Gait & Posture 39 (2014) 1115–1121
Contents lists available at ScienceDirect
Gait & Posture journal homepage: www.elsevier.com/locate/gaitpost
The influence of ankle muscle activation on postural sway during quiet stance Meagan J. Warnica a,b, Tyler B. Weaver a,b, Stephen D. Prentice b, Andrew C. Laing a,b,* a b
Injury Biomechanics and Aging Laboratory, University of Waterloo, Waterloo, Ontario, Canada Department of Kinesiology, University of Waterloo, Waterloo, Ontario, Canada
A R T I C L E I N F O
A B S T R A C T
Article history: Received 9 May 2013 Received in revised form 21 January 2014 Accepted 24 January 2014
Although balance during quiet standing is postulated to be influenced by multiple factors, including ankle stiffness, it is unclear how different mechanisms underlying increases in stiffness affect balance control. Accordingly, this study examined the influence of muscle activation and passive ankle stiffness increases on the magnitude and frequency of postural sway. Sixteen young adults participated in six quiet stance conditions including: relaxed standing, four muscle active conditions (10%, 20%, 30% and 40% maximum voluntary contraction (MVC)), and one passive condition wearing an ankle foot orthotic (AFO). Kinetics were collected from a force plate, while whole-body kinematics were collected with a 12sensor motion capture system. Bilateral electromyographic signals were recorded from the tibialis anterior and medial gastrocnemius muscles. Quiet stance sway amplitude (range and root mean square) and frequency (mean frequency and velocity) in the sagittal plane were calculated from time-varying centre of gravity (COG) and centre of pressure (COP) data. Compared to the relaxed standing condition, metrics of sway amplitude were significantly increased (between 37.5 and 63.2%) at muscle activation levels of 30% and 40% MVC. Similarly, frequency measures increased between 30.5 and 154.2% in the 20– 40% MVC conditions. In contrast, passive ankle stiffness, induced through the AFO, significantly decreased sway amplitude (by 23–26%), decreased COG velocity by 13.8%, and increased mean COP frequency by 24.9%. These results demonstrate that active co-contraction of ankle musculature (common in Parkinson’s Disease patients) may have differential effects on quiet stance balance control compared to the use of an ankle foot orthotic (common for those recovering from stroke). Crown Copyright ß 2014 Published by Elsevier B.V. All rights reserved.
Keywords: Ankle stiffness Ankle-foot orthotic Centre of pressure Centre of gravity Inverted pendulum model
1. Introduction Balance during quiet stance is controlled by vestibular, visual, proprioceptive, tactile and muscular factors [1]. While difficult to measure all aspects of balance, previous literature has suggested that ankle stiffness plays a significant contribution in the ability to maintain balance during quiet stance [1–3]. There are several distinct contributors to ankle stiffness. Passive ankle stiffness refers to a resistance to postural deviation caused by contributions from non-contractile elements including passive tissues (e.g. ligaments, non-activated muscles) or external devices (e.g. an ankle–foot orthotic) that span the ankle joint. In contrast, active ankle stiffness is generated through increased muscle activation, specifically the co-contraction of antagonistic muscles
* Corresponding author at: Department of Kinesiology, Faculty of Applied Health Sciences, University of Waterloo, 200 University Avenue West Waterloo, Ontario N2L 3G1, Canada. Tel.: +1 519 888 4567x38947. E-mail address: [email protected] (A.C. Laing). http://dx.doi.org/10.1016/j.gaitpost.2014.01.019 0966-6362/Crown Copyright ß 2014 Published by Elsevier B.V. All rights reserved.
