Amputee locomotion: Frequency content of prosthetic vs. intact limb vertical ground reaction forces during running and the effects of filter cut-off frequency

A 0.05 m Change in Inertial Measurement Unit Placement Alters Time and Frequency Domain Metrics during Running

Inertial measurement units (IMUs) provide exciting opportunities to collect large volumes of running biomechanics data in the real world. IMU signals may, however, be affected by variation in the initial IMU placement or movement of the IMU during use. To quantify the effect that changing an IMU's location has on running data, a reference IMU was 'correctly' placed on the shank, pelvis, or sacrum of 74 participants. A second IMU was 'misplaced' 0.05 m away, simulating a 'worst-case' misplacement or movement. Participants ran over-ground while data were simultaneously recorded from the reference and misplaced IMUs. Differences were captured as root mean square errors (RMSEs) and differences in the absolute peak magnitudes and timings. RMSEs were ≤1 g and ~1 rad/s for all axes and misplacement conditions while mean differences in the peak magnitude and timing reached up to 2.45 g, 2.48 rad/s, and 9.68 ms (depending on the axis and direction of misplacement). To quantify the downstream effects of these differences, initial and terminal contact times and vertical ground reaction forces were derived from both the reference and misplaced IMU. Mean differences reached up to -10.08 ms for contact times and 95.06 N for forces. Finally, the behavior in the frequency domain revealed high coherence between the reference and misplaced IMUs (particularly at frequencies ≤~10 Hz). All differences tended to be exaggerated when data were analyzed using a wearable coordinate system instead of a segment coordinate system. Overall, these results highlight the potential errors that IMU placement and movement can introduce to running biomechanics data.

Anterior-posterior ground reaction forces across a range of running speeds in unilateral transfemoral amputees

As a fundamental motor pattern, the ability to run at a range of constant speeds is a prerequisite for participating in competitive games and recreational sports. However, it remains unclear how unilateral transfemoral amputees modulate anterior and posterior ground reaction force impulses (GRFIs) in order to maintain constant running speeds. The purpose of this study was to investigate anterior and posterior GRFIs across a wide range of constant running speeds in unilateral transfemoral amputees wearing a running-specific prosthesis. Eleven runners with unilateral transfemoral amputation ran on an instrumented treadmill at 5 different speeds (30%, 40%, 50%, 60%, and 70% of the average velocity of their 100-m personal records). Anterior-posterior ground reaction forces (GRFs) were measured at 1000 Hz over 14 consecutive steps. Impulse, magnitude, and duration of anterior and posterior GRFs were compared between the affected and unaffected limbs at each speed. The net anterior-posterior GRFI, reflecting the changes in horizontal running velocity, was consistently positive (propulsion) in the affected limb and negative (braking) in the unaffected limb at all speeds. Regardless of running speed, unilateral transfemoral amputees maintain constant running speeds not over each step, but over 2 consecutive steps (i.e., one stride).

Unsupervised Gait Event Identification with a Single Wearable Accelerometer and/or Gyroscope: A Comparison of Methods across Running Speeds, Surfaces, and Foot Strike Patterns

We evaluated 18 methods capable of identifying initial contact (IC) and terminal contact (TC) gait events during human running using data from a single wearable sensor on the shank or sacrum. We adapted or created code to automatically execute each method, then applied it to identify gait events from 74 runners across different foot strike angles, surfaces, and speeds. To quantify error, estimated gait events were compared to ground truth events from a time-synchronized force plate. Based on our findings, to identify gait events with a wearable on the shank, we recommend the Purcell or Fadillioglu method for IC (biases +17.4 and -24.3 ms; LOAs -96.8 to +131.6 and -137.0 to +88.4 ms) and the Purcell method for TC (bias +3.5 ms; LOAs -143.9 to +150.9 ms). To identify gait events with a wearable on the sacrum, we recommend the Auvinet or Reenalda method for IC (biases -30.4 and +29.0 ms; LOAs -149.2 to +88.5 and -83.3 to +141.3 ms) and the Auvinet method for TC (bias -2.8 ms; LOAs -152.7 to +147.2 ms). Finally, to identify the foot in contact with the ground when using a wearable on the sacrum, we recommend the Lee method (81.9% accuracy).

