THE ANATOMICAL RECORD 298:346–353 (2015)

Age-Related Difference in Postural Control During Recovery from Posterior and Anterior Perturbations MADELINE LOUISE SINGER, LORINDA K. SMITH, LELAND E. DIBBLE, AND K. BO FOREMAN* Department of Physical Therapy, University of Utah, 520 Wakara Way, Salt Lake City, Utah

ABSTRACT Decreased reactive postural responses in elderly adults may place them at increased risk for falls and related injuries. The first step in addressing the high rate of falls in the elderly population is to determine a baseline for postural response in healthy young and healthy elderly individuals. To determine these age-related differences in reactive postural responses during recovery from posterior and anterior perturbations, we used the tether-release method in conjunction with a motion analysis system to evaluate overall movement latencies, overall movement amplitude and velocity, and joint-specific amplitude and velocity in healthy young (n 5 10, mean age525 6 5) and healthy elderly participants (n 5 10, mean age 5 67 6 6). During posterior perturbations, healthy elderly participants had increased recovery time (P 5 0.01) and ratio of center of mass to step length (P 5 0.013) when compared with young participants. Elderly participants also had decreased step length (P 5 0.006), peak COM velocity (P 5 0.01), peak knee flexion angle (P 5 0.002), and decreased hip (P 5 0.005) and knee (P 5 0.0005) average angular velocity when compared with young participants. We conclude that these movement deficiencies at the hip and knee limited the length of the recovery step. With this restricted step, elderly participants could not achieve adequate mechanical advantage to counteract the displacement of their COM using a single step. During anterior perturbations, elderly participants did not exhibit any significant differences compared to young participants in overall movement variables. This understanding of postural responses in healthy individuals is clinically relevant to the development of rehabilitation programs for individuals at high fall risk. C 2014 Wiley Periodicals, Inc. Anat Rec, 298:346–353, 2015. V

Key words: Postural control; falls; elderly; motion capture; biomechanics

INTRODUCTION Elderly individuals are at an increased risk of falling and experiencing injury from falls. Previous research reports that 30%–40% of adults over the age of 65 fall each year (Campbell et al., 1981; Blake et al., 1988) and these falls pose a large health and financial problem for our society (Schiller et al., 2007; Białoszewski et al., 2008; Czerwi nski et al., 2008). In the elderly population, falls can result in injury, hospitalization, and even death C 2014 WILEY PERIODICALS, INC. V

This article includes AR WOW Videos, which can be viewed at http://bcove.me/smqcj96j; http://bcove.me/4ofw2wm3 and http://bcove.me/uwp6l5bd Grant sponsor: College of Health, University of Utah. *Correspondence to: K. Bo Foreman, PT, PhD, 520 Wakara Way, Salt Lake City, UT 84108. E-mail: [email protected] Received 10 October 2013; Accepted 2 April 2014. DOI 10.1002/ar.23043 Published online 30 August 2014 in Wiley Online Library (wileyonlinelibrary.com).

AGE-RELATED POSTURAL CONTROL DURING FALLS

(Tinetti et al., 1988; Campbell et al., 1990). For elderly individuals, the injury rates due to falls are nearly three times that of younger individuals (Schiller et al., 2007). In addition to hospitalization and injury, falls can reduce autonomy, mobility, and the overall quality of life for the falling individual (Yoshida, 2007; Białoszewski et al., 2008; Czerwi nski et al., 2008). Falls in the elderly population also place a large financial burden on the health care industry. The cost of fall related injuries is expected to increase from $27.3 billion in 1994 to $43.8 billion by 2020 (Englander et al., 1996). To further exacerbate the burden of falls on our society, the high fall risk population (adults over 65 years) is expected to increase from 35 million in 2000 to 55 million in 2020 (Administartion on Aging, 2009). Previous research has demonstrated a marked agerelated decline in postural control that affects the ability of elderly individuals to successfully recover from a fall (Overstall et al., 1977; Horak et al., 1989; Gu et al., 1996; Hall and Jensen, 2002). In a dynamic situation, such as a fall, the Center of Mass (COM) is displaced from the Center of Pressure (COP), causing a loss of balance. If the displacement is large enough, a compensatory stepping response is required to regain control (Mathiyakom and McNitt-Gray, 2008) of the COM. During this response, coordination between neural and musculoskeletal systems must occur quickly and efficiently to appropriately position the stepping limb (Hall and Jensen, 2002; Mathiyakom and McNitt-Gray, 2008). Many methods have been used to evaluate the stepping response to determine the risk factors contributing to the age-related decline in postural control. The clinical methods used to evaluate fall risk and postural instability include the Pull Test (Hunt et al., 2006) and the Push and Release Test (Valkovicˇ et al., 2008). Both tests use examiner produced perturbations that require the patient to take a compensatory step to regain balance. These tests are typically used in a clinical setting to evaluate symptoms of degenerative neurological disorders such as Parkinson disease and multiple sclerosis (Visser et al., 2003; Jankovic, 2008). However, these clinical measures are not ideal in a research setting where a repeatable method of creating perturbations must be used consistently across participants. Several experimental methods have successfully produced consistent and repeatable perturbations in the laboratory. The two most prominent methods are the sliding force plate method and the tether-release method. The sliding force plate method creates a sudden translational movement of the base of support (BOS) to initiate a loss of balance. However this method is expensive and lacks ecological validity. The tether-release method (Do et al., 1982), on the other hand, is a relatively inexpensive method that is used to simulate the body position that typically occurs during a fall. Thus, this method has been frequently used to evaluate an individual’s recovery response after a simulated slip (posterior perturbation) (Hsiao and Robinovitch, 2001) or a trip (anterior perturbation) (Thelen et al., 1997, 2000; Wojcik et al., 1999, 2001; Madigan and Lloyd, 2005a, 2005b). The majority of these tether-release studies have focused on determining the age-related biomechanical differences between elderly and young individuals during recovery from an anterior perturbation (Thelen et al., 1997, 2000; Wojcik et al., 1999, 2001; Madigan

