Early evidence of gait impairments following concussion were reported by McCrory based on video analysis of concussions sustained in the mid-1990s by competitors in Australian rules football [107]. Gait impairments, operationally defined as ataxic, stumbling, or unsteady gait, were noted post-injury in 41 % of concussed athletes with the majority manifesting symptoms immediately post-injury; however, a small percentage, 14 %, had a minimal delay of 10–20 s prior to the onset of gait unsteadiness [107]. While not specifically studied, it was speculated that gait unsteadiness involved a brainstem pathology and was multifactorial including postural tone, cerebellar, and labyrinthine function [107]. This study, while limited to gross video observations without true biomechanical assessment, provided foundational evidence of post-concussion gait impairments.

Compared to other commonly investigated neurological pathologies (e.g., Parkinson disease, elderly fallers, stroke, amputee), investigations of gait to identify impairments in postural control post-concussion have been fairly limited. Indeed, there are more review articles (e.g., systematic reviews, meta-analyses) on gait in the elderly than original research articles related to concussion and gait. The majority of gait studies were performed at one laboratory and were largely delimited to grade II concussions, as defined by the American Academy of Neurology (no LOC and symptoms persisting longer than 15 min) [108], had fairly homogeneous and small (n = 10–17/group) participant populations for most studies, lacked within-subject pre-injury data, and have involved a variety of gait tasks including single-task gait, dual-task gait with working memory challenges, and obstacle avoidance tasks [98–104, 109–111]. Finally, not all raw data is provided for all dependent variables of interest limiting the ability to perform a meta-analysis of the findings.

Utilizing traditional clinical measures of balance (e.g., BESS), large cohort investigators have suggested that postural control returns to its baseline value within 3–5 days post-injury [53]. The post-concussion gait studies which have been currently published are limited by lack of within-subjects baseline data; however, they are generally tightly matched to otherwise healthy control subjects. Within this context, gait velocity generally appears to return to a normal value by day 5 or 6 post-injury despite still experiencing concussion-related symptoms [99–101, 111], although in one study it had not recovered by day 28 [110]. This finding and other similar findings need to be taken in context as an apparent practice effect was potentially a confounding variable as the gait velocity steadily increased with each testing session in the healthy control group. However, by day 28 the concussion subjects had still not reached the initial and lowest gait velocity of the control subjects [110]. Similar findings were noted in the stepping characteristics (stride time, width, and length) [111] and sagittal plane COM measurements (anterior displacement and velocity of the COM and the anterior center of pressure [COP]–COM separation) [99, 111]. Frontal plane kinematics may be a more challenging task post-concussion as there is a limited base of support during the single-support phase of gait [112]. During single-task gait, post-concussion participants demonstrated limited increases in the medial to lateral COM range of motion and velocity [98, 102, 104, 111]. These impairments appear to persist for up to 28 days post-injury despite apparent recovery on the traditional clinical assessment battery [104, 111]. Interestingly, in many of these studies there was an apparent recovery on many dependent variables by day 5 post-injury, but significant differences reemerged 2–4 weeks later. These findings suggest either potential differences on day 5 are not statistically significant due to small groups and the possibility of the study being underpowered or some residual consequence of delayed impairment following a concussion. Overall, these gait studies suggest that a conservative gait strategy has been adopted post-concussion, although the rationale for these strategies remains unknown.

Adopting from methodologies utilized with elderly and diseased state patients, the addition of a cognitive challenge to the motor task of gait is beginning to be explored post-concussion. Both impairments in postural control and cognitive processing are known acute consequences of a concussion and thus not surprisingly both are abnormal when tested within 24 h of the injury. Post-concussion gait studies utilizing a dual-task paradigm have largely focused on utilizing working memory challenges (e.g., reciting the months of the year backwards) from the mini mental examination to assess cognitive performance while performing either level over ground gait or obstacle avoidance [113]. A recent systematic review and meta-analysis by Lee et al. [96] suggested that gait velocity and frontal plane range of motion are sensitive markers of dual-task interference in post-concussion individuals (Fig. 8.3).

