Abstract
BACKGROUND:
One third of individuals after stroke report an inability to walk in the community. Community mobility requires walking adaptability — the ability to adjust one’s stepping pattern to meet environmental demands and task goals. Walking on uneven terrain (e.g. grass, gravel) has unique requirements and is a critical component of walking adaptability that has not been investigated in the post-stroke population.
OBJECTIVE:
To summarize current knowledge of biomechanical and neuromuscular modifications during uneven terrain negotiation in healthy individuals and discuss implications of post-stroke impairments.
METHODS:
Review of eleven studies, identified through a search of relevant literature on PubMed and CINAHL.
RESULTS:
On uneven terrain, healthy adults demonstrate numerous gait modifications including a lowered center of mass, increased muscle co-contraction during stance and exaggerated or increased toe clearance during swing. After stroke, changes in muscle activity and limb coordination will likely result in difficulty or inability performing these modifications that healthy adults use to maintain stability and safety when walking on uneven terrain.
CONCLUSIONS:
Studies of biomechanical and neuromuscular control of walking on uneven terrain are needed to quantify mobility limitations in adults post-stroke. Such investigations will contribute to the understanding of mobility impairments after stroke and the design of critically important intervention strategies.
Introduction
Walking at home and in the community requires walking adaptability, defined as the ability to adjust the basic stepping pattern to meet environmental demands and task goals (e.g., obstacle negotiation, dual-tasking) (Balasubramanian, 1999). Individuals after stroke often demonstrate considerable difficulty with walking adaptability, even if they have relatively good recovery of steady state walking ability (Hill, Ellis, Bernhardt, Maggs, & Hull, 1997). In fact, individuals with post-stroke walking deficits frequently report the avoidance of conditions requiring adaptability (Robinson, Matsuda, Ciol, & Shumway-Cook, 2013). Subsequently, over one-third of individuals post-stroke are unable to walk in the community (Lord, McPherson, McNaughton, Rochester, & Weatherall, 2004) and three quarters report difficulty in an “outdoor community” setting (Robinson, Shumway-Cook, Ciol, & Kartin, 2011). Recovery of community mobility after stroke is critical because it is associated with the health-related benefits of increased physical activity, improved quality of life through participation in life roles and a decreased risk for depression (Carod-Artal, Egido, Gonzalez, & Varela de Seijas, 2000). While current physical therapy practice and rehabilitation emphasize the recovery of ambulation after stroke, minimal therapy time is spent on training the ability to adapt walking (Latham et al., 2005). To address this critical gap in the recovery of walking and design rehabilitation interventions that target community mobility after stroke, an important first step is to understand how stroke affects the ability to perform adaptations.
One component of walking adaptability that is underinvestigated in the post-stroke population is negotiation of uneven terrain. Walking on uneven terrain (e.g. gravel, grass) is unique because the surfaces are typically compliant, non-uniform and, therefore, unpredictable. The characteristics of uneven terrain, as well as the transitions between various surfaces, make anticipation of task demands particularly challenging. For instance, during walking on uneven terrain healthy adults modify both step parameters and limb kinematics to meet the demands of the task (MacLellan & Patla, 2010). Furthermore, healthy adults seem to rely on somatosensory inputs to adjust their walking strategy on uneven terrain, as a lack of visualization of the surface has no impact on step parameters (Rogers, 2005). Overall, numerous strategies are used by healthy individuals during walking on uneven terrain to maintain stability and safety. Due to a critical gap in the understanding of mobility post-stroke, it is not known if people post-stroke would demonstrate similar adaptation strategies. It is likely that stroke-related impairments in motor, sensory and cognitive function will severely impair walking control during uneven terrain negotiation and adults post-stroke may adopt unique strategies.
To begin addressing this gap in the evidence, it is essential to have a good understanding of both the normal modifications used in uneven terrain negotiation and the implications of post-stroke walking deficits on these strategies. This step is critical prior to the development of research that investigates the changes that occur after stroke. Therefore, the first section is a scoping review which summarizes studies that have investigated gait modifications required for healthy adults to walk on uneven terrain. Scoping reviews employ a systematic search for evidence to address an explicit question, summarize findings in tabular form, and include a “narrative integration of the relevant evidence” (Dijkers, 2015). A search for relevant studies was conducted through PubMed and CINAHL using the search term “walking” with “uneven terrain,” “irregular surface,” or “compliant surface.” Additionally, the reference lists of the identified articles were explored for other pertinent studies. Articles were excluded if the paradigm involved a single unexpected perturbation, a slipping motion or a pathological population such as amputees or those with a neuropathy. The studies examined walking on a variety of surfaces (e.g. stones, foam and other combinations of materials) which were integrated into both overground and treadmill paradigms. Gait measures such as spatiotemporal parameters, lower extremity kinematics and muscle activity were examined. A table is included that summarizes the experimental design and key findings of each study (Table 1). This first section on gait modifications performed by healthy adults is separated into two categories: biomechanical and neuromuscular modifications. Subsequently, the second section provides a perspective on how these findings integrate with our understanding of common gait impairments after stroke. It concludes by providing insights into future steps needed to design investigations that will move the field towards enhanced knowledge about the mechanisms and recovery of the control of walking on uneven terrain in individuals post-stroke.