347

and Lloyd, 2005a). These studies have concluded that younger individuals can successfully recover from greater forward lean angles than elderly individuals (Thelen et al., 1997; Wojcik et al., 1999). The literature has also presented a variety of spatiotemporal (Thelen et al., 1997; Wojcik et al., 1999), kinematic (Madigan and Lloyd, 2005a), and kinetic (Wojcik et al., 2001; Madigan and Lloyd, 2005b) factors contributing to this conclusion. However, the discrepancy amongst studies is vast, and the specific biomechanical characteristics that contribute to the decreased postural control and increased fall risk in the elderly are debated (HsiaoWecksler, 2008). Furthermore, few studies have examined the recovery response of individuals during a posterior perturbation (Hsiao-Wecksler, 2008), and no published studies, to our knowledge, have used the tether-release method to compare the biomechanical characteristics of healthy elderly to healthy young individuals during recovery from a posterior perturbation. Due to this paucity of research concerning backward falls, little is known about the mechanics and neuromuscular control that occurs during recovery from a posterior perturbation (Mathiyakom and McNitt-Gray, 2008). This knowledge is essential because posterior falls can result in hip fracture (Kurz et al., 2013), head trauma (Jager et al., 2000), and other serious injuries resulting in hospitalization and even death (Yoshida, 2007). Additionally, individuals are more likely to fall from a posterior perturbation than from an anterior perturbation of the same magnitude (Hsiao and Robinovitch, 1998). This increased likelihood of a posterior fall is primarily due to the reduced postural control in the posterior direction created by the decreased ability to control the posterior movement of the COM (Winter et al., 1996). Therefore, the purpose of this study was to determine the age-related biomechanical differences in reactive postural responses during recovery from posterior and anterior perturbations. Based on the high incidence of falls in the elderly population, we hypothesized that (1) healthy elderly participants would exhibit increased overall movement latencies during recovery from posterior and anterior perturbations when compared with healthy young participants, (2) healthy elderly participants would exhibit decreased overall movement amplitude and velocity during recovery from posterior and anterior perturbations when compared with healthy young participants, and (3) healthy elderly participants would exhibit decreased joint-specific movement amplitude and velocity during recovery from posterior and anterior perturbations when compared with healthy young participants. Results from testing the aforementioned hypotheses will provide a baseline understanding for perturbation recovery in healthy young and healthy elderly individuals.

METHODS Participant Selection Potential participants were a sample of convenience recruited from the University community. The inclusion criteria for the healthy young group included individuals between the ages of 20–40 years. For the healthy elderly group, the inclusion criteria included individuals between the ages of 60–80 years. The exclusion criterion

348

SINGER ET AL.

for both groups was a history of neurological disease or medical conditions (orthopedic, cardiovascular, or otherwise) that would limit the participant’s ability to recover from a fall normally.

Measures To consistently and repeatedly evaluate the perturbation recovery response in elderly and young participants, we modified a method from the literature called the tether-release method (Do et al., 1982). The equipment used in this method includes an electromagnet affixed to the wall with a tether extending from it to a chest harness worn by the participant. Also attached to the chest harness is a fall restraint tether that extends from the ceiling for safety in the event that the participant is unable to recover. Additionally, there is a small force sensor in-line with the tether and attached to the chest harness to measure lean forces. The method requires the participant to lean against the tether, displacing the COM from the COP, which occurs in a slip or a trip (Mathiyakom and McNitt-Gray, 2008). Once the threshold is reached, the electromagnet is remotely turned off and the tether releases, resulting in a perturbation (Video 1).