Specifically, a pooled mean decrease in gait velocity of 0.13 m/s was noted across the meta-analysis which is similar to a noted decrease of 0.17 m/s across a diverse population of neurologically impaired participants [96, 114]. In the acute recovery phase following a concussion, similar to single-task gait, multiple impairments are noted during dual-task gait. Specifically, when compared to tightly matched control subjects, reductions in stride length, anterior velocity of the COM, and COM displacement in the frontal plane have been reported [98–103, 110]. Consistent with many dual-task paradigms, most kinematic characteristics of gait were reduced with the addition of a cognitive task in both the recently concussed and healthy control groups [98–103, 110]. The recovery patterns of dual-task gait were similar to, but expand upon, the single-task gait with apparent lingering deficits still present up to 28 days post-injury [99–101, 110]. Once again, the frontal plane kinematics appeared most sensitive to the identification of delayed recovery following a concussion [112]. These findings support the necessity of a multifaceted concussion assessment as most cognitive and postural control assessments are recovered far before 28 days post-injury [53]. Further, when compared to a commonly utilized computerized neuropsychological test, there was little relationship with the dual-task gait performance [99]. The authors speculated that dynamic motor tasks, such as dual-task gait, are potentially more complex and challenging than traditional computerized neuropsychological tests and may better approximate the demands experienced during sports participation [99].

The results of these combined studies suggest that impairments in postural control persist for up to 1 month post–injury despite resolution on the traditional clinical assessment battery. Further, an interesting, but unexplained, finding was the apparent recovery within a week post-injury, but residual impairment in performance which persisted up to a month. Future research needs to elucidate the reasons for this altered performance.

The remaining postural control data presented in this chapter is derived from the Georgia Southern University Concussion Management research protocol. This data represents 84 participants (Ht: 1.74 + 0.13 m; weight: 79.7 + 23.5 kg; age: 19.6 + 1.4 years; 50.7 % with a previous history of concussion [0.8 + 1.1 overall]) who suffered a sports-related concussion. The concussion initial presentations (9.5 % LOC; 34.5 % posttraumatic amnesia) and recovery timelines (symptom-free: 4.8 + 3.1 days; BESS recovery: 2.9 + 2.8 days; standard assessment of concussion recovery [SAC]: 2.2 + 1.8 days) are consistent with previous large epidemiological studies [32, 33, 53]. All participants completed a graduated and progressive return to participation exercise protocol, generally consistent with the third International Consensus Statement on Concussion in Sport (3rd CIS) [7], and the average time to unrestricted return to participation was 12.6 + 5.1 days.

The post-injury assessment protocol has been modified over the years as new information and recommendations have been incorporated. Specifically, the protocol was established in the late 2008 and did not incorporate computerized neuropsychological testing until 2010. The postural control testing occurred in the biomechanics laboratory which contains four force plates (AMTI, Watertown, MA) and an instrumented walkway (GAITRite; CIR Systems, Sparta, NJ); see also Fig. 8.4. Following a concussion, injured student-athletes performed the BESS, SAC, and a graded symptom checklist (22-items, 0–6 Likert scale) daily until they achieved their baseline values on each specific test. Participants were tested daily with over 90 % compliance.

The control of posture and locomotion are interdependent at several levels of the central nervous system [115]. Therefore, impaired posture and gait components may contribute to deficits in locomotion due to adaptive changes in neural control [115]. Many post-concussion balance assessments (e.g., BESS, SOT) are novel challenges and, as described, are subject to a substantial practice effect with repeat administration [79–83]. Thus, we have opted to utilize what we refer to as “non-novel” tasks—these are tasks which are performed as regular activities of daily living and therefore not subject to a practice or learning effect. One task commonly utilized to investigate the interactions between posture and locomotion components is gait initiation (GI). Indeed, GI, the phase between motionless standing and steady state locomotion requiring the generation of propulsive forces, has been shown to be a sensitive indicator of balance dysfunction [116]. GI challenges the postural control systems as it is a volitional transition from a large stable base of support to a smaller continuously unstable posture during gait [117]. From a motor control perspective, GI requires the central nervous system to regulate the spatial and temporal relationship between the position and motion of the COM [118]. Therefore, GI has been used to quantify impairments in postural instability amongst elderly, Parkinson disease, stroke, and amputee patient groups [119–125].

