Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Jan 22.
Published in final edited form as: J Biomech. 2020 Dec 3;115:110139. doi: 10.1016/j.jbiomech.2020.110139

Mechanisms used to increase propulsive forces on a treadmill in older adults

Erica A Hedrick 1,*, Sheridan M Parker 1, HaoYuan Hsiao 2, Brian A Knarr 1
PMCID: PMC7855876  NIHMSID: NIHMS1654802  PMID: 33321429

Abstract

Older adults typically demonstrate reductions in overground walking speeds and propulsive forces compared to young adults. These reductions in walking speeds are risk factors for negative health outcomes. Therefore, this study aimed to determine the effect of an adaptive speed treadmill controller on walking speed and propulsive forces in older adults, including the mechanisms and strategies underlying any change in propulsive force between conditions. Seventeen participants completed two treadmill conditions, one with a fixed comfortable walking speed and one with an adaptive speed controller. The adaptive speed treadmill controller utilized a set of inertial-force, gait parameters, and position-based controllers that respond to an instantaneous anterior inertial force. A biomechanical-based model previously developed for individuals post-stroke was implemented for older adults to determine the primary gait parameters that contributed to the change in propulsive forces when increasing speed. Participants walked at faster average speeds during the adaptive speed controller (1.20 m/s) compared to the fixed speed controller conditions (0.98 m/s); however, these speeds were not as fast as their overground speed (1.44 m/s). Although average trailing limb angle (TLA) (p<0.001) and ankle moment (p=0.020) increased when speed also increased between treadmill conditions, increasing TLA contributed more to the increased propulsive forces seen during faster treadmill speeds. Our findings show that older adults chose faster walking speeds and increased propulsive force when walking on an adaptive speed treadmill compared to a fixed speed treadmill, suggesting that an adaptive speed treadmill controller has the potential to be a beneficial alternative to current exercise interventions for older adults.

Keywords: adaptive speed treadmill, trailing limb angle, ankle moment, propulsion, walking speed

1. Introduction

Older adults typically have a slower preferred walking speed compared to younger adults (Kerrigan et al., 1998). A slower walking speed is related to cognitive impairment, fall risk, disability, and overall mortality in older adults (Abellan Van Kan et al., 2009). Compared to individuals with lower functional fitness, older adults with higher functional fitness and overall health have been shown to have higher walking speeds (Paulson and Gray, 2015; Studenski et al., 2011). Thus, improving the functional fitness level of older adults should increase their walking speed (Liu and Latham, 2009) and in turn may reduce the risk factors listed above. Treadmill walking is a common tool used to increase fitness and to reduce fall risk in older adults (Gerards et al., 2017; Shimada et al., 2004). However, older adults typically have a decreased preferred walking speed on a treadmill compared to overground walking (Malatesta et al., 2017; Nagano et al., 2013). One reason could be that typical treadmills impose a constant speed that can constrain movements (Terrier and Deriaz, 2011). For example, older adults have been shown to walk with short step lengths and to increase the proportion of double support when walking on the treadmill, presumably in order to preserve balance(Nagano et al., 2013). Not constraining walking speed on a treadmill could lead to increases in preferred walking speeds by allowing older adults to alter their gait patterns.

One mechanism that can be used to attempt to increase walking speed is to use a treadmill with variable speed. The idea is to better emulate overground walking while on a treadmill. Having a variable speed allows the individual to vary their movements to fluctuate their speed, which is more comparable to overground walking than a treadmill with a fixed speed. A previous study investigated the changes in propulsive forces using an adaptive speed treadmill controller that utilized a set of inertial-force, gait parameters, and position-based controllers that respond to an instantaneous anterior inertial force (Ray et al., 2020, 2018). This study found that young adults walked faster on an adaptive speed treadmill controller compared to a fixed speed controller as well as increased propulsive forces (Ray et al., 2018). Therefore, this type of treadmill controller could allow older adults to increase their walking speed, making it a promising exercise intervention for older adults.

The altered kinematics and kinetics of older adults can contribute to their slower gait speeds. During walking at their preferred speeds, older adults demonstrate reductions in peak propulsive forces, hip range of motion, ankle plantarflexor moment and ankle power compared to younger adults (Anderson and Madigan, 2014; Conway and Franz, 2020; Franz et al., 2014; Franz and Kram, 2013). While these reductions in forces are known, previous literature has shown that older adults have a propulsive force reserve that provides them with the ability to increase or manipulate their propulsive forces (Browne and Franz, 2018; Franz et al., 2014) and ankle power generation (Browne and Franz, 2019) with the use of visual biofeedback or more demanding tasks. Additionally, when recalling this increase in ankle power during overground walking (without visual biofeedback), they were able to walk 11% faster than their preferred walking speed (Browne and Franz, 2019). Thus, a device that engages greater propulsive force generation could promote faster walking speed in older adults. Walking on an adaptive speed treadmill has the potential to utilize this propulsive force reserve in older adults, but if so, what strategy they use to increase their propulsive forces is unknown.