Studies investigating the effects of uneven terrain on gait
Studies investigating the effects of uneven terrain on gait
Step parameter modifications
When healthy adults walk on uneven terrain, changes in step parameters depend on the features of the uneven surface. For instance, walking on foam resulted in increased step length, width and time when compared to walking on level surface at the same speed (MacLellan & Patla, 2006). The wider and longer steps while walking on foam may provide an increased base of support making it easier to control the center of mass. While walking on two different surfaces, one composed of railroad ballast (irregularly shaped crushed stone) and the other of artificial grass over foam and small objects, a different adaptation pattern was observed; step time increased, similar to the response on foam, but step length decreased (Menant, 2010). It is possible that the reduced speed associated with walking under these conditions may have influenced the step length. However, Wade et al. (2010) described this strategy of shorter, slower steps as a “cautious gait” pattern. Walking with a shortened step length and increased step time could decrease the risk of slipping on surfaces where the foot has the potential to shift. Interestingly, unlike the overground walking studies that reported increases in step time, individuals adjusted with a decrease in step time as well as reduced step length during walking on an uneven treadmill surface (Voloshina et al., 2013). This strategy of reduced step time and length may have emerged if the individuals were unfamiliar with a treadmill paradigm or because they were unable to alter their speed.
Another alteration found during negotiation of uneven terrain is increased movement variability, evidenced in parameters such as step length, width and time (Gates, 2013). Previously, increased step parameter variability has been associated with significant increases in the metabolic cost of walking (O’Connor, Xu, & Kuo, 2012). Therefore, the increased movement variability observed during uneven terrain negotiation is likely a contributing factor to the substantial increase in metabolic cost when walking on a surface with a small increase in surface variability (Voloshina et al., 2013). Without consistent and appropriate timing of events between lower extremities during the gait cycle, the energy requirement increases due to lost efficiency in the step to step energy transfer and joint work must increase to accommodate. For instance, if the work performed in the trailing leg during push-off does not equal the work of the leading leg during loading response, or if the events are not simultaneous, additional energy is required during single limb stance to maintain forward progression (Kuo, Donelan, & Ruina, 2005). Another likely explanation for the increase in metabolic cost during uneven terrain negotiation is the energy lost to the compliant environment, such as shifting of stones or sand under the feet, which reduces forward propulsion (Lejeune, Willems, & Heglund, 1998).
Kinematic modifications
Similar to the changes observed in step parameters, healthy adults demonstrate modifications in gait kinematics to improve balance and prevent a fall when walking on uneven terrain. During stance, improved stability on an uneven surface is achieved by reducing the height of the center of mass (MacLellan & Patla, 2006) through a crouched position with increased hip and knee flexion (Gates et al., 2012). In addition, trunk and lower extremity adaptations assist in maintaining a steady head position to allow the visual and vestibular system to operate from a stable position and contribute more effectively to maintaining balance. Compared to walking on a level surface, walking on uneven terrain results in significantly increased pelvic accelerations (Menant et al., 2003a), but healthy young individuals demonstrate no significant change in head accelerations (Menz et al., 2003a). This suggests that trunk accommodations likely have a considerable role in successfully attenuating the perturbations created by the uneven terrain. In healthy adults who are older (aged 61–85), these trunk accommodations are impaired (Marigold & Patla, 2003b). Since the magnitude of accelerations increases exponentially as speed of walking increases (Menz et al., 2003a), the slower speed adopted by older adults on uneven terrain may be a strategy to maintain a tolerable level of head movement (Menz et al., 2003b). For example, the reduced speed adopted by older adults on uneven terrain has been found to result in lower pelvic acceleration magnitudes compared to younger individuals walking faster, but this reduction was not maintained at the head (Marigold & Patla, 2008) Instead, there was no significant difference between the groups in the magnitude of head accelerations in two planes, so the reduced speed was necessary to maintain the stable head position (Marigold & Patla, 2008).
Stability and safety improvements have shown to be achieved in swing through modifications that decrease the risk of tripping and slipping. Increased hip and knee flexion and dorsiflexion improves toe clearance, decreasing the risk of catching the toes and subsequently tripping on uneven terrain (Gates et al., 2011). Though dorsiflexion increased when averaged across the gait cycle, decreased dorsiflexion was found in terminal swing resulting in a “flat-footed” approach rather than a true heel strike (Gates et al., 2012). Gates et al. (Gates et al., 2012) proposed that the flat-footed approach reduces the coefficient of friction required at the shoe-floor interface and therefore prevents slipping.