crossing the ankle joint. There has been support for the contribution of increased ankle muscle activation on postural sway variables during quiet stance, especially in the presence of consistent passive properties [4]. One approach to characterizing ankle stiffness utilizes timevarying underfoot centre of pressure (COP) and whole-body centre of gravity (COG) data during postural sway. As stiffness determines the acceleration of the COG, and based on the inverted-pendulum model (where the body is assumed to be rigid above a single articulation at the ankle), COP–COG is proportional to the acceleration of the COG, stiffness can be estimated through the analysis of the amplitude spectrum of the COP–COG signal [3]. As such, Winter, Patla, Prince, Ishac and Gielo-Perczak [3] demonstrated that, during relaxed quiet stance conditions, anterior–posterior (AP) sway amplitude (i.e., COG displacement) is inversely proportional to the estimated stiffness about the ankle, while the frequency of postural sway increases with stiffness [3]. This suggests that stiffness at the ankle is preset by the central nervous system to control the body’s COG during quiet stance and assumes that the muscles behave as springs to keep the COP and
1116
M.J. Warnica et al. / Gait & Posture 39 (2014) 1115–1121
COG in phase as the body sways. The inverted-pendulum model proposes that increases in ankle joint stiffness (achieved through either passive or active approaches) should result in decreased sway amplitude and increased sway frequency. These relationships are governed by the equations,
vn ¼
0:5 Ke ðV 0 ðIsa Þ0:5 Þ and X ¼ Isa Ke0:5
where the frequency of oscillation (vn) is a function of the effective stiffness (Ke) and moment of inertia (Isa), while the amplitude of oscillation (X) is a function of the horizontal velocity of the COG (V0), Isa and Ke [3]. However, despite these theoretical relationships, it has never been demonstrated whether increases in stiffness (either active or passive in nature) influence parameters of sway or postural control strategies during the task of quiet stance. Alternative explanations suggest that active anticipatory control strategies must act in conjunction with pre-set passive stiffness characteristics to maintain balance during quiet stance [1,5–8]. The effects of rocking frequency on the COP and COG dynamics have also been studied, with increasing voluntary rocking frequency causing peak COG displacement to decrease with no concomitant effect on peak COP displacement [9]. However, despite the wealth of literature on quiet stance, there is still an overall lack of understanding of how both increased ankle muscle activation and passive ankle stiffness influence neuromechanical balance control dynamics during quiet stance. Additionally, it is unclear whether changes in co-contraction of the ankle musculature would alter system characteristics sufficiently to promote a shift in balance control away from the ankle strategy which typically dominates during quiet stance. There is potential clinical value in characterizing the effects of increased ankle muscle activation and passive stiffness on balance control parameters for any pathology which alters ankle stiffness. For example, individuals with Parkinson’s Disease (PD) have been reported to exhibit increased active ankle stiffness compared to healthy controls as a consequence of co-contraction [10,11]. In addition, it is potentially important to examine how postural sway could be influenced by mechanically-induced increases in passive ankle stiffness which can be created by an external device such as an ankle foot orthotic (AFO). Clinical populations such as those with cerebral palsy, muscular dystrophy, multiple sclerosis and stroke survivors commonly use such devices to decrease passive ‘toe-drop’ during activities of daily living (ADLs) including locomotion. Many ADLs involve standing for extended periods (e.g., standing at the sink to wash dishes, standing at a bus stop). Unfortunately, there is currently a lack of information on how devices such as AFOs influence postural sway during quiet standing, and how this compares to active mechanisms of increased ankle stiffness. The present study examined the influence of ankle muscle activation and passive ankle stiffness on the magnitude and frequency of postural sway variables in the sagittal plane. The study’s primary goal was to test the hypotheses that, during quiet stance, increases in ankle muscle activation and passive stiffness (through the use of an AFO) would result in: (1) decreased amplitude (i.e., displacement range and root mean square) and (2) increased frequency (i.e., mean velocity and frequency) of COP and COG sway in the sagittal plane. A secondary goal of the study was to explore potential mechanisms underlying any changes observed in COP or COG sway parameters. 2. Methods 2.1. Participants Sixteen healthy young adults (eight males, eight females) participated in this study (mean (SD)) age, height, and body mass
