared across the swing phase using Statistical Parametric Mapping (SPM1D) to determine if adding WBAM significantly improved prediction.

3. Key Results

Kinematic Responses to Perturbations

• Translation Perturbations: Caused large deviations in CoM position (0.2 m) and velocity (0.4 m/s) but minimal changes in WBAM.
• Rotation Perturbations: Induced strong WBAM deviations (5 kg·m²/s) but smaller CoM deviations (0.05 m) compared to translation.
• Recovery Dynamics:
  • Step Width: Participants stepped in the direction of the pull immediately (Post 1). For translation, this effect persisted to Post 2. For rotation, step width returned to baseline by Post 2.
  • Margins of Stability (MoS): Reduced significantly in Post 1 for all conditions. Translation perturbations showed prolonged recovery (significant differences up to Post 3), whereas rotation perturbations recovered faster (mostly baseline by Post 3).

Foot Placement Prediction Models

• Translation Perturbations: Including WBAM in Model 2 did not significantly improve the prediction of foot placement compared to Model 1 (CoM only) at any point in the swing phase.
• Rotation Perturbations: Including WBAM provided a statistically significant but small improvement in explained variance ($R^2$) only during early swing (approximately 10–20% of the swing phase) at normal walking speeds.
• Regression Coefficients: The coefficient for WBAM ($\beta_3$) was time-specific and generally small, indicating a limited role in determining final foot placement.

4. Key Contributions

1. Temporal Hierarchy of Control: The study provides evidence that humans prioritize the correction of whole-body angular momentum during the stance phase (immediately after perturbation onset) rather than relying on foot placement adjustments in the subsequent swing phase.
2. Dominance of Linear Momentum: Despite the application of pure rotational perturbations, mediolateral foot placement is predominantly governed by linear momentum dynamics (CoM position and velocity). WBAM contributes minimally to foot placement prediction once the stance-phase corrections are made.
3. Methodological Insight: The findings suggest that discrepancies with previous studies (e.g., Leestma et al., 2023) regarding the importance of WBAM may stem from perturbation timing. By applying perturbations at heel strike, this study allowed for immediate stance-phase corrections, whereas later perturbations (mid-stance) might force reliance on foot placement to correct angular momentum.

5. Significance and Implications

• Theoretical Impact: This challenges the notion that foot placement is a primary mechanism for correcting angular momentum deviations in the frontal plane. Instead, it supports a control strategy where stance-phase muscle activity (likely hip abductors and ankle plantar flexors) actively restores angular momentum and body orientation before the swing leg is positioned.
• Clinical Relevance: Understanding that angular momentum is corrected early in the gait cycle (stance) rather than late (swing) is crucial for rehabilitation strategies. For populations with impaired stance-phase control (e.g., stroke, spinal cord injury), the inability to correct angular momentum early may force reliance on compensatory, potentially less stable, foot placement strategies.
• Future Directions: The authors note that perturbation timing, magnitude, and walking speed are critical variables. Future research should investigate perturbations applied later in the stance phase to see if WBAM becomes a stronger predictor of foot placement when stance-phase correction is constrained.

Conclusion: The study concludes that while whole-body angular momentum is critical for gait stability, its deviations are largely corrected via stance-phase mechanisms before foot placement occurs. Consequently, mediolateral foot placement is primarily a mechanism for controlling linear momentum, with whole-body angular momentum playing only a limited, time-specific role in its prediction.