Accurate estimation of peak vertical ground reaction force using the duty factor in level treadmill running

This study aimed to (1) construct a statistical model (SMM) based on the duty factor (DF) to estimate the peak vertical ground reaction force ( F v , max ) and (2) to compare the estimated F v , max to force plate gold standard (GSM). One hundred and fifteen runners ran at 9, 11, and 13 km/h. Force (1000 Hz) and kinematic (200 Hz) data were acquired with an instrumented treadmill and an optoelectronic system, respectively, to assess force-plate and kinematic based DFs. SMM linearly relates F v , max to the inverse of DF because DF was analytically associated with the inverse of the average vertical force during ground contact time and the latter was very highly correlated to F v , max . No systematic bias and a 4% root mean square error (RMSE) were reported between GSM and SMM using force-plate based DF values when considering all running speeds together. Using kinematic based DF values, SMM reported a systematic but small bias (0.05BW) and a 5% RMSE when considering all running speeds together. These findings support the use of SMM to estimate F v , max during level treadmill runs at endurance speeds if underlying DF values are accurately measured.

A Single Sacral-Mounted Inertial Measurement Unit to Estimate Peak Vertical Ground Reaction Force, Contact Time, and Flight Time in Running

Peak vertical ground reaction force (Fz,max), contact time (tc), and flight time (tf) are key variables of running biomechanics. The gold standard method (GSM) to measure these variables is a force plate. However, a force plate is not always at hand and not very portable overground. In such situation, the vertical acceleration signal recorded by an inertial measurement unit (IMU) might be used to estimate Fz,max, tc, and tf. Hence, the first purpose of this study was to propose a method that used data recorded by a single sacral-mounted IMU (IMU method: IMUM) to estimate Fz,max. The second aim of this study was to estimate tc and tf using the same IMU data. The vertical acceleration threshold of an already existing IMUM was modified to detect foot-strike and toe-off events instead of effective foot-strike and toe-off events. Thus, tc and tf estimations were obtained instead of effective contact and flight time estimations. One hundred runners ran at 9, 11, and 13 km/h. IMU data (208 Hz) and force data (200 Hz) were acquired by a sacral-mounted IMU and an instrumented treadmill, respectively. The errors obtained when comparing Fz,max, tc, and tf estimated using the IMUM to Fz,max, tc, and tf measured using the GSM were comparable to the errors obtained using previously published methods. In fact, a root mean square error (RMSE) of 0.15 BW (6%) was obtained for Fz,max while a RMSE of 20 ms was reported for both tc and tf (8% and 18%, respectively). Moreover, even though small systematic biases of 0.07 BW for Fz,max and 13 ms for tc and tf were reported, the RMSEs were smaller than the smallest real differences [Fz,max: 0.28 BW (11%), tc: 32.0 ms (13%), and tf: 32.0 ms (30%)], indicating no clinically important difference between the GSM and IMUM. Therefore, these results support the use of the IMUM to estimate Fz,max, tc, and tf for level treadmill runs at low running speeds, especially because an IMU has the advantage to be low-cost and portable and therefore seems very practical for coaches and healthcare professionals.

Relevance of Frequency-Domain Analyses to Relate Shoe Cushioning, Ground Impact Forces and Running Injury Risk: A Secondary Analysis of a Randomized Trial With 800+ Recreational Runners

