Coordination between motor and cognitive tasks in dual task gait

BackgroundDual Task (DT) paradigms are frequently used by researchers and clinicians to examine the integrity of motor processes in many movement disorders. However, the mechanism of this interaction is not fully understood. Therefore, the aim of this study was to examine the within-stride interactions between cognitive and motor processes during dual task gait (DT).Research questionDo healthy young adults coordinate gait with secondary task processing? If so, is cognitive task processing capability associated with the coordination observed?MethodsNineteen healthy young adults walked for two minutes on a motorized treadmill whilst counting backwards in sevens from three-digit numbers. The coordination of calculation verbalizations with gait parameters were assessed across six phases of the gait cycle. Mid verbalization time points (VER_Mid) were used as points of high cognitive processing of the dual task and compared with the end of the verbalizations (VER_End) as points of low cognitive processing.ResultsVER_Mid and VER_End did not systematically occur in any phase of the gait cycle. However, 10/19 and 9/19 participants showed non-random distributions of verbalizations for VER_Mid and VER_End time points respectively (p < 0.01), indicating that these walkers coordinated gait with the cognitive task. Analysis of subgroups of Verbalization Coordinators and Non-Coordinators showed slower verbalization response durations (VRD) for VER_Mid Coordinators compared to VER_Mid Non-Coordinators, indicating that VER_Mid Coordinators found the cognitive tasks more demanding. No differences were found in VRD for VER_End Coordinators and VER_End Non-Coordinators.SignificanceIt was found that cognitive processing is coordinated with gait phases in some but not all healthy young adults during DT gait. When demands on cognitive processes are high, healthy young adults coordinate cognitive processing with phases of gait. Analysis of within-stride coordination may be of use for studying clinical conditions where gait and attentional cognition performance breaks down.     

Gaze diversion affects cognitive and motor performance in young adults when stepping over obstacles

BackgroundIn many common multi-tasks, vision is used for two or more of the tasks, such as viewing cars, traffic signals, and the sidewalk curb at a crosswalk.Research QuestionHow does gaze diversion affect adaptive locomotion in young adults?MethodsSeventeen young adults completed a simple reaction time (RT) task while (1) standing and (2) during the approach to an obstacle on an 8 m walkway. Participants pressed a remote switch in response to a light cue (activated once during approach phase). The light cue was located either (1) on the obstacle (gaze diverted to obstacle) or (2) at eye level (gaze diverted away from obstacle). A gait baseline task with no RT task was included.ResultsAn interaction was observed (task (standing versus walking) by gaze location (on versus away from obstacle), p = 0.01), where RT was not affected by the gaze location in the standing task, but RT was longer when gaze was diverted away from the obstacle in the gait task. Furthermore, trail foot placement was closer to the obstacle when the gaze was diverted away from the obstacle (p = 0.002), which increased risk of tripping.SignificanceGaze diversion did not affect cognitive performance in the standing task, as information regarding the obstacle was not relevant for the standing task. However, completing a simple discrete visual cognitive task during obstacle crossing impaired both cognitive and gait performance, but only when gaze was diverted away from the obstacle. The impaired performance is likely due to the larger amount of structural interference when gaze was diverted away from the obstacle. These findings highlight the critical role of vision during the approach phase to an obstacle.     

Effects of motor practice on learning a dynamic balance task in healthy young adults: A wavelet-based time-frequency analysis