Neuromuscular modifications during uneven terrain negotiation
Neuromuscular modifications evident in healthy adults while walking on uneven terrain include increased muscle activity, greater contraction amplitude variability between steps and increased co-contraction in lower extremity muscles. These findings, measured with surface electromyography (EMG), are consistent with the kinematic changes, step to step variability and increased demands associated with walking on uneven terrain. For instance, the increased variability of muscle contraction amplitude is congruent with the increased variability in spatiotemporal parameters (Voloshina et al., 2013). Nearly all muscles demonstrated this increase in variability, although the increase was greater in thigh than leg muscles. Variability in thigh muscle activation may be related to the increased pelvic accelerations evident when walking on uneven terrain. As discussed earlier, increased movement variability correlates with increased energy expenditure while negotiating uneven terrain (Voloshina et al., 2013). Increased muscle activity and co-contraction also add to this increase in energy expenditure (Voloshina et al., 2010). Both mean and peak muscle activities increase and the contraction bursts demonstrate an earlier onset and longer duration compared to activity recorded during walking on level surface (Wade et al., 2010). Voloshina et al. (Voloshina et al., 2013) suggested that increases in muscle activity during ambulation on uneven terrain may be a reflection of the increased demand created by maintaining a more crouched position. The increased co-contraction recorded in both the knee flexors/extensors and ankle plantarflexors/dorsiflexors during stance supports this argument (Voloshina et al., 2010). Another possible explanation, proposed by Wade et al. (Wade et al., 2010), is that increased co-contraction stabilizes the joint in anticipation of the perturbations created by the surface and distributes the increased joint forces throughout the articular surface. Increased co-contraction of ankle invertors and evertors shortly after initial contact supports this reasoning (Wade et al., 2010). One exception to the increase in muscle activity has been identified by separating the analysis of muscle activity by gait phase, rather than averaging over the entire gait cycle. Tibialis anterior activity decreases in late swing, consistent with the biomechanical modification described earlier where dorsiflexion is reduced in late swing when approaching an unstable surface (Voloshina et al., 2013).
Implications of gait adaptations during uneven terrain negotiation for people after stroke
The evidence on healthy individuals serves as a basis for understanding the gait modifications required to successfully navigate uneven terrains. The following section aims to consider these findings from the perspective of known impairments after stroke. At this time no studies have directly assessed how individuals after stroke adapt their walking to uneven terrain. An understanding of both the modifications required and the specific post-stroke impairments that may interfere allows for an appreciation of the challenges this environment creates. The effect of this challenge is evident in the high proportion of individuals who require assistance and the use of an assistive device when walking outside of the home environment (Skolarus, Burke, & Freedman, 2014) as well as the increased falls after stroke (Jorgensen, Engstad, & Jacobsen, 2002).
A primary characteristic of post-stroke gait is asymmetry which is most apparent in the increased stance time on the non-paretic lower extremity and in the differences between paretic and non-paretic joint kinematics, particularly knee flexion in swing (Olney & Richards, 2009). Much of this asymmetry stems from marked impairment in the ability to support one’s own body weight during single limb stance and produce propulsion of the body during terminal stance and pre-swing (Chen, Patten, Kothari, & Zajac, 2005). Both decreased stance time and reduced control of single limb stance on the paretic side affect the individual’s ability to generate non-paretic swing. On uneven terrain, this may reduce the non-paretic limb’s capacity to produce the swing phase adaptations that are required for safe foot clearance. An individual after stroke also demonstrates reduced dorsiflexion, knee flexion and hip flexion in swing when ambulating on a level surface (Burdett, Borello-France, Blatchly, & Potter, 1988). This is problematic considering that healthy individuals utilize an increase in these parameters to successfully negotiate uneven terrain (Gates et al., 2012). In addition, gait speed on level surfaces is already reduced by 50% or more in individuals with moderate to severe impairments after stroke (Olney & Richards, 1996). It is therefore likely that gait speed will be further reduced on uneven terrain as post-stroke impairments limit adaptation to an unpredictable surface.