During static stance, the COP and COM are tightly coupled and located just anterior to the malleolus [124]. To initiate gait, they must decouple to generate forward momentum while maintaining upright balance [124, 126]. Initially, the COP moves posteriorly and laterally towards the initial swing limb (Fig. 8.5). This anticipatory postural adjustment (APA), controlled by the supplementary motor area and/or premotor area, involves bilateral tibialis anterior activation and soleus inhibition [127–129]. The initial posterior COP movement generates the forward momentum needed to separate the COP and COM while the lateral COP displacement, controlled by the gluteus medias, propels the COM towards the initial stance limb [130]. This momentum generation is necessary to achieve successful forward locomotion while maintaining upright balance. Thus, the initial posterior and lateral COP displacements are sensitive indicators of balance dysfunction [119, 120, 125, 131].

Following a sports-related concussion, impairments in GI have been noted. A typical healthy adult will displace his or her COP approximately 5–7 cm both posteriorly and laterally during the APA phase of GI. One day post-concussion, the otherwise healthy adults’ APA posterior displacement was 2.59 + 1.62 cm; a 131 % decrease compared to a normal healthy adult. Similarly, the lateral displacement of the COP post-concussion is reduced to 3.43 + 1.92 cm; a decrease of ~75 % from a healthy adult. As the posterior and lateral displacement during the APA is believed to generate the momentum needed to accelerate the COM forward, it is not surprising that the initial step length (0.60 m) and velocity (0.58 m/s) are substantially reduced compared to population norms [125]. An exemplar COP displacement trace is provided in Fig. 8.6 and particular attention should be paid to the APA phase noting that the COP at movement initiation is nearly identical between traces. This postural conservative strategy, unlikely to be associated with a fear of falling as is commonly suggested in neurologically impaired older adults, is consistent with gait-based studies comparing post-concussion postural control to healthy adults.

While comparison to healthy individuals is valid, comparing the individuals to their own premorbid performance is ideal. While the observed differences may appear small (i.e., only a few centimeters), the effect size of these differences needs to be considered. Effect size is a measure of the magnitude of the difference between groups and a value of 0.2 is considered a small effect, 0.5 a medium effect size, and 0.8 a large effect size. There is some debate on using effect size on within-subjects measures, but this is largely focused on varying treatment effects which are not present within this data set [132]. Following a sports-related concussion, individuals reduce their APA posterior displacement from a premorbid value of 5.46–2.34 cm, statistically significant (p < 0.001) with a large effect size (d) of 1.99. Similarly, the lateral displacement during the APA phase is reduced from 5.55 to 3.25 cm, statistically significant (p < 0.001) with a moderate effect size of 0.51. The reductions in APA COP displacement are likely associated with the reduction in initial step length (PRE: 0.68 + 0.11 m and day 1: 0.60 + 0.09, p = 0.001, d = 0.37) and step velocity (PRE: 0.67 + 0.17 m/s and day 1: 0.58 + 0.15 m/s, p = 0.021, d = 0.27). These results suggest that the largest impairments are noted in the APA component of GI as opposed to the resulting stepping characteristics. This likely occurs as the APA component is a supraspinal or central control process whereas the stepping characteristics are likely controlled at multiple levels including supraspinal (motor cortex), spinal (central pattern generators), and peripheral (local neuromuscular adaptations) [127–129, 133–135].