The majority of changes in propulsive forces can be attributed to trailing limb angle (TLA) and ankle moment, for which a model has been validated in young adults and individuals post-stroke (Hsiao et al., 2015; Hsiao et al., 2015). When older adults walk at their maximum speed, they can improve their push-off intensity by increasing their TLA and/or ankle moment. However, peak ankle plantarflexor moment at fast walking speeds in older adults is still lower than that of younger adults at a preferred walking speed (Conway and Franz, 2020). Finding how both TLA and ankle moment contribute to propulsive forces and their effect on increasing walking speed could provide useful information that guides intervention efforts.

The purpose of the study was to determine the effect of an adaptive speed treadmill controller on an older adult’s walking speed and propulsive force. A secondary purpose was to quantify the relative contribution of ankle moment and TLA to changes in propulsive force. We hypothesized that older adults would increase their walking speed and propulsive forces when walking on the adaptive speed treadmill as compared to walking at a fixed speed because the adaptive speed treadmill controller allows for more unconstrained changes in walking speed. This in theory would better resemble overground walking, and this relationship has been seen in young adult populations that have used this controller (Ray et al., 2018). Additionally, we hypothesized that propulsive force would be increased mainly by an increase in trailing limb angle because older adults have been shown to overcome deficits, compared to young adults, in TLA when walking faster (Conway and Franz, 2020).

2. Methods

2.1. Data Collection

Seventeen older adult participants were included in this analysis (70.24 ± 5.25 yrs., 73.68 ± 12.52 kg, 1.64 ± 0.09 m). They were recruited from wellness centers, recruitment flyers, and by word of mouth. General inclusion criteria required subjects to be between 60–85 years old. Exclusion criteria were a diagnosis of osteoporosis, and/or a diagnosis of a neurological disorder including but not limited to stroke, traumatic brain injury, Alzheimer’s, and dementia. Eighteen total participants were recruited and completed the study, but data from one subject were excluded due to equipment malfunctions. Informed consent was collected from all participants and the study was approved by the University of Nebraska Medical Center Institutional Review Board.

Participants were asked to complete two treadmill walking conditions: one fixed speed and one adaptive speed condition, along with an overground 10-meter walk test. The two treadmill walking conditions were randomized to account for a potential crossover or learning effect. Lower extremity kinematics and kinetics were collected using a 14-camera motion capture system (Vicon, Oxford, UK) with a 100 Hz capture rate and an instrumented split-belt treadmill (Bertec, Columbus OH, USA) with a 1000 Hz sampling rate. Motion capture markers were placed using a custom marker set with markers placed bilaterally at the feet, ankles, knees, greater trochanters, and pelvis. Marker shells were placed bilaterally at the thigh and shank segments.

For the fixed speed walking condition, participants walked at their self-selected walking speed for three minutes. To determine the participant’s self-selected walking speed, the treadmill was initially set to 0.5 m/s and increased by 0.1 m/s every ten seconds until the participant verbally indicated they were walking at a comfortable speed (Gault et al., 2013; Mannering et al., 2017), and the participant proceeded to walk at their comfortable speed for thirty seconds before stopping the treadmill. Once the acclimatization was over, the treadmill was restarted for the three-minute trial at the self-selected speed.

For the adaptive speed treadmill walking condition, a novel treadmill controller was used (Ray et al., 2018). The adaptive speed treadmill controller started at 1 m/s (Zeni Jr. and Higginson, 2009) and changed both belts speed in response to the impulse of the instantaneous anterior inertial force, step length and duration, and the position of the participant relative to the center of the treadmill, to determine each participant’s self-selected walking speed. Calculations for the controller are made concurrently for both limbs concerning foot placement on the treadmill and walking phase in D-Flow (Motek Medical, Norwell, MA, USA) with a 300Hz sampling rate. A minimum two minute acclimation period was performed on the adaptive speed treadmill, and more time was given if needed (Lee and Hidler, 2008). Once comfortable, the treadmill was stopped and then restarted for the actual data trial. Participants were instructed to walk at a comfortable pace “like walking in the park” for three minutes. Participants were encouraged to not use the handrails for either treadmill condition but had the option to if needed.

For the overground walking speed, a 10-meter walk test was performed over a flat walkway. Participants were asked to walk at a comfortable pace “like walking in the park”. The time to complete the middle 6 meters was collected using a stopwatch. Participants were asked to perform this test twice.