Cushioning systems in running shoes are used assuming that ground impact forces relate to injury risk and that cushioning materials reduce these impact forces. In our recent trial, the more cushioned shoe version was associated with lower injury risk. However, vertical impact peak force was higher in participants with the Soft shoe version. The primary objective of this study was to investigate the effect of shoe cushioning on the time, magnitude and frequency characteristics of peak forces using frequency-domain analysis by comparing the two study groups from our recent trial (Hard and Soft shoe group, respectively). The secondary objective was to investigate if force characteristics are prospectively associated with the risk of running-related injury. This is a secondary analysis of a double-blinded randomized trial on shoe cushioning with a biomechanical running analysis at baseline and a 6-month follow-up on running exposure and injury. Participants (n = 848) were tested on an instrumented treadmill at their preferred running speed in their randomly allocated shoe condition. The vertical ground reaction force signal for each stance phase was decomposed into the frequency domain using the discrete Fourier transform. Both components were recomposed into the time domain using the inverse Fourier transform. An analysis of variance was used to compare force characteristics between the two study groups. Cox regression analysis was used to investigate the association between force characteristics and injury risk. Participants using the Soft shoes displayed lower impact peak force (p < 0.001, d = 0.23), longer time to peak force (p < 0.001, d = 0.25), and lower average loading rate (p < 0.001, d = 0.18) of the high frequency signal compared to those using the Hard shoes. Participants with low average and instantaneous loading rate of the high frequency signal had lower injury risk [Sub hazard rate ratio (SHR) = 0.49 and 0.55; 95% Confidence Interval (CI) = 0.25–0.97 and 0.30–0.99, respectively], and those with early occurrence of impact peak force (high frequency signal) had greater injury risk (SHR = 1.60; 95% CI = 1.05–2.53). Our findings may explain the protective effect of the Soft shoe version previously observed. The present study also demonstrates that frequency-domain analyses may provide clinically relevant impact force characteristics.

Clinical Trial Registration:https://clinicaltrials.gov/, identifier: 9NCT03115437.

Low-pass filter cutoff frequency affects sacral-mounted inertial measurement unit estimations of peak vertical ground reaction force and contact time during treadmill running

Inertial measurement units (IMUs) are popular tools for estimating biomechanical variables such as peak vertical ground reaction force (GRF_v) and foot-ground contact time (t_c), often by using multiple sensors or predictive models. Despite their growing use, little is known about the effects of varying low-pass filter cutoff frequency, which can affect the magnitude of force-related dependent variables, the accuracy of IMU-derived metrics, or if simpler methods for such estimations exist. The purpose of this study was to investigate the effects of varying low-pass filter cutoff frequency on the correlation of IMU-derived peak GRF_v and t_c to gold-standard lab-based measurements. Thirty National Collegiate Athletics Association Division 1 cross country runners ran on an instrumented treadmill at a range of speeds while outfitted with a sacral-mounted IMU. A simple method for estimating peak GRF_v from the IMU was implemented by multiplying the IMU's vertical acceleration by the runner's body mass. Data from the IMU were low-pass filtered with 5, 10, and 30 Hz cutoffs. Pearson correlation coefficients were used to determine how well the IMU-derived estimates matched gold-standard biomechanical estimations. Correlations ranged from very weak to moderate for peak GRF_v and t_c. For peak GRF_v, the 10 Hz low-pass filter cutoff performed best (r = 0.638), while for t_c the 5 Hz cut-off performed best (r = 0.656). These results suggest that IMU-derived estimates of force and contact time are influenced by the low-pass filter cutoff frequency. Further investigations are needed to determine the optimal low-pass filter cutoff frequency or a different method to accurately estimate force and contact time is suggested.

Vibration Suppression of a Composite Prosthetic Foot Using Piezoelectric Shunt Damping: Implications to Vibration-Induced Cumulative Trauma

OBJECTIVE

Energy-storage-and-return (ESAR) prosthetic feet have improved amputee mobility due to their efficient conversion of strain energy to mechanical work. However, this efficiency is typically achieved using light-weight, high-stiffness materials, which generate high-frequency vibrations that are potentially injurious if transmitted to biological tissues. To reduce the vibration which may cause cumulative tissue trauma, high-frequency vibration suppression by piezoelectric shunt damping patches on a commercial ESAR foot was evaluated.

METHODS

Two patches with either passive or active shunt circuits were placed on the foot to investigate vibration suppression during experimental tests where a plastic hammer was used to hit a clamped ESAR foot on the free end. Prosthesis bending moments at each modal frequency were obtained by finite element methods to identify piezoelectric patch placement.

RESULTS

Both shunt circuits decreased vibration amplitudes at specific modes better than the no shunt case, but also increased the amplitude at specific frequencies. The vibration suppression performance of the active shunt circuit deteriorated at the second mode, while the vibration suppression performance of the passive shunt circuit deteriorated at all frequencies above the third mode.

CONCLUSIONS

These results indicate piezoelectric shunt patches may be a viable strategy for decreasing vibrations of an ESAR foot, with active methods more efficient at suppressing high-frequency vibrations. Additional research is necessary to fine-tune the method for maximal vibration suppression.

SIGNIFICANCE

Overall, this study indicates that high-frequency vibration suppression is possible using piezoelectric patches, possibly decreasing the cumulative tissue damage that may occur with repetitive exposure to vibration.

Filtration Selection and Data Consilience: Distinguishing Signal from Artefact with Mechanical Impact Simulator Data

A large variety of data filtration techniques exist in biomechanics literature. Data filtration is both an 'art' and a 'science' to eliminate noise and retain true signal to draw conclusions that will direct future hypotheses, experimentation, and technology development. Thus, data consilience is paramount, but is dependent on filtration methodologies. In this study, we utilized ligament strain, vertical ground reaction force, and kinetic data from cadaveric impact simulations to assess data from four different filters (12 vs. 50 Hz low-pass; forward vs. zero lag). We hypothesized that 50 Hz filtered data would demonstrate larger peak magnitudes, but exhibit consilience of waveforms and statistical significance as compared to 12 Hz filtered data. Results demonstrated high data consilience for matched pair t test correlations of peak ACL strain (≥ 0.97), MCL strain (≥ 0.93) and vertical ground reaction force (≥ 0.98). Kinetics had a larger range of correlation (0.06-0.96) that was dependent on both external load application and direction of motion monitored. Coefficients of multiple correlation demonstrated high data consilience for zero lag filtered data. With respect to in vitro mechanical data, selection of low-pass filter cutoff frequency will influence both the magnitudes of discrete and waveform data. Dependent on the data type (i.e., strain and ground reaction forces), this will not likely significantly alter conclusions of statistical significance previously reported in the literature with high consilience of matched pair t-test correlations and coefficients of multiple correlation demonstrated. However, rotational kinetics are more sensitive to filtration selection and could be suspect to errors, especially at lower magnitudes.

Accelerometer-based prediction of running injury in National Collegiate Athletic Association track athletes

Running-related injuries (RRI) may result from accumulated microtrauma caused by combinations of high load magnitudes (vertical ground reaction forces; vGRFs) and numbers (strides). Yet relationships between vGRF and RRI remain unclear - potentially because previous research has largely been constrained to collecting vGRFs in laboratory settings and ignoring relationships between RRI and stride number. In this preliminary proof-of-concept study, we addressed these constraints: Over a 60-day period, each time collegiate athletes (n = 9) ran they wore a hip-mounted activity monitor that collected accelerations throughout the entire run. Accelerations were used to estimate peak vGRF, number of strides, and weighted cumulative loading (sum of peak vGRFs weighted to the 9th power) across the entirety of each run. Runners also reported their post-training pain/fatigue and any RRI that prevented training. Across 419 runs and >2.1 million strides, injured (n = 3) and uninjured (n = 6) participants did not report significantly different pain/fatigue (p = 0.56) or mean number of strides per run (p = 0.91). Injured participants did, however, have significantly greater peak vGRFs (p = 0.01) and weighted cumulative loading per run (p < 0.01). Results from this small but extensively studied sample of elite runners demonstrate that loading profiles (load magnitude-number combinations) quantified with activity monitors can provide valuable information that may prove essential for: (1) testing hypotheses regarding overuse injury mechanisms, (2) developing injury-prediction models, and (3) designing and adjusting athlete- and loading-specific training programs and feedback.