Procedures Prior to testing, Institutional Review Board (IRB) approval for the study was obtained. Following recruitment, the purposes and procedures were outlined and participants were provided with an opportunity to ask questions prior to signing the IRB consent form. All testing was completed in a single session at the Motion Capture Facility at the University of Utah Department of Physical Therapy. Data were captured using a 10-camera motion analysis system at 200 Hz (Vicon Motion Systems; Oxford, UK) and two force platforms at 1000 Hz (AMTI; Watertown, Mass, USA). Prior to the participant’s entry into the laboratory, the system was calibrated. Demographics and anthropometrics were collected from the participant, and the participant was fitted with the chest harness and 63 reflective markers, allowing for tracking of 15 body segments. After preparation of the system and the participant, a detailed description of the procedure was provided. Once the participant gave verbal confirmation that they understood the test, the participant was positioned on the first force platform, and the tether and fall-restraint tether were attached to the front of the chest harness in preparation for a posterior perturbation. Tape was used to mark the starting position of the participant’s feet, and this starting position was used for all trials. Prior to initiation of the trials, the participant was instructed to try and recover their balance using a single step. The participant was then asked to lean against the tether until the small in-line force sensor registered a lean force equal to approximately 10% of their body weight (range5 9%– 11%) (Madigan and Lloyd, 2005a; Hsiao and Robinovitch, 2001). This was signaled by an audible cue emitted from the computer providing the participant with feedback. The tether was released at a randomized time between 10 and 30 sec after the threshold was reached (Video 1) (Hsiao and Robinovitch, 2001). After the release of the tether, the participant fell posteriorly and initiated the

Swing Phase of the recovery step, which involved the movement of the stepping limb towards the rear force plate for recovery. Foot Strike, as registered by the rear force plate, signified the end of the Swing Phase and the beginning of the Support Phase of the recovery step, which involved the production of forces through the lower limb to counteract the displacement and velocity of the COM. The Support Phase ended with the successful stop of the COM in a posterior direction (Video 2). Participants completed two practice trials before completing five additional trials. The same procedure was repeated for anterior perturbations in which the tether was attached to the back of the harness. The stepping response was characterized using seven dependent biomechanical variables segregated into three categories. The overall temporal and kinematic variables were selected to examine major differences in the stepping response from a whole-body perspective. The jointspecific variables were used to pinpoint joint-specific problems that may contribute to any major differences observed in stepping strategy. 1. Overall Movement Latencies a. Reaction Time: time (sec) from tether release to initial foot movement of the stepping limb. b. Step Time: time (sec) from initial foot movement of the stepping limb to initial contact of the stepping limb on the rear force plate. c. Recovery Time: time (sec) from tether release to when the COM stopped moving in the sagittal plane. 2. Overall Movement Amplitude and Velocity a. Step Length: The vector displacement (m) of the heel marker of the stepping limb from the position at tether release to the position at initial contact of the stepping limb on the rear force plate. Step length was normalized to height. b. Peak COM Velocity: The peak velocity of the COM in the sagittal plane from the time of initial movement of the stepping limb to foot strike (m/sec). c. Ratio of COM displacement to Step Length: COM displacement divided by step length (m/m). During the stepping response, the recovery step must have sufficient length to create a lever arm large enough to stop the COM displacement in a single step. If the COM travels beyond the length of the recovery step, then single step recovery is impossible regardless of the forces exerted by the stepping limb. The ratio of COM displacement to step length represents this relationship. A smaller ratio implies that the recovery step is more mechanically advantageous, while a ratio greater than 1 implies that recovery is unsuccessful. 3. Joint-Specific Amplitude and Velocity a. Peak Lower Extremity (LE) Angle: Maximum flexion/extension angle (deg) at the hip, knee, and ankle during the time from initial foot movement of the stepping limb to initial contact of the stepping limb on the rear force plate b. Average LE Angular Velocity: Average angular velocity of the hip, knee, and ankle flexion/extension angle during the time from initial foot movement of the stepping to initial contact of the stepping limb on the rear force plate (deg/sec). Angular velocity graph was rectified to account for both flexion and extension angles (Video 3).

349

AGE-RELATED POSTURAL CONTROL DURING FALLS

TABLE 1. Outcome Variables for Healthy Young and Healthy Elderly Participants During Posterior and Anterior Perturbation Recovery Posterior perturbation recovery

Temporal Step Time (s) Reaction Time (s) Recovery time (s) Spatial Step Length (m) Peak COM Velocity (m/s) COM Disp/Step Length Ratio Peak Joint Angle Hip Extension ( ) Knee Flexion ( ) Ankle Flexion ( ) Average Angular Velocity Hip Flexion/Extension ( /s) Knee Flexion/Extension ( /s) Ankle Flexion/Extension ( /s)

Anterior perturbation recovery

Healthy young mean 6 SD

Healthy elderly mean 6 SD

Healthy young mean 6 SD

Healthy elderly mean 6 SD

0.43 6 0.04 0.18 6 0.01 0.95 6 0.20

0.40 6 0.05 0.17 6 0.04 1.19 6 0.26a

0.44 6 0.07 0.14 6 0.05 0.92 6 0.24

0.49 6 0.10 0.18 6 0.10 1.05 6 0.23

0.54 6 0.08 0.75 6 0.08 0.25 6 0.03

0.39 6 0.14a 0.64 6 0.10a 0.32 6 0.08a

0.48 6 0.08 0.71 6 0.11 0.30 6 0.05

0.48 6 0.09 0.75 6 0.14 0.34 6 0.09

6.38 6 13.04 64.16 6 6.87 16.58 6 4.23

4.83 6 10.28 48.65 6 12.54b 13.07 6 4.54

46.20 6 11.16 57.51 6 7.10 17.05 6 2.88

40.91 6 12.25 47.25 6 11.01 15.38 6 4.85

157.49 6 33.70 275.23 6 42.22 72.50 6 22.79

110.94 6 34.85b 186.30 6 55.73b 60.73 6 19.82

145.45 6 23.71 299.36 6 45.24 74.49 6 14.44

118.37 6 26.05a 224.95 6 49.21a 69.43 6 20.56

a

Statistically significant result compared to healthy young participants (P  0.017). Statistically significant result compared to healthy young participants (P  0.025).