2.2. Data Analysis

Kinematic and kinetic data were processed using Nexus (VICON, Oxford, UK) and calculations were performed in Visual 3D software (C-Motion, Inc., Germantown, MD, USA) as well as MATLAB (Mathworks, Natick, MA, USA). Kinetic data were filtered at 60 Hz while marker data were filtered at 6 Hz, both using a 4th order low pass Butterworth filter. To determine preferred overground walking speed, the average of the two trials to travel the middle 6 meters of the 10-meter walking task was used. The preferred walking speed on the fixed speed controller condition was read directly from the treadmill. For the adaptive speed controller, the average speed from 30 seconds to three minutes of walking was calculated from the real-time speed component of the controller, to disregard the initial acceleration of the treadmill.

Ankle moment, TLA, and lever arm of the GRF have been shown to significantly predict changes in propulsive forces for individuals post-stroke (Hsiao et al., 2015). Therefore, peak propulsive force, ankle moment, TLA, and lever arm at the peak propulsive force were extracted from the data. Peak propulsive force was defined as the maximum anterior GRF during the second half of stance phase (propulsion phase). Ankle moment was defined as the ankle plantarflexion moment at the instant of peak propulsion during stance. TLA was defined in the sagittal plane as the angle between the laboratory’s vertical axis and the vector joining the greater trochanter with the center of pressure and defined at the instant of peak propulsion (Figure 1). The lever arm at the peak propulsive force was defined as the perpendicular distance from the ankle joint to the vector joining the greater trochanter with the COP. It was assumed that these healthy older adults had symmetrical gait mechanics; therefore the right and left sides were combined, and the averages of the combination were calculated.

Figure 1.

Figure 1.

Open in a new tab

Diagram illustrating the variables of interest, including the propulsive force, ankle moment, TLA, and lever arm, adapted from (Hsiao et al., 2015).

Each variable’s contribution to the changes in propulsive force during increases in speed was determined using the model presented in a previous study (Hsiao et al., 2015). In order to determine the variables’ contribution to any change in speed between conditions, the calculations were done by comparing the faster speed of the two treadmill conditions to the slower speed per subject. The four components assessed for contributing to the change in propulsive force were: TLA, ankle moment, the interaction between changes in TLA and ankle moment, and lever arm length. A decrease in the contributing variable while the propulsive force increased would suggest that this variable had no contribution to the increase in propulsive force. Therefore, all negative values were set to 0, signifying no contribution to the increase in propulsive force between conditions, and the relative contributions were calculated by dividing each term by the sum of all terms.

2.3. Statistical Analysis

A single factor ANOVA was used to determine if there was a difference in average walking speed between the three conditions. If a main effect was found, a post-hoc paired t-test, with a Bonferroni correction, was done between each condition. To determine if the walking speeds between the two treadmill conditions or the treadmill controller itself caused differences in the participant’s gait mechanics, a multivariate regression was used to quantify the portion of the change in peak propulsive force, ankle moment and TLA at peak propulsion. The coefficients for each variable as well as the overall coefficient of determination (adjusted R2) of the model were used to determine these relationships. Additionally, a linear regression was used to determine the relationship between speed and both peak propulsion and TLA.

To evaluate the model used to predict the treadmill conditions peak propulsive forces, a Pearson’s correlation coefficient (r) was used to assess the relationship between the predicted and the measured peak propulsive force for each condition. This analysis was repeated for the change in measured and predicted peak propulsive force. Multiple paired t-tests were performed to detect the change in peak propulsion, TLA, ankle moment, and lever arm for the slower average speed condition compared to the faster average speed condition per subject. The significance level was set at an alpha of 0.05. Statistical analysis was performed in MATLAB (Mathworks, Natick, MA, USA) and IBM SPSS Statistics (IBM Corp., Armonk, NY, USA).

3. Results

Participants had significantly different average walking speeds between conditions (F=17.01, p<0.001). Average speed increased on the adaptive speed treadmill controller (1.20 ± 0.30 m/s) compared to the fixed speed (0.98 ± 0.20 m/s) treadmill (p=0.003) (Figure 2). On an individual basis, thirteen of the seventeen subjects increased their speed on the adaptive speed compared to the fixed speed treadmill controller. However, participants walked faster overground (1.44 ± 0.18 m/s) than on the adaptive (p=0.003) and fixed speed (p<0.001) treadmill controller. Results of the multivariate regression revealed that the difference in speed between the two treadmill conditions explained a significant percentage of the variance in the main dependent variables (peak propulsive force, TLA, ankle moment), while the controller did not (Table 1). Specifically, 68% (p<0.001) of the variance in peak propulsive force, 56% (p<0.001) of the variance in TLA at peak propulsion, and 27% (p=0.002) of the variance in ankle moment at peak propulsion was explained by the difference in speed between the treadmill conditions. Speed and peak propulsive force (R2 =0.698) as well as speed and TLA (R2=0.578) had a significant linear relationship from the linear regression (Figure 3).