BackgroundPrevious research showed changes in amplitude- or time-derived measures of electromyographic (EMG) activity with motor learning. However, an analysis of the EMG spectral content (e.g., via wavelet technique) has not been included in these investigations yet.ObjectiveThe aim of this study was to use conventional, amplitude-derived EMG parameters along with modern, wavelet-based time-frequency EMG measures to assess the effects of motor practice on learning a dynamic balance task.MethodsNineteen young male adults (mean age: 26 ± 6 years) practiced a dynamic balance task for two days. Delayed retention test was performed on the third day. On a behavioral level, the root-mean-square error (RMSE) of the stability platform angle was calculated and used as outcome measure. On a neuromuscular level, EMG data from the tibialis anterior (TA) and the gastrocnemius medialis (GM) muscle were unilaterally recorded and analysed by calculating the integrated EMG (iEMG) and the EMG intensity (via continuous wavelet transforms).ResultsTwo days of practice resulted in significantly improved balance performance (i.e., lower RMSE) and TA/GM activation (i.e., reduced iEMG and EMG intensity) that was still present during the retention test on day 3. There was also evidence of practice-related changes in the EMG intensity pattern as indicated by an intensity shift from higher to lower frequency components.ConclusionsWe conclude that motor practice leads to improvements in movement effectiveness as indicated by reduced RMSE and in movement efficiency (i.e., decreased iEMG and EMG intensity, intensity shift). In addition to conventional amplitude-derived EMG parameters, modern, wavelet-based time-frequency EMG measures are appropriate to detect practice-related changes in muscle activation.     

Cognitive and visual demands,but not gross motor demand,of concurrent smartphone use affect laboratory and free-living gait among young and older adults

BackgroundAs smartphones are an integral part of daily activities, understanding the underlying mechanism associated with concurrent cell phone use while walking may help reduce the risks of injury.Research questionThis study examined the effect of cognitive, visual, and gross motor demands while using a phone during gait among young and older adults in the laboratory and free-living environments.MethodsTwelve young and twelve older adults walked along a 10-m walkway under five conditions: single-task walking (Walk), walking and bi-manually holding a phone (Walk-Hold), walking while looking at a phone held in front of the participants (Walk-Look), walking while answering questions (Walk-Answer), and walking while texting (Walk-Text). All conditions were performed in laboratory and free-living environments. Gait velocity, step time, step length, and cadence were obtained using a smartphone with a built-in accelerometer attached to the body. The dual-task cost (DTC) was also assessed. A three-way ANOVA was utilized for all parameters.ResultsWhile no three-way interactions were found for any parameter, group × condition interactions were significant for gait velocity, step time, step length, cadence and their corresponding DTC. Decreased gait velocity, step length and cadence, with increased step time was demonstrated during Walk-Look, Walk-Answer, and Walk-Text, compared to Walk and Walk-Hold. While older adults markedly changed their gait during Walk-Answer and Walk-Text, these changes were less pronounced among young adults.SignificanceVisual and cognitive demand while concurrently using a phone influenced gait, especially among the elderly. Environment did not accentuate gait alterations during concurrent phone use. Therefore, smartphone technology should be developed to detect dual-task walking and temporarily modify functionality to reduce risk of injury from divided attention.     

Selective effect of static stretching,concentric contractions,and a one-leg balance task on ankle motion sense in young and older adults

BackgroundBeing aware of ankle movement and motor control has a critical role in maintaining balance during functional activities such as standing, walking, and running. Since the somatosensory system declines with aging, this is even more important for older adults.Research questionHow do different exercise modalities (static stretching, one-leg balance task, concentric contractions, and control) acutely influence ankle motion sense in young and older adults?MethodsSeventeen young and fifteen older participants performed four different intervention protocols (static stretching, one-leg balance task, concentric contractions, and control) in random order. Each session comprised measurements of ankle motion sense in plantar flexion (PF) and dorsal flexion (DF) directions prior to and after an intervention protocol. Average threshold levels (in degrees) of motion sense detection were calculated from three trials in each direction (PF/DF).ResultsA lower threshold of motion ankle sense was observed for young adults compared to older adults regardless of the exercise modality and the direction of the movement (p < 0.001). However, the changes in PF and DF ankle motion senses followed a similar trend in both groups during the three exercise modalities: static stretching increased ankle motion sense threshold (PF: 14% and 5%; DF: 19% and 11% in young and older adults, respectively), concentric contractions decreased ankle motion sense threshold (PF: −24% and −14%; DF: −19% and −21% in young and older adults, respectively), and the one-leg balance task did