were 22.6(1.4) years, 173.3(11.1) cm, and 70.7(12.9) kg, respectively). All participants were free of any injuries or health issues with the potential to influence balance control. Informed consent was obtained from all participants, and the study protocol was approved by the Office of Research Ethics at the University of Waterloo. 2.2. Experimental protocol Testing sessions began with instrumentation setup and calibration. The first step involved bilateral preparation (i.e. shaving and cleaning) and application of electrodes (Noraxon electrodes, Noraxon, Scottsdale, AZ, USA) on the skin overlying the tibialis anterior (TA) and medial gastrocnemius (MG). Throughout the session, electromyography signals were collected from these sites (Bortec AMT-8, Bortec, Calgary, AB, Canada). Three maximal voluntary contraction (MVC) trials were recorded from the TA and MG muscles. The MVCs trials were 5 s in duration, and were obtained via standing isometric toe and heel lifts against resistance provided by straps across the toes and over the shoulders, respectively. EMG data from these trials were processed by subtracting the mean from each trial, full wave rectifying, and linear enveloping using a 2nd order Butterworth filter with a 4 Hz cutoff frequency [12]. MVC values for each muscle were defined as the maximum value observed during the middle three seconds across all MVC trials. Experimental trials (described below) were similarly processed, and then normalized by participant and muscle-specific MVC values. Nineteen active infrared markers (Smartmarkers, Northern Digital Inc., Waterloo, Ontario, Canada) were placed on each participant to create an 11-segment model. Specific marker placements included the skin over the posterior aspect of the seventh cervical vertebra, and bilaterally on the big toe, fifth metatarsal, lateral malleolus, calcaneus, femoral condyle, greater trochanter, acromion, lateral epicondyle, and radial styloid. Throughout the session, kinematic data were collected using a 12-sensor motion capture system (Optotrak Certus, Northern Digital Inc., Waterloo, Ontario, Canada). For the experimental session, each participant stood quietly on a force plate (OR6–3, AMTI, Watertown, MA, USA) with their arms crossed and hands lightly touching the opposite shoulders facing a computer monitor which displayed real-time feedback of the linearenveloped activation level of the right MG muscle (processed as described above). Although only a single muscle was used as feedback for the participants to achieve the desired activation levels (to keep the task as simple as possible), participants were not told that the signal was derived from a single site. Rather, they were instructed to ‘‘contract all the muscles around their ankles to reach the desired range of muscle activity as shown by the feedback on the computer monitor’’. Participants were provided with a five minute familiarization period to practice reaching the muscle activation levels. Finally, experimental trials involving quiet stance were conducted. All data sources were synchronized using NDI First Principles (Northern Digital Inc., Waterloo, Ontario, Canada). The six experimental conditions included: a relaxed (baseline) condition, four muscle activation conditions (10%, 20%, 30% and 40% of MVC) in which the participants matched their muscle activity to the appropriate target level (5%) on the computer monitor, and one passive condition wearing a common AFO (CA508-LL Ankle Foot Orthosis, Sammons Preston Canada Inc, Toronto, Ontario, Canada). Four trials were completed for each condition. The active conditions were chosen to span and exceed activation levels observed in clinical populations including PD and CP [13], while still being feasible for participants to maintain for the duration of the 30 s trials. Although previous research has recommended a sampling duration of at least 60 s to capture both the low and high frequency content of the COP signal [14], the current study employed trials