b

Data Analysis For each dependent variable, the first three successful trials were averaged. A successful trial was one in which all markers were visible, and the participant successfully recovered their balance with at least one step inside the boundaries of the rear force plate. Several participants were unable to recover from this magnitude of perturbation using a single step and thus required multiple steps for successful recovery. However, for these participants only the first step of their multi-step recovery was analyzed. Processing and extraction of the biomechanical variables was completed using Visual 3D (C-Motion, Germantown, MD, USA) (Videos 2 and 3). Descriptive statistics were completed for the demographic variables. Due to our small sample sizes and the potential of nonnormally distributed data, healthy young and healthy elderly groups were compared using separate Mann Whitney U-tests for each dependent variable. The experimental wide level of significance was set at P  0.05. However, to adjust for multiple comparisons, the level of significance for each group of variables was adjusted using a Bonferroni correction as follows: Overall Movement Latencies: P  0.017, Overall Movement Amplitude and Velocity: P  0.017, Joint-Specific Movement Amplitude and Velocity: P  0.025. All statistical analyses were performed using SPSS 19 (IBM Inc; Armonk, NY) for Macintosh.

RESULTS Recruitment Ten healthy young participants (mean age: 25 6 5), and 10 healthy elderly participants (mean age: 67 6 6) enrolled in this study. All participants exercised regularly and they reported that they did not have any balance deficits.

multiple steps to fully regain balance and stop the movement of the COM. During posterior perturbation recovery, 40% of the healthy elderly participants took multiple recovery steps during all 3 of their trials, and 30% had a mixed stepping response, signifying they took multiple recovery steps during one or two of their trials. Only 10% of the healthy young participants required multiple steps to recover from a posterior perturbation. For anterior perturbation recovery, the distribution of the stepping response was less varied between healthy elderly and healthy young participants. Of the healthy elderly participants, only 20% required more than one step to successfully recover during all three of their trials, and 10% had a mixed stepping response. The anterior perturbation recovery distribution for healthy young participants was the same as that for posterior perturbation recovery, with only 10% of healthy young participants taking multiple steps during recovery.

Overall Movement Latencies The variables measured in this study to examine the temporal components of the whole-body stepping response included reaction time, step time, and recovery time (Table 1). For posterior perturbation recovery, there were no significant differences in reaction time and step time between healthy elderly and healthy young participants (Fig. 1A). However, healthy elderly participants had a significantly longer recovery time during posterior perturbations when compared with healthy young participants (Fig. 1A, P 5 0.012). For anterior perturbation recovery, there were no significant differences in movement latencies between healthy elderly and healthy young participants.

Stepping Response Distribution During perturbation recovery, several participants were unable to successfully recover from the perturbation using a single step. Thus, several participants took

Overall Movement Amplitude and Velocity The variables measured to examine the movement amplitude and velocity of the whole-body stepping

350

SINGER ET AL.

Fig. 1. Posterior and Anterior Perturbation Recovery Outcome Variables. A: Overall movement latencies including reaction time, step time, and recovery time of healthy young participants (black) and healthy elderly participants (white) during posterior perturbation recovery. B: Overall movement amplitude and velocity of healthy young (black) and healthy elderly (white) participants during posterior perturbation recovery. (1) Length of the recovery step (m). (2) Peak Center of Mass (COM) velocity (m/s) during the swing phase of recovery. (3) Ratio of COM displacement to Step Length. An increase in this ratio suggests a less mechanically advantageous recovery strategy. C: Joint-specific

amplitude and velocity of healthy young participants (black) and healthy elderly participants (white) during posterior perturbation recovery. (1) Peak hip extension, knee flexion, and ankle flexion angle during swing phase of recovery (2) Average angular velocity achieved at the hip, knee, and ankle throughout the flexion/extension (FL/EX) motion of the swing phase. D: Average angular velocity at the hip, knee, and ankle throughout flexion/extension (FL/EX) of healthy young participants (black) and healthy elderly participants (white) during the swing phase of anterior perturbation recovery. An asterisk indicates statistically significant result compared to healthy young participants.

response included step length, peak COM velocity, and the ratio of COM displacement to step length (Table 1). For posterior perturbation recovery, healthy elderly participants had a significantly shortened step length when compared with healthy young participants (Fig. 1B(1), P 5 0.0025). Additionally, healthy elderly participants had a significantly decreased peak COM velocity when compared with healthy young participants (Fig. 1B(2), P 5 0.013). Healthy elderly participants also had a significantly increased ratio of COM displacement to step length (Fig. 1B(3), P 5 0.013). For anterior perturbation recovery, there were no significant differences in movement amplitude and velocity between healthy elderly and healthy young participants.