Figure 2.

Figure 2.

Open in a new tab

The average walking speed was significantly different between conditions (p<0.001). Specifically, the walking speed for the adaptive speed treadmill condition was greater than the fixed speed condition (p=0.003) and the average walking speed overground was greater than the adaptive speed treadmill condition (p=0.003) and the fixed speed treadmill condition (p<0.001). The individual data points represent each participant. The filled in black circles are the participants that used handrails while walking and the filled in white circles are the participants that walked without handrails. *p<0.0167

Table 1.

Results from the multiple linear regression. Speed had a significant effect on TLA, peak propulsive force (p<0.001), and ankle moment (p=0.002) while the controller did not (p>0.05).

Dependent Variable Predictor Variables R R2 B SE B β P
TLA at Peak Propulsion (deg) Intercept 0.770 0.567 4.395 1.272 0.002
Controller −0.299 0.652 −0.057 0.650
Speed (m/s) 7.724 1.221 0.792 <0.001
Peak Propulsion (N/BW) Intercept 0.841 0.689 0.024 0.015 0.127
Controller −0.002 0.008 −0.028 0.796
Speed (m/s) 0.118 0.015 0.852 <0.001
Ankle Moment at Peak Propulsion (Nm/BW) Intercept 0.525 0.229 0.083 0.009 <0.001
Controller −0.004 0.004 −0.143 0.398
Speed (m/s) 0.028 0.008 0.566 0.002
Open in a new tab

Figure 3.

Figure 3.

Open in a new tab

The relationship between walking speed and (A) peak propulsive force normalized to the participant’s body weight (R2=0.698) and (B) TLA at peak propulsion (R2=0.578). Individual data points for the fixed speed (FS) are in blue and adaptive speed (AST) treadmill condition are in red.

The model used to predict peak propulsive force was the combination of ankle moment, sin(TLAcop), the interaction of ankle moment and sin(TLAcop), and the inverse of the lever arm length (d). The measured values and predicted values from the model were strongly correlated for each treadmill condition, as well as the change in between conditions (Figures 4 & 5). The model predicted more than 83% of the variance in propulsive forces for the adaptive speed condition and 92% of the variance in the fixed speed condition. Additionally, the model predicted almost 81% of the variance in the changes in propulsive forces between conditions. Peak propulsive force (97.01 ± 26.88 N vs 121.54 ± 31.53 N, p<0.001), TLA at peak propulsion (11.66 ± 2.47 deg vs 13.67 ± 2.49 deg, p<0.001), and ankle moment at peak propulsion (77.98 ± 14.54 Nm vs 83.26 ± 19.01 Nm, p=0.020) all significantly increased with an increase in speed from the slower treadmill condition to the faster treadmill condition, while inverse lever arm did not (6.80 ± 0.52 m vs 6.68 ± 0.77 m, p=0.31) (Figure 6).

Figure 4.

Figure 4.

Open in a new tab

Relationship between the measured propulsive force and the predicted peak propulsive force from the model equation normalized to the participant’s body weight for both (A) fixed speed (blue) and (B) adaptive speed (red). The model could accurately predict the peak propulsive force for both treadmill conditions (p<0.001). The best fit line to the data is represented by the dotted line.

Figure 5.

Figure 5.

Open in a new tab

Relationship between the change in measured propulsive force between the two treadmill conditions and the predicted change in propulsive force between the two conditions. The model can predict about 80% of the changes in peak propulsive force between the two conditions. The best fit line to the data is represented by the dotted line.

Figure 6.

Figure 6.

Open in a new tab

Average and standard deviations of (A) peak propulsive force normalized to participant’s body weight, (B) participant’s TLA at peak propulsive force, (C) ankle moment at peak propulsive force normalized to body weight and (D) inverse lever arm length at peak propulsive force for each subject’s slow and fast condition. Peak propulsive force (p<0.001), TLA at peak propulsive force (p<0.001), and ankle moment at peak propulsive force (p=0.020) all significantly increased with the faster condition. The individual data points represent each participant. The filled in black circles are the participants that used handrails while walking and the filled in white circles are the participants that walked without handrails. *p<0.05

On average, the contribution of TLA and ankle moment to increases in propulsive forces were 60%, and 22%, respectively. Four of the seventeen participants reported greater speeds on the fixed speed compared to the adaptive speed treadmill and had contributions from both TLA (41 ± 31%) and ankle moment (29% ± 24%), or an interaction of both (30 ± 47%)). All thirteen participants that increased speed on the adaptive speed treadmill mainly utilized increasing their trailing limb angle to increase propulsion (66 ± 19%). The relative contributions to propulsive forces, based on the four components mentioned previously, are quantified in Table 2.