M.J. Warnica et al. / Gait & Posture 39 (2014) 1115–1121
lasting only 30 s in an effort to minimize fatigue due to participants performing multiple contractions of the ankle musculature. Additionally, sitting breaks were given to the participant after every five trials. The relaxed and muscle active conditions were randomized for each participant. The AFO condition was always completed at the end of the test session. During all trials, kinetic and EMG data were collected at 2160 Hz, and kinematic data were collected at 60 Hz. 2.3. Data analysis All data analysis was completed with custom software (MatLab vR2007b, Mathworks, Natick, MA). Gaps in kinematic marker data (
Contents lists available at ScienceDirect
Gait & Posture journal homepage: www.elsevier.com/locate/gaitpost
The influence of ankle muscle activation on postural sway during quiet stance Meagan J. Warnica a,b, Tyler B. Weaver a,b, Stephen D. Prentice b, Andrew C. Laing a,b,* a b
Injury Biomechanics and Aging Laboratory, University of Waterloo, Waterloo, Ontario, Canada Department of Kinesiology, University of Waterloo, Waterloo, Ontario, Canada
A R T I C L E I N F O
A B S T R A C T
Article history: Received 9 May 2013 Received in revised form 21 January 2014 Accepted 24 January 2014
Although balance during quiet standing is postulated to be influenced by multiple factors, including ankle stiffness, it is unclear how different mechanisms underlying increases in stiffness affect balance control. Accordingly, this study examined the influence of muscle activation and passive ankle stiffness increases on the magnitude and frequency of postural sway. Sixteen young adults participated in six quiet stance conditions including: relaxed standing, four muscle active conditions (10%, 20%, 30% and 40% maximum voluntary contraction (MVC)), and one passive condition wearing an ankle foot orthotic (AFO). Kinetics were collected from a force plate, while whole-body kinematics were collected with a 12sensor motion capture system. Bilateral electromyographic signals were recorded from the tibialis anterior and medial gastrocnemius muscles. Quiet stance sway amplitude (range and root mean square) and frequency (mean frequency and velocity) in the sagittal plane were calculated from time-varying centre of gravity (COG) and centre of pressure (COP) data. Compared to the relaxed standing condition, metrics of sway amplitude were significantly increased (between 37.5 and 63.2%) at muscle activation levels of 30% and 40% MVC. Similarly, frequency measures increased between 30.5 and 154.2% in the 20– 40% MVC conditions. In contrast, passive ankle stiffness, induced through the AFO, significantly decreased sway amplitude (by 23–26%), decreased COG velocity by 13.8%, and increased mean COP frequency by 24.9%. These results demonstrate that active co-contraction of ankle musculature (common in Parkinson’s Disease patients) may have differential effects on quiet stance balance control compared to the use of an ankle foot orthotic (common for those recovering from stroke). Crown Copyright ß 2014 Published by Elsevier B.V. All rights reserved.
Keywords: Ankle stiffness Ankle-foot orthotic Centre of pressure Centre of gravity Inverted pendulum model
1. Introduction Balance during quiet stance is controlled by vestibular, visual, proprioceptive, tactile and muscular factors [1]. While difficult to measure all aspects of balance, previous literature has suggested that ankle stiffness plays a significant contribution in the ability to maintain balance during quiet stance [1–3]. There are several distinct contributors to ankle stiffness. Passive ankle stiffness refers to a resistance to postural deviation caused by contributions from non-contractile elements including passive tissues (e.g. ligaments, non-activated muscles) or external devices (e.g. an ankle–foot orthotic) that span the ankle joint. In contrast, active ankle stiffness is generated through increased muscle activation, specifically the co-contraction of antagonistic muscles
* Corresponding author at: Department of Kinesiology, Faculty of Applied Health Sciences, University of Waterloo, 200 University Avenue West Waterloo, Ontario N2L 3G1, Canada. Tel.: +1 519 888 4567x38947. E-mail address: [email protected] (A.C. Laing). http://dx.doi.org/10.1016/j.gaitpost.2014.01.019 0966-6362/Crown Copyright ß 2014 Published by Elsevier B.V. All rights reserved.