joint angular velocity at the hip, knee, and ankle during the swing phase of recovery (Video 3, Table 1). For posterior perturbation recovery, there were no significant differences in peak hip extension angle and ankle flexion angle during the swing phase of recovery between healthy elderly and healthy young participants (Fig. 1C(1)). However, healthy elderly participants had significantly decreased peak knee flexion angle in the swing phase when compared with healthy young participants (Fig. 1C(1), P 5 0.002). Additionally, no significant difference was found between groups for average angular velocity at the ankle (Fig. 1C(2)). However, healthy elderly participants had significantly decreased average angular velocity at the hip (P 5 0.005) and knee (P 5 0.0005) when compared with healthy young participants (Fig. 1C(2)). For anterior perturbation recovery, there was no significant difference in peak hip extension, knee flexion, or ankle flexion angle between healthy elderly and healthy young participants. Additionally, there was no

Joint-Specific Amplitude and Velocity The variables measured to examine the amplitudes and velocities of the lower extremity joints during the stepping response included peak joint angle and average

AGE-RELATED POSTURAL CONTROL DURING FALLS

significant difference in average angular velocity at the ankle between healthy elderly and healthy young participants (Fig. 1D). However, healthy elderly participants had a significantly decreased average angular velocity at the hip (P 5 0.0115) and knee (P 5 0.004) during the swing phase when compared with healthy young participants (Fig. 1D, Video 3).

DISCUSSION The purpose of this study was to determine the agerelated biomechanical differences in reactive postural responses during recovery from posterior and anterior perturbations. Specifically, we hypothesized that (1) healthy elderly participants would exhibit slower movement latencies, (2) decreased overall movement amplitude and velocity, and (3) decreased joint-specific movement amplitude and velocity during recovery from posterior and anterior perturbations when compared with healthy young participants. From the results of testing these hypotheses, we obtained a valuable understanding of the baseline recovery response exhibited by both healthy young and healthy elderly participants. For posterior perturbation recovery, these hypotheses are largely supported. Concerning movement latencies, the results show that both healthy elderly and healthy young participants reacted to the perturbation and completed the first recovery step in the same amount of time, but the healthy elderly participants required a longer amount of time to stop the posterior progression of the COM and fully recover their balance. Additionally, healthy elderly participants have several deficiencies in their overall movement amplitude, as exhibited by their shorter steps and increased ratio of COM displacement to step length. These results demonstrate that healthy elderly participants are not able to complete as long of a recovery step during the same time as healthy young participants and as a result are unable to control their COM efficiently. Furthermore, several differences in the movement amplitude and velocity of the lower extremity joints are evident in the healthy elderly perturbation. These results reveal that healthy elderly participants moved the hip and knee of their stepping limb more slowly and did not achieve as much knee flexion as the healthy young participants. These joint-specific differences restrict the length of the recovery step the healthy elderly participants can achieve in the given amount of time. With this shortened step, the lower limb is not in a position to counteract their COM displacement in the posterior direction. Thus, a second or even third step is needed to stop the COM progression and recover. For anterior perturbation recovery, the majority of our hypotheses are rejected because there are no significant differences in overall movement latencies, amplitudes, or velocities between healthy elderly and healthy young participants. However, there are several significant differences in joint-specific movement amplitudes and velocities between groups. Healthy elderly participants have less hip and knee average angular velocity during the swing phase of anterior perturbation recovery. Despite these joint-specific differences, the stepping response distribution for anterior perturbation recovery displays a relatively similar number of healthy elderly and healthy young participants recover in a single step. This similarity suggests that the differences at the hip

351

and knee are not substantial enough to affect the overall success of single step recovery in healthy elderly participants. Our results differ from the numerous studies that observed significant biomechanical differences between healthy elderly and healthy young participants during anterior perturbation recovery. (Thelen et al., 1997, 2000; Wojcik et al., 1999, 2001; Pavol et al., 2001; Madigan and Lloyd, 2005a, 2005b; Madigan, 2006). Many of those studies examined the recovery response after the release from the maximal lean angle that the participant could successfully recover. For these studies the participants leaned at an angle of approximately 30 , which is roughly equivalent to a lean of 36% of the participant’s body against the tether (Madigan and Lloyd, 2005a, 2005b). The studies concluded that many age-related differences in anterior perturbation recovery are only apparent during high-demand recovery (Thelen et al., 1997; Wojcik et al., 1999). Our study only used a lean of 10% of the participant’s body weight against the tether, which can be considered a mild perturbation. Potentially, this magnitude of perturbation is not great enough to reproduce the statistical differences seen in previous studies. During recovery from mild to moderate perturbations, such as those in our methodology, healthy elderly participants appear to have directionally specific postural instability. The decrease in postural control in healthy elderly participants when compared with healthy young participants is more evident in their recovery from posterior perturbations than in their recovery from anterior perturbations of the same magnitude. The predominant characteristic affecting posterior instability is the amplitude of their movement in the posterior direction, specifically the length of the first posterior recovery step. This shortened step limited their ability to quickly control the posterior movement of their COM. Furthermore, the shortened recovery step exhibited by the healthy elderly participants may be due to the decreased velocity of movement at the knee and hip.