The post-concussion response to gait has been well established through a series of studies conducted in Li-Shan Chou’s lab at the University of Oregon and has been discussed previously [98–105, 109–111]. Our data is similar with noted deficits on the day following the concussion. Specifically, substantial decrements in performance were noted in gait velocity (PRE: 1.49 + 0.13 m/s and day 1: 1.17 + 0.13 m/s, p < 0.001, d = 0.77), mean step length (PRE: 0.76 + 0.04 m and day 1: 0.64 + 0.05 m, p = 0.002, d = 0.76), percentage of the gait cycle in double support (PRE: 22.48 + 2.38 % and day 1: 24.59 + 2.23, p = 0.035, d = 0.42), and percentage of the gait cycle in the swing phase (PRE: 38.75 + 1.12 % and day 1: 37.7 + 1.21 %, p = 0.05, d = 0.41). In a small subgroup, increases in gait variability, expressed as a coefficient of variation, have been identified post-concussion. Gait variability is an indicator of the rhythmicity and gait stability and is known to be impaired in elderly individuals or those with neurological impairments [136–139]. A variability of greater than 7 % has been associated with impairments in postural control in older individuals with neurological impairments; however, in healthy adults normal variability is below 3 % [140–144]. While the post-concussion individuals in this study did not exceed the 7 % threshold, there were increases from baseline (<3 %) to day 1 post-injury (>3 %). Consistent with the Oregon findings, these results suggest a conservative gait strategy is adopted following a concussion. However, the neurophysiological explanation has not been fully elucidated.

These previous findings suggest that, acutely post-concussion, impairments in postural control are identified with single (motor)-task challenges; however, emerging evidence suggests that reallocation of attentional resources and/or neural plasticity may allow the individual to overcome simple single-task challenges [62, 97, 145]. Dual-task challenges examine the effect of executing a secondary cognitive task (e.g., mental processing) on the concurrent performance of a primary motor task (e.g., walking) [145]. Even routine motor activities, such as sitting, standing quietly, or walking, require cognitive processing [146]. Previous investigations noted impaired postural control during both quiet stance and gait in healthy young adults under dual-task conditions [147, 148]. Simultaneous performance of a motor task and a cognitive task may interfere with the performance of one or both, probably due to competing demands for inherently limited attentional resources [149–151]. Utilizing working memory challenges (e.g., serial 7’s), post-concussion participants had further reductions in the displacement of the COP during the APA phase of GI and took shorter and slower steps than during single-task GI. Consistent with the findings from the Oregon studies, those differences were also present during gait with significant reductions in gait velocity, stride length, single-support phase, and swing phase when compared to healthy young adults. Interestingly, there were no differences noted within subjects when comparing single- and dual-task gait; potentially due to an inverted “ceiling” effect whereby individual’s performance during single task was already dramatically impaired. Indeed, both single- (1.17 m/s) and dual-task (1.15 m/s) gait velocities fell below the 1.2 m/s threshold often cited for a healthy gait in elderly and neurologically impaired individuals [152, 153].

Gait termination (GT) is not a mirror image of GI [154]. Rather, GT is a process by which the central nervous system anticipates, controls, and arrests the forward momentum of the COM without exceeding the borders of the base of support [155, 156]. Further, GT has a known and invariant set of parameters that constrains the multiple degrees of freedom within the lower extremity [157–159]. However, GT poses a unique challenge to the postural control systems because the COM is often located outside the base of support at the onset of GT [126]. As a result, GT is an excellent model for investigating alterations in motor programming and neurologic dysfunction. The central neurophysiologic control of GT is more elusive than GI; however, an fMRI study has suggested that the prefrontal area, specifically the inferior frontal gyrus and the pre-supplementary motor area, likely controls GT [154]. Indeed, GT has quantified impairments in postural control amongst the aging, people with Parkinson disease, amputee groups, chronic ankle instability, and those with general balance disorders [158–168].

The termination of gait requires the coordinated activity of both legs. Indeed, force production is modulated bilaterally such that the lead limb (limb behind the COM) reduces foot push-off propulsive forces as the swing limb (limb in front of the COM) concurrently increases vertical and anteroposterior braking forces [157, 158, 169]. Reduced propulsive forces are caused by soleus inhibition and increased activation of the tibialis anterior while concurrent increases in braking forces are due to an increased soleus activity and inhibition of the tibialis anterior [166, 170]. Failure to reduce lead limb propulsive forces results in an increased reliance on a single-limb stopping strategy and subsequently longer termination times and a greater number of steps required to control the COM [160]. Therefore, increases in propulsive and braking forces have been identified as sensitive indicators of balance dysfunction and alterations in the cortical control of GT [158, 160, 165]; see also Table 8.1.