crossing the ankle joint. There has been support for the contribution of increased ankle muscle activation on postural sway variables during quiet stance, especially in the presence of consistent passive properties [4]. One approach to characterizing ankle stiffness utilizes timevarying underfoot centre of pressure (COP) and whole-body centre of gravity (COG) data during postural sway. As stiffness determines the acceleration of the COG, and based on the inverted-pendulum model (where the body is assumed to be rigid above a single articulation at the ankle), COP–COG is proportional to the acceleration of the COG, stiffness can be estimated through the analysis of the amplitude spectrum of the COP–COG signal [3]. As such, Winter, Patla, Prince, Ishac and Gielo-Perczak [3] demonstrated that, during relaxed quiet stance conditions, anterior–posterior (AP) sway amplitude (i.e., COG displacement) is inversely proportional to the estimated stiffness about the ankle, while the frequency of postural sway increases with stiffness [3]. This suggests that stiffness at the ankle is preset by the central nervous system to control the body’s COG during quiet stance and assumes that the muscles behave as springs to keep the COP and
1116
M.J. Warnica et al. / Gait & Posture 39 (2014) 1115–1121
COG in phase as the body sways. The inverted-pendulum model proposes that increases in ankle joint stiffness (achieved through either passive or active approaches) should result in decreased sway amplitude and increased sway frequency. These relationships are governed by the equations,
vn ¼
0:5 Ke ðV 0 ðIsa Þ0:5 Þ and X ¼ Isa Ke0:5
where the frequency of oscillation (vn) is a function of the effective stiffness (Ke) and moment of inertia (Isa), while the amplitude of oscillation (X) is a function of the horizontal velocity of the COG (V0), Isa and Ke [3]. However, despite these theoretical relationships, it has never been demonstrated whether increases in stiffness (either active or passive in nature) influence parameters of sway or postural control strategies during the task of quiet stance. Alternative explanations suggest that active anticipatory control strategies must act in conjunction with pre-set passive stiffness characteristics to maintain balance during quiet stance [1,5–8]. The effects of rocking frequency on the COP and COG dynamics have also been studied, with increasing voluntary rocking frequency causing peak COG displacement to decrease with no concomitant effect on peak COP displacement [9]. However, despite the wealth of literature on quiet stance, there is still an overall lack of understanding of how both increased ankle muscle activation and passive ankle stiffness influence neuromechanical balance control dynamics during quiet stance. Additionally, it is unclear whether changes in co-contraction of the ankle musculature would alter system characteristics sufficiently to promote a shift in balance control away from the ankle strategy which typically dominates during quiet stance. There is potential clinical value in characterizing the effects of increased ankle muscle activation and passive stiffness on balance control parameters for any pathology which alters ankle stiffness. For example, individuals with Parkinson’s Disease (PD) have been reported to exhibit increased active ankle stiffness compared to healthy controls as a consequence of co-contraction [10,11]. In addition, it is potentially important to examine how postural sway could be influenced by mechanically-induced increases in passive ankle stiffness which can be created by an external device such as an ankle foot orthotic (AFO). Clinical populations such as those with cerebral palsy, muscular dystrophy, multiple sclerosis and stroke survivors commonly use such devices to decrease passive ‘toe-drop’ during activities of daily living (ADLs) including locomotion. Many ADLs involve standing for extended periods (e.g., standing at the sink to wash dishes, standing at a bus stop). Unfortunately, there is currently a lack of information on how devices such as AFOs influence postural sway during quiet standing, and how this compares to active mechanisms of increased ankle stiffness. The present study examined the influence of ankle muscle activation and passive ankle stiffness on the magnitude and frequency of postural sway variables in the sagittal plane. The study’s primary goal was to test the hypotheses that, during quiet stance, increases in ankle muscle activation and passive stiffness (through the use of an AFO) would result in: (1) decreased amplitude (i.e., displacement range and root mean square) and (2) increased frequency (i.e., mean velocity and frequency) of COP and COG sway in the sagittal plane. A secondary goal of the study was to explore potential mechanisms underlying any changes observed in COP or COG sway parameters. 2. Methods 2.1. Participants Sixteen healthy young adults (eight males, eight females) participated in this study (mean (SD)) age, height, and body mass