CONCLUSION AND CLINICAL IMPLICATIONS In conclusion, this study provides insight into the baseline perturbation recovery response in healthy elderly participants. The main characteristic of instability for healthy elderly participants, especially during posterior perturbations, appears to be lower amplitude of movement during recovery. Healthy elderly participants demonstrated slower knee and hip movements (bradykinesia) along with decreased movement at the knee (hypokinesia). These movement deficits appeared to limit the length of recovery step and possibly the amount of foot clearance during recovery. Without a sufficient step, the healthy elderly participants could not produce adequate mechanical advantage to counteract the displacement of their COM, thus requiring multiple steps for successful recovery. During perturbations of greater magnitude, this multistep strategy could result in more falls because it takes longer to control the COM during a multistep recovery. Also, during greater perturbations, the COM may reach increased magnitudes and velocities, potentially making successful recovery impossible. The bradykinetic and hypokinetic movements demonstrated by our healthy elderly participants may provide targets for clinical interventions. Recent randomized trials and meta-analyses suggest that exercise regimens for

352

SINGER ET AL.

elderly individuals at high fall risk should incorporate strength and balance training (Sherrington et al., 2008; Clemson et al., 2012). However, programs focused on task specific step training may enable elderly individuals to achieve recovery steps fast enough and long enough to successfully restrain their COM (J€obges et.al.,2004).

LIMITATIONS AND DIRECTIONS FOR RESEARCH Despite the statistical significance that our study detected, several limitations exist in both recruitment and methods. For recruitment, our sample sizes are small, and our healthy elderly participants did not represent a high fall risk population. However, there were still clear agerelated movement abnormalities in these participants who did not report balance deficiencies. We suspect that these movement abnormalities will be even more significant in balance-impaired elderly individuals. The tetherrelease method that we used to create the repeatable perturbations in this study did not precisely simulate a fall that would occur outside of the laboratory. Despite the randomized release of the tether, the participants were aware that a perturbation would take place and in which direction the perturbation would occur. In addition, the order of perturbations was not randomized. For each participant, posterior perturbation occurred first, then anterior perturbation. Furthermore, there was no initial velocity during our induced perturbation, as there would be during a trip or slip outside of the laboratory. For these reasons, the stepping response examined in this study may not fully represent the response that typically occurs. Future studies should include larger samples at greater risk of falls (e.g., such as those with Parkinson disease or Multiple Sclerosis). Given its responsiveness to age-related differences, the tether-release method may be a relevant outcome measure to examine improvements in reactive postural responses. Lastly, further research is needed to determine the efficacy of training compensatory stepping, using tether-release like models, as a means of reducing fall risk (J€obges et.al.,2004; Roger et al., 2003).

VIDEO LEGENDS Video 1. Methods using the Tether-Release Model. This video depicts a healthy young participant recovering from simulated posterior and anterior perturbations. The participant is instrumented with 63 reflective markers in a modified Plug-in Gait formation with clusters and a chest harness. The subject is positioned on one of two in-ground AMTI force plates. The tether-release method is used to initiate the perturbations. The method consists of an electromagnet fixed to the wall with a tether extending from it to the chest harness worn by the participant. Also attached to the chest harness is a yellow safety tether. Additionally, a small force sensor is in-line with the tether and attached to the harness. This sensor measures the amount of force applied to the system when the participant leans against the tether. An audible cue is emitted from the computer when the force sensor registers a lean force equal to approximately 10% (range 5 9%–11%) of the participant’s body weight. Once the threshold is reached, the tether is released by remotely turning off the electromagnet, and the participant falls, initiating a recovery step onto the adjacent force plate.