Table 2.

Walking speeds and relative contribution to increases in propulsive forces from changes in each variable. TLA = trailing limb angle, Ma = ankle moment, Mix = interaction of ankle moment and TLA, d= lever arm of GRF. Relative contributions are calculated from the treadmill condition with the lower speed to the treadmill condition at the higher speed. Four subjects had the fixed speed treadmill controller as their faster condition (FS), and thirteen subjects had the adaptive speed treadmill controller as their faster condition (AST).

Subject Age Weight (kg) Height (m) Faster Condition Overground Speed (m/s) Fixed Speed (m/s) AST Speed (m/s) Relative Contribution to propulsion Handrail Use
TLA Ma mix d
1 72 102.5 1.9 FS 1.57 1.10 0.86 0.56 0.33 0.11 0.00 N
2 74 61.235 1.615 AST 1.83 1.30 1.63 0.98 0.00 0.00 0.02 N
3 67 68.89 1.66 FS 1.36 1.10 0.99 0.38 0.57 0.05 0.00 N
4 69 74.4 1.53 AST 1.33 1.10 1.21 0.51 0.34 0.02 0.13 N
5 70 68.5 1.62 AST 1.67 0.80 1.49 0.46 0.39 0.14 0.00 N
6 61 68.8 1.72 AST 1.47 1.10 1.53 0.58 0.38 0.05 0.00 N
7 76 49.9 1.55 FS 1.32 0.70 0.68 0.00 0.00 1.00 0.00 N
8 64 70.76 1.65 FS 1.36 0.80 0.76 0.71 0.25 0.04 0.00 N
9 76 56.7 1.62 AST 1.50 1.00 1.36 1.00 0.00 0.00 0.00 N
10 70 78.02 1.57 AST 1.71 0.90 0.94 0.70 0.00 0.00 0.30 N
11 75 78.02 1.66 AST 1.32 0.80 0.96 0.51 0.44 0.05 0.00 N
12 61 62.6 1.63 AST 1.59 1.30 1.51 0.75 0.23 0.02 0.00 N
13 67 77.11 1.76 AST 1.29 0.90 1.27 0.66 0.26 0.08 0.00 N
14 68 84.82 1.6 AST 1.25 0.60 1.09 0.74 0.17 0.10 0.00 N
15 70 83.5 1.62 AST 1.34 1.10 1.41 0.79 0.17 0.04 0.00 Y
16 79 82.55 1.655 AST 1.26 1.00 1.51 0.48 0.18 0.05 0.30 Y
17 75 84.3 1.58 AST 1.34 1.10 1.16 0.44 0.00 0.00 0.56 Y
Mean 70.24 73.68 1.64 1.44 0.98 1.20 0.60 0.22 0.10 0.08
SD 5.26 12.52 0.09 0.18 0.20 0.30 0.24 0.18 0.24 0.16
Max 79.00 102.50 1.90 1.83 1.30 1.63 1.00 0.57 1.00 0.56
Min 61.00 49.90 1.53 1.25 0.60 0.68 0.00 0.00 0.00 0.00
Open in a new tab

4. Discussion

The purpose of the current study was to determine the effect of the adaptive speed treadmill on walking speed and propulsive forces for older adults. In support of our hypothesis, older adults were able to increase their walking speed and propulsive force when walking on the adaptive speed treadmill compared to the fixed speed treadmill. The increase in average walking speed from the fixed (0.98 m/s) to the adaptive (1.20 m/s) speed treadmill conditions was greater than the minimal detectable change for older adults (0.124 m/s) (Hars et al., 2013). This suggests that adaptive speed treadmill walking could be an effective approach for locomotor training that targets walking speed and better encourages individuals to walk at speeds closer to their overground walking speed. However, the average adaptive speed treadmill controller speed was significantly lower than their average overground speed. This difference in speed is in agreement with a study that had individuals post-stroke use this same treadmill controller (Ray et al., 2020). Furthermore, the participants were mainly active adults, and previous research has shown that more active adults have higher overground walking speeds compared to lower-functioning groups (Paulson and Gray, 2015). While treadmill speed is correlated with overall activity level (Carlson et al., 2012), older adults also take longer to adapt to walking on a treadmill, compared to overground (Wass et al., 2005). This could explain why these participants were able to increase their speed slightly on the adaptive speed treadmill, but they may not have adapted to the treadmill enough to increase their speed to their overground speed.