Changes in neuromuscular control after stroke will likely disrupt specific adaptation strategies used by healthy individuals. For instance, individuals after stroke demonstrate prolonged muscle activity, as well as an increase in the proportion of the gait cycle that antagonistic muscles are simultaneously active (Den Otter, Geurts, Mulder, & Duysens, 2007) and a reduction in the complexity of intermuscular coordination after stroke (Clark, Ting, Zajac, Neptune, & Kautz, 2010). The compliance and unpredictability of an uneven terrain surface will likely amplify the effects of these impairments in timing and coordination because the individual would be unable to produce the variable and task-specific modulation of muscle activity needed to adapt to the surface. In addition, the increased variability in spatiotemporal step characteristics present in steady state walking after stroke (Balasubramanian, Neptune, & Kautz, 2009) may amplify the increased gait variability on uneven terrain as reported in healthy individuals. Therefore, while increased muscle activation and greater co-contraction are may be necessary to safely and successfully negotiate uneven terrain, individuals post-stroke already demonstrate reduced muscle activity and decreased co-contraction in the paretic limb during single limb support (Hirschberg & Nathanson, 2000). Furthermore, deficits in trunk control and changes in posture after stroke (Karatas, Cetin, Bayramoglu, & Dilek, 2004) would likely compromise the ability to dampen accelerations at the head as was seen in older adults (Menz et al., 2003b). Reducing head movement during gait allows the visual and vestibular systems to operate from a more stable platform as they contribute to the maintenance of balance. This deficit may be particularly detrimental for those individuals with stroke that already have a reduction in visual, vestibular and/or somatosensory inputs to balance.
The multitude of gait adjustments required to walk on uneven terrain may also increase the demand for attentional resources for individuals post-stroke. This is evident from studies demonstrating increased cortical activity during the performance of walking tasks that require adaptability (Clark, 2014). The complexity of the environment and the demands placed on the individual while walking on uneven terrain will likely result in heightened attentional demand and potential interference with control processes for any other simultaneous motor or cognitive task (Clark, 2015). This degrades performance and places the individual at increased risk for falling (Hyndman, 2006).
Although walking on uneven terrain has not been investigated in individuals post-stroke, some insights can be gained from the performance changes after stroke in other walking tasks requiring adaptability. For example, while individuals post-stroke were able to successfully clear obstacles of two different heights, they completed this task with significantly less knee flexion and dorsiflexion and therefore relied on compensatory strategies to lift the foot (MacLellan, Richards, Fung, & McFadyen, 2013). There was also a decrement in performance as evidenced by the greater movement variability. Examination of step characteristics also revealed an ineffective strategy; participants had increased vertical foot clearance but reduced post-obstacle foot clearance (Said, Goldie, Patla, & Sparrow, 2001). Overall, it is likely that not only are these decrements in performance and compensations after stroke seen across varying walking adaptability tasks, but also may be particularly evident on uneven terrain given the demands of an unpredictable and variable surface.
Interestingly, despite the sensorimotor limitations post-stroke, it is also possible that when challenged to walk on uneven terrains, individuals post-stroke may actually be able to generate appropriate task-specific modifications required to successfully adapt their walking. In support of this concept, when children with spastic diplegia walked on uneven terrain, they demonstrated appropriate adaptations of greater hip flexion, knee flexion and dorsiflexion rather than the hypothesized compensatory strategy (Bohm, Hosl, Schwameder, & Doderlein, 2014). This task-specific and appropriate adaptation is notable since children with spastic diplegia exhibit reduced hip flexion, knee flexion and dorsiflexion during typical walking. Recently, a similar finding in people after stroke supports the idea that performance of walking tasks that require adaptability may elicit appropriate improvements in muscle activation (Clark, Neptune, Behrman, & Kautz, 2016). Clark et al. (2016) showed that, when performing a long step leading with the non-paretic leg, individuals consistently demonstrated an appropriately timed increase in plantarflexor activity to improve paretic leg propulsion and achieve an elongated step. Therefore, interventions that include uneven terrain negotiation may facilitate greater gains in performance than those focused only on steady state walking practice.
Conclusion
Successful navigation of uneven terrain requires a variety of adaptations to the basic stepping pattern. As demonstrated in healthy adults, these modifications improve safety by avoiding slips or trips and provide the needed muscle activity and movement variability to allow the body to progress across the surface with control. Many of these modifications require motor, sensory and cortical resources that are affected by stroke. These post-stroke impairments may interfere with the adaptation strategies required to safely and successfully negotiate uneven terrains. Therefore, people after stroke may use altered and ineffective strategies to walk on uneven terrain, resulting in greater mobility deficits on uneven terrain compared to their walking ability on level surfaces.
To advance the study of walking adaptability, and specifically that of uneven terrain negotiation, it is important to investigate and quantify step parameter and kinematic modifications used by individuals after stroke. Quantification of stroke-specific strategies may be challenging because of inter-individual variability of gait mechanics in individuals post-stroke. However, investigation and quantification of adaptability strategies for individuals post-stroke will enable development of specific intervention approaches that promote the recovery of uneven terrain negotiation. Studies should also move beyond biomechanical measures to quantify the neural control strategies, such as the role of cortical activity in the performance of necessary gait adaptations. These types of investigations can advance our understanding of walking adaptability after stroke and move us towards the goal of improved community mobility for this population.
Conflict of interest
No potential conflict of interest was reported by the authors.