Video 2. Events Determination for a Typical Healthy Young and Healthy Elderly Participant as Depicted in Visual 3D Software. After motion capture data were collected for each participant, Visual 3D (V3D) software was used to fit the participant with a biomechanical model based on reflective marker placement and anthropometrics. Using V3D, the stepping response was then analyzed and the variables were extracted. This video depicts a healthy young and healthy elderly participant recovering from posterior and anterior perturbations as visualized in V3D. The skeleton is a biomechanical representation of the participant based on the marker locations. The purple squares on the floor represent the force plates. The participant starts on FP1 and recovers onto FP2. The blue arrows represent the ground reaction force as measure by the force plates. The yellow dot on the force plates represents the movement of the Center of Mass (COM) projected onto the floor. In addition, the COM is also represented by the large blue sphere located in the pelvic region of the subject. Additionally, four real-time graphs of the trial accompany the video of the participant. These graphs depict 4 signals that were used to define several events and stages of the recovery process. The trial and graphs are zoomed into the section of interest, and all graphs show time along the X-axis. Graph A depicts the analog signal from the small in-line force sensor (Fz3). This sensor registers the amount of force applied to the system when the participant leans against the tether. When the sensor measures a threshold equal to approximately 10% of the participant’s body weight range (9%–11%), the tether is released and the force applied to the sensor plummets to zero. This signal is used to determine when the tether is released from the electromagnet, signifying the initiation of the perturbation. This “Tether Release” event is shown in the graphs as a yellow dash. Graph B depicts the velocity of the lateral ankle marker of the stepping leg in the posterior/anterior direction (RANK_vel). This signal is used to determine when the participant initiates the stepping response. To calculate this “Foot Off” event, depicted by the red dash, the mean and standard deviation of the static baseline signal is measured, then the event is placed at the point where the signal exceeds two standard deviations above the baseline mean. Graph C depicts force data of the rear force plate in the Z direction (FP1Z). This signal is used to determine when the participant makes contact with the rear force plate and completes the swing phase of recovery. This “Foot Strike” event is depicted by the blue dash in the graph. The blue arrow in the capture volume represents the ground reaction force registered by the force plate. Graph D depicts the location of the Center of Mass (COM) in the Y direction (COMY). The COM can be visualized in the model of the participant as the large blue sphere located in the pelvic region. The movement of the COM in the posterior/anterior direction is used to determine when the participant stops the movement of the COM, successfully regaining balance. This point occurs when the COM positions stops moving in the positive Y direction and starts to move in the negative Y direction. The “COM Stop” event is depicted by the green dash in the graphs. Video 3. Joint Angular Velocity of a Typical Healthy Young and Healthy Elderly Participant as Depicted in Visual 3D Software.

AGE-RELATED POSTURAL CONTROL DURING FALLS

This video depicts the recovery response of a healthy young and a healthy elderly participant during a simulated posterior and anterior perturbation as visualized in Visual 3D. Accompanying the video of each participant are 3 graphs depicting the flexion/extension joint angular velocity (JAVX) of the hip (a), knee (b), and ankle (c) of the stepping leg during the swing phase of recovery. Each graph depicts time along the X-axis. Tether Release is marked by the yellow dash. Foot Off is marked by the red dash. Foot Strike is marker by the blue dash. COM Stop is marked by the green dash. These events are further explained in Video 2. Graph A demonstrates the joint angular velocity (JAV) of the hip of the swing leg around the X-axis throughout flexion and extension from Foot Off to Foot Strike. Positive angular velocity is hip flexion, and negative angular velocity is hip extension. Graph B demonstrates the joint angular velocity (JAV) of the knee of the swing leg around the X-axis throughout flexion and extension from Foot Off to Foot Strike. Positive angular velocity is knee extension, and negative angular velocity is knee flexion. Graph C demonstrates the joint angular velocity (JAV) of the ankle of the swing leg around the X-axis throughout flexion and extension from Foot Off to Foot Strike. Positive angular velocity is dorsiflexion, and negative angular velocity is plantarflexion.

LITERATURE CITED Administration on Aging. 2009. A profile of older Americans: 2009. Washington, DC. Białoszewski D, Słupik A, Lewczuk E, Gotlib J, Mosiołek A, Mierzwi nska A. 2008. Incidence of falls and their effect on mobility of individuals over 65 years of age relative to their place of residence. Ortop Traumatol Rehabil 10:441-448. Blake AJ, Morgan K, Bendall MJ, Dallosso H, Ebrahim SB, Arie TH, Fentem PH, Bassey EJ. 1988. Falls by elderly people at home: prevalence and associated factors. Age Ageing 17:365-372. Campbell AJ, Borrie MJ, Spears GF, Jackson SL, Brown JS, Fitzgerald JL. 1990. Circumstances and consequences of falls experienced by a community population 70 years and over during a prospective study. Age Ageing 19:136-141. Campbell AJ, Reinken J, Allan BC, Martinez GS. 1981. Falls in old age: a study of frequency and related clinical factors. Age Ageing 10:264-270. Clemson L, Fiatarone Singh MA, Bundy A, Cumming RG, Manollaras K, O’Loughlin P, Black D. 2012. Integration of balance and strength training into daily life activity to reduce rate of falls in older people (the LiFE study): randomised parallel trial. BMJ 345:e4547-e4547. Czerwi nski E, Białoszewski D, Borowy P, Kumorek A, Białoszewski A. 2008. Epidemiology, clinical significance, costs and fall prevention in elderly people. Ortop Traumatol Rehabil 10:419-428. Do MC, Breniere Y, Brenguier P. 1982. A biomechanical study of balance recovery during the fall forward. J Biomech 15:933-939. Englander F, Hodson TJ, Terregrossa RA. 1996. Economic dimensions of slip and fall injuries. J Forensic Sci 41:733-746. Gu MJ, Schultz AB, Shepard NT, Alexander NB. 1996. Postural control in young and elderly adults when stance is perturbed: dynamics. J Biomech 29:319-329. Hall CD, Jensen JL. 2002. Age-related differences in lower extremity power after support surface perturbations. J Am Geriatr Soc 50:1782-1788. Horak FB, Shupert CL, Mirka A. 1989. Components of postural dyscontrol in the elderly: a review. Neurobiol Aging 10:727-738. Hsiao ET, Robinovitch SN. 1998. Common protective movements govern unexpected falls from standing height. J Biomech 31:1-9.