One strength of this study was the ability to isolate the effect of the adaptive speed treadmill controller itself since both treadmill conditions were done on the same treadmill. We found that speed was a significant predictor of propulsive forces and TLA, not the differences between the participants or the controller itself. This agrees with a previous study with young adults using the same treadmill controller (Ray et al., 2018). While the treadmill controller itself did not significantly contribute to the changes in propulsive force, the majority of individuals performed a faster walking speed when walking on the adaptive speed controller than they did during the self-selected fixed speed condition, which is beneficial for exercise interventions.

There are a variety of user-driven treadmill controllers that have tried to emulate overground walking by not constraining walking speed to a fixed speed. These previous treadmills employ different strategies, including changing speed based on feedback from kinetic (Hejrati et al., 2015; Koenig et al., 2009) or kinematic measures (Kim et al., 2013; Minetti et al., 2003; Yoon et al., 2012), but none have combined kinetic and kinematic measures with a controller. The adaptive speed treadmill used in this study employs a multi-faceted approach, using an instrumented treadmill to capture forces and center of pressure to combine inertial-force, position, and gait parameters to alter the speed (Ray et al., 2020, 2018). Thus, the treadmill controller can change speed based on changes in propulsive force and step length. Using multiple variables to change speed can give more information on walking performance and could allow a more rapid and fluid response from the user. It also has the potential to have a more usable experience and may be a better translation of overground walking. Though, futures studies directly comparing adaptive speed treadmills with different inputs and parameters weighting will need to be done in order to determine the most effective control paradigm that emulates overground walking.

The model used (Hsiao et al., 2015) accurately describes the relationship between ankle moment, TLA, the lever arm of the GRF, and propulsive forces for older adults. Participants in the current study walked at a variety of walking speeds (0.6 to 1.63 m/s), therefore the model works for a wide range of walking speeds. In agreement with previous studies (Conway and Franz, 2020), we found older adults increased their TLA, peak ankle moment, and peak propulsive force when increasing speed. The change in TLA had a greater contribution (60%) to increased propulsion than ankle moment (22%). This agrees with previous studies that indicated young adults and individuals post-stroke mainly use a change in TLA to increase propulsive forces (Hsiao et al., 2015; Hsiao et al., 2015). Hsiao and colleagues reported that in younger adults, 65% of the increase in propulsion was contributed from the increase in TLA and 33.7% of the increase in peak propulsion was contributed from the increase in ankle moment. Our results, therefore, indicate that older adults appeared to have a similar capacity to increase TLA but the ability to modulate ankle moment was reduced compared to younger adults. These findings support previous observation that older adults can overcome age-related deficits in trailing limb extension and propulsive force when walking faster, but their ankle moment is still significantly lower than younger adults (Conway and Franz, 2020). This supports our findings that older adults would use primarily a TLA strategy to increase their propulsive forces.

There were limitations to this study. A relatively slow self-selected fixed walking speed was determined using only an increasing speed protocol instead of using a combination of an increasing and decreasing speed protocol or an overground protocol. Thus, we cannot exclude the possibility that between-condition effects in walking speed and our translational recommendations are sensitive to these methodological decisions. Further study is certainly warranted. Subjects were encouraged not to use the handrails while walking, but three subjects did end up using handrails (Table 2). While we do not know how much force was placed on the handrails or how this might have affected their gait mechanics, the three subjects did use the handrails for both conditions. We believe most of the handrail use effect of handrail use was washed out when comparing the conditions, with the assumption that the force applied to the handrails was similar between conditions. Moreover, these subjects that utilized the handrails employed a TLA strategy for increasing propulsion, which is a similar strategy compared to the rest of the participants. Additionally, as seen in Figures 1 and 5, the individuals that used handrails, shown by the black circles, are within the range of those that did not use handrails, shown by the white circles. Therefore, we concluded that handrail use was not correlated to a specific mechanism of increasing propulsion (Table 2). Another limitation of this study would be a small sample size (N=17). However, despite a small sample size, we were still able to achieve significant increases in speed between the controllers and significant differences in gait mechanics between slow and fast walking speeds.

This study was able to increase the walking speed of healthy older adults on a treadmill by using an adaptive speed controller compared to a fixed speed condition. Most of the participants relied on increasing TLA to increase propulsive force between the two conditions, which confirms that older adults have a propulsive force reserve that can be utilized by increasing their TLA (Conway and Franz, 2020). Walking on the adaptive speed treadmill controller encourages older adults to utilize their propulsive force reserve, which can, in turn, increase their walking speed and functional fitness level. The direct feedback from altering their movements and feeling the change in speed could help older adults better translate the training to overground walking. Therefore, the adaptive speed treadmill has the potential to promote a more beneficial gait strategy during exercise interventions for older adults. Future studies can investigate how a prolonged training period with the adaptive speed treadmill controller affects preferred walking speed during overground walking as well as everyday life in order to solidify the direct translation of this adaptive speed treadmill.