were 22.6(1.4) years, 173.3(11.1) cm, and 70.7(12.9) kg, respectively). All participants were free of any injuries or health issues with the potential to influence balance control. Informed consent was obtained from all participants, and the study protocol was approved by the Office of Research Ethics at the University of Waterloo. 2.2. Experimental protocol Testing sessions began with instrumentation setup and calibration. The first step involved bilateral preparation (i.e. shaving and cleaning) and application of electrodes (Noraxon electrodes, Noraxon, Scottsdale, AZ, USA) on the skin overlying the tibialis anterior (TA) and medial gastrocnemius (MG). Throughout the session, electromyography signals were collected from these sites (Bortec AMT-8, Bortec, Calgary, AB, Canada). Three maximal voluntary contraction (MVC) trials were recorded from the TA and MG muscles. The MVCs trials were 5 s in duration, and were obtained via standing isometric toe and heel lifts against resistance provided by straps across the toes and over the shoulders, respectively. EMG data from these trials were processed by subtracting the mean from each trial, full wave rectifying, and linear enveloping using a 2nd order Butterworth filter with a 4 Hz cutoff frequency [12]. MVC values for each muscle were defined as the maximum value observed during the middle three seconds across all MVC trials. Experimental trials (described below) were similarly processed, and then normalized by participant and muscle-specific MVC values. Nineteen active infrared markers (Smartmarkers, Northern Digital Inc., Waterloo, Ontario, Canada) were placed on each participant to create an 11-segment model. Specific marker placements included the skin over the posterior aspect of the seventh cervical vertebra, and bilaterally on the big toe, fifth metatarsal, lateral malleolus, calcaneus, femoral condyle, greater trochanter, acromion, lateral epicondyle, and radial styloid. Throughout the session, kinematic data were collected using a 12-sensor motion capture system (Optotrak Certus, Northern Digital Inc., Waterloo, Ontario, Canada). For the experimental session, each participant stood quietly on a force plate (OR6–3, AMTI, Watertown, MA, USA) with their arms crossed and hands lightly touching the opposite shoulders facing a computer monitor which displayed real-time feedback of the linearenveloped activation level of the right MG muscle (processed as described above). Although only a single muscle was used as feedback for the participants to achieve the desired activation levels (to keep the task as simple as possible), participants were not told that the signal was derived from a single site. Rather, they were instructed to ‘‘contract all the muscles around their ankles to reach the desired range of muscle activity as shown by the feedback on the computer monitor’’. Participants were provided with a five minute familiarization period to practice reaching the muscle activation levels. Finally, experimental trials involving quiet stance were conducted. All data sources were synchronized using NDI First Principles (Northern Digital Inc., Waterloo, Ontario, Canada). The six experimental conditions included: a relaxed (baseline) condition, four muscle activation conditions (10%, 20%, 30% and 40% of MVC) in which the participants matched their muscle activity to the appropriate target level (5%) on the computer monitor, and one passive condition wearing a common AFO (CA508-LL Ankle Foot Orthosis, Sammons Preston Canada Inc, Toronto, Ontario, Canada). Four trials were completed for each condition. The active conditions were chosen to span and exceed activation levels observed in clinical populations including PD and CP [13], while still being feasible for participants to maintain for the duration of the 30 s trials. Although previous research has recommended a sampling duration of at least 60 s to capture both the low and high frequency content of the COP signal [14], the current study employed trials
M.J. Warnica et al. / Gait & Posture 39 (2014) 1115–1121
lasting only 30 s in an effort to minimize fatigue due to participants performing multiple contractions of the ankle musculature. Additionally, sitting breaks were given to the participant after every five trials. The relaxed and muscle active conditions were randomized for each participant. The AFO condition was always completed at the end of the test session. During all trials, kinetic and EMG data were collected at 2160 Hz, and kinematic data were collected at 60 Hz. 2.3. Data analysis All data analysis was completed with custom software (MatLab vR2007b, Mathworks, Natick, MA). Gaps in kinematic marker data (