353

Hsiao ET, Robinovitch SN. 2001. Elderly subjects’ ability to recover balance with a single backward step assocaites with body configuration at step contact. J Gerontol 56:42-47. Hsiao-Wecksler ET. 2008. Biomechanical and age-related differences in balance recovery using the tether-release method. J Electromyogr Kinesiol 18:179-187. Hunt AL, Sethi KD. 2006. The pull test: a history. Mov Disord 21: 894–899. Jager TE, Weiss HB, Coben JH, Pepe PE. 2000. Traumatic brain injuries evaluated in U.S. Emergency Departments, 1992-1994. Acad Emergency Med 7:134-140. Jankovic J. 2008. Parkinson’s disease: clinical features and diagnosis. J Neurol Neurosurg Psychiatry 79:368-376. J€ obges M, Heuschkel G, Pretzel C, Illhardt C, Renner C, Hummelsheim H. 2004. Repetitive training of compensatory steps: a therapeutic approach for postural instability in Parkinson’s disease. J Neurol Neurosurg Psychiatry 75:1682-7. Kurz I, Oddsson L, Melzer I. 2013. Characteristics of balance control in older persons who fall with injury—a prospective study. J Electromyogr Kinesiol 23:814-819. Madigan ML. 2006. Age-related differences in muscle power during single-step balance recovery. J Appl Biomech 22:186-193. Madigan ML, Lloyd EM. 2005a. Age and stepping limb performance differences during a single-step recovery from a forward fall. J Gerontol A Biol Sci Med Sci 60:481-485. Madigan ML, Lloyd EM. 2005b. Age-related differences in peak joint torques during the support phase of single-step recovery from a forward fall. J Gerontol A Biol Sci Med Sci 60:910-914. Mathiyakom W, McNitt-Gray JL. 2008. Regulation of angular impulse during fall recovery. JRRD 45:1237. Overstall PW, Exton-Smith AN, Imms FJ, Johnson AL. 1977. Falls in the elderly related to postural imbalance. Br Med J 1:261-264. Pavol MJ, Owings TM, Foley KT, Grabiner MD. 2001. Mechanisms leading to a fall from an induced trip in healthy older adults. J Gerontol A Biol Sci Med Sci 56:M428-M437. Rogers MW, Johnson ME, Martinez KM, Mille ML, Hedman LD, 2003. Step training improve the speed of voluntary step initiation in aging. J Gerontol A Biol Sci Med Sci 58(1):46-51 Schiller JS, Kramarow EA, Dey AN. 2007. Fall injury episodes among noninstitutionalized older adults: United States, 20012003. Adv Data:1-16. Sherrington C, Whitney JC, Lord SR, Herbert RD, Cumming RG, Close JCT. 2008. Effective exercise for the prevention of falls: a systematic review and meta-analysis. J Am Geriatr Soc 56:2234-2243. Thelen DG, Muriuki M, James J, Schultz AB, Ashton-Miller JA, Alexander NB. 2000. Muscle activities used by young and old adults when stepping to regain balance during a forward fall. J Electromyogr Kinesiol 10:93-101. Thelen DG, Wojcik LA, Schultz AB, Ashton-Miller JA, Alexander NB. 1997. Age differences in using a rapid step to regain balance during a forward fall. J Gerontol A Biol Sci Med Sci 52:M8-M13. Tinetti ME, Speechley M, Ginter SF. 1988. Risks factors for falls in the elderly. N Eng J Med 219:1701-1707. Valkovicˇ P, Brozov a H, B€ otzel K, R˚uzicˇka Ee, Benetin J. 2008. Push and release test predicts better Parkinson fallers and nonfallers than the pull test: Comparison in OFF and ON medication states. Mov Disord 23:1453-1457. Visser M, Marinus J, Bloem BR, Kisjes H, van den Berg BM, van Hilten JJ. 2003. Clinical tests for the evaluation of postural instability in patients with parkinson’s disease. Arch Phys Med Rehabil 84:1669-1674. Winter DA, Prince F, Frank JS, Powell C, Zabjek KF. 1996. Unified theory regarding A/P and M/L balance in quiet stance. J Neurophysiol 75:2334-2343. Wojcik LA, Thelen DG, Schultz AB, Ashton-Miller JA, Alexander NB. 1999. Age and gender differences in single-step recovery from a forward fall. J Gerontol A Biol Sci Med Sci 54:M44-M50. Wojcik LA, Thelen DG, Schultz AB, Ashton-Miller JA, Alexander NB. 2001. Age and gender differences in peak lower extremity joint torques and ranges of motion used during single-step balance recovery from a forward fall. J Biomech 34:67-73. Yoshida S. 2007. A Global Report on Falls Prevention: Epidemiology of Falls. World Health Organisation.