Acknowledgments

The authors acknowledge the support of NIH P20 GM109090, R15 HD094194, & U54 GM115458.

Footnotes

Conflict of Interest Statement

The authors have no conflict of interests.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Abellan Van Kan G, Rolland Y, Andrieu S, Bauer J, Beauchet O, Bonnefoy M, Cesari M, Donini LM, Gillette-Guyonnet S, Inzitari M, Nourhashemi F, Onder G, Ritz P, Salva A, Visser M, Vellas B, 2009. Gait speed at usual pace as a predictor of adverse outcomes in community-dwelling older people an International Academy on Nutrition and Aging (IANA) task force. Journal of Nutrition, Health and Aging 13, 881–889. 10.1007/s12603-009-0246-z [DOI] [PubMed] [Google Scholar]
  2. Anderson DE, Madigan ML, 2014. Healthy older adults have insufficient hip range of motion and plantar flexor strength to walk like healthy young adults. Journal of Biomechanics 47, 1104–1109. 10.1016/j.jbiomech.2013.12.024 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Browne MG, Franz JR, 2019. Ankle power biofeedback attenuates the distal-to-proximal redistribution in older adults. Gait and Posture 71, 44–49. 10.1016/j.gaitpost.2019.04.011 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Browne MG, Franz JR, 2018. More push from your push-off : Joint-level modifications to modulate propulsive forces in old age. Plos One 13, e0201407 10.1371/journal.pone.0201407 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Carlson RH, Huebner DR, Hoarty CA, Whittington J, Haynatzki G, Balas MC, Schenk AK, Goulding EH, Potter JF, Bonasera SJ, 2012. Treadmill gait speeds correlate with physical activity counts measured by cell phone accelerometers. Gait and Posture 36, 241–248. 10.1016/j.gaitpost.2012.02.025 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Conway KA, Franz JR, 2020. Increasing the Propulsive Demands of Walking to Their Maximum Elucidates Functionally Limiting Impairments in Older Adult Gait. Journal of Aging and Physical Activity 28, 1–8. 10.1123/japa.2018-0327 [DOI] [PubMed] [Google Scholar]
  7. Franz JR, Kram R, 2013. Advanced age affects the individual leg mechanics of level, uphill, and downhill walking. Journal of Biomechanics 46, 535–540. 10.1016/j.jbiomech.2012.09.032 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Franz JR, Maletis M, Kram R, 2014. Real-time feedback enhances forward propulsion during walking in old adults. Clinical Biomechanics 29, 68–74. 10.1016/j.clinbiomech.2013.10.018 [DOI] [PubMed] [Google Scholar]
  9. Gault ML, Clements RE, Willems MET, 2013. Cardiovascular responses during downhill treadmill walking at self-selected intensity in older adults. Journal of Aging and Physical Activity 21, 335–347. 10.1123/japa.21.3.335 [DOI] [PubMed] [Google Scholar]
  10. Gerards MHG, McCrum C, Mansfield A, Meijer K, 2017. Perturbation-based balance training for falls reduction among older adults: Current evidence and implications for clinical practice. Geriatrics and Gerontology International 17, 2294–2303. 10.1111/ggi.13082 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Hars M, Herrmann FR, Trombetti A, 2013. Reliability and minimal detectable change of gait variables in community-dwelling and hospitalized older fallers. Gait and Posture 38, 1010–1014. 10.1016/j.gaitpost.2013.05.015 [DOI] [PubMed] [Google Scholar]
  12. Hejrati B, Crandall KL, Hollerbach JM, Abbott JJ, 2015. Kinesthetic force feedback and belt control for the treadport locomotion interface. IEEE Transactions on Haptics 8, 176–187. 10.1109/TOH.2015.2404357 [DOI] [PubMed] [Google Scholar]
  13. Hsiao Haoyuan, Knarr BA, Higginson JS, Binder-Macleod SA, 2015. Mechanisms to increase propulsive force for individuals poststroke. Journal of NeuroEngineering and Rehabilitation 12, 1–8. 10.1186/s12984-015-0030-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Hsiao HaoYuan, Knarr BA, Higginson JS, Binder-Macleod SA, 2015. The relative contribution of ankle moment and trailing limb angle to propulsive force during gait. Human Movement Science 39, 212–221. 10.1016/j.humov.2014.11.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Kerrigan DC, Todd MK, Della Croce U, Lipsitz LA, Collins JJ, 1998. Biomechanical gait alterations independent of speed in the healthy elderly: Evidence for specific limiting impairments. Archives of Physical Medicine and Rehabilitation 79, 317–322. 10.1016/S0003-9993(98)90013-2 [DOI] [PubMed] [Google Scholar]
  16. Kim J, Gravunder A, Stanley CJ, Park HS, 2013. Low-cost implementation of a self-paced treadmill by using a commercial depth sensor, in: 35th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) IEEE, pp. 874–877. 10.1109/EMBC.2013.6609640 [DOI] [PubMed] [Google Scholar]
  17. Koenig A, Member S, Binder C, Zitzewitz J, Omlin X, Bolliger M, Riener R, 2009. Voluntary gait speed adaptation for robot-assisted treadmill training, in: IEEE International Conference on Rehabilitation Robotics pp. 419–424. [Google Scholar]
  18. Lee SJ, Hidler J, 2008. Biomechanics of overground vs. treadmill walking in healthy individuals. Journal of Applied Physiology 104, 747–755. 10.1152/japplphysiol.01380.2006 [DOI] [PubMed] [Google Scholar]
  19. Liu CJ, Latham NK, 2009. Progressive resistance strength training for improving physical function in older adults. Cochrane Database of Systematic Reviews. 10.1002/14651858.CD002759.pub2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Malatesta D, Canepa M, Menendez Fernandez A, 2017. The effect of treadmill and overground walking on preferred walking speed and gait kinematics in healthy, physically active older adults. European Journal of Applied Physiology 117, 1833–1843. 10.1007/s00421-017-3672-3 [DOI] [PubMed] [Google Scholar]
  21. Mannering N, Young T, Spelman T, Choong PF, 2017. Three-dimensional knee kinematic analysis during treadmill gait: Slow imposed speed versus normal self-selected speed. Bone and Joint Research 6, 514–521. 10.1302/2046-3758.68.BJR-2016-0296.R1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Minetti AE, Boldrini L, Brusamolin L, Zamparo P, McKee T, 2003. A feedback-controlled treadmill (treadmill-on-demand) and the spontaneous speed of walking and running in humans. Journal of Applied Physiology 95, 838–843. 10.1152/japplphysiol.00128.2003 [DOI] [PubMed] [Google Scholar]
  23. Nagano H, Begg RK, Sparrow WA, Taylor S, 2013. A comparison of treadmill and overground walking effects on step cycle asymemtry in young and older individuals. Journal of Applied Biomechanics 29, 188–193. 10.1123/jab.29.2.188 [DOI] [PubMed] [Google Scholar]
  24. Paulson S, Gray M, 2015. Parameters of gait among community-dwelling older adults. Journal of Geriatric Physical Therapy 38, 28–32. 10.1519/JPT.0000000000000018 [DOI] [PubMed] [Google Scholar]
  25. Ray NT, Knarr BA, Higginson JS, 2018. Walking speed changes in response to novel user-driven treadmill control. Journal of Biomechanics 78, 143–149. 10.1016/j.jbiomech.2018.07.035 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Ray NT, Reisman DS, Higginson JS, 2020. Walking speed changes in response to user-driven treadmill control after stroke. Journal of Biomechanics 101, 109643 10.1016/j.jbiomech.2020.109643 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Shimada H, Obuchi S, Furuna T, Suzuki T, 2004. New intervention program for preventing falls among frail elderly people: The effects of perturbed walking exercise using a bilateral separated treadmill. American Journal of Physical Medicine and Rehabilitation 83, 493–499. 10.1097/01.PHM.0000130025.54168.91 [DOI] [PubMed] [Google Scholar]
  28. Studenski S, Perera S, Patel K, Rosano C, Faulkner K, Inzitari M, Brach J, Chandler J, Cawthon P, Connor EB, Nevitt M, Visser M, Kritchevsky S, Badinelli S, Harris T, Newman AB, Cauley J, Ferrucci L, Guralnik J, 2011. Gait speed and survival in older adults. Journal of the American Medical Association 305, 50–58. 10.1001/jama.2010.1923 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Terrier P, Deriaz O, 2011. Kinematic variability, fractal dynamics and local dynamic stability of treadmill walking. Journal of NeuroEngineering & Rehabilitation 8 10.1186/1743-0003-8-12 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Wass E, Taylor NF, Matsas A, 2005. Familiarisation to treadmill walking in unimpaired older people. Gait and Posture 21, 72–79. 10.1016/j.gaitpost.2004.01.003 [DOI] [PubMed] [Google Scholar]
  31. Yoon J, Park HS, Damiano DL, 2012. A novel walking speed estimation scheme and its application to treadmill control for gait rehabilitation. Journal of NeuroEngineering and Rehabilitation 9 10.1186/1743-0003-9-62 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Zeni JA Jr., Higginson JS, 2009. Differences in gait parameters between healthy subjects and persons with moderate and severe knee osteoarthritis: A result of altered walking speed? Clinical Biomechanics 24, 372–378. 10.1016/j.clinbiomech.2009.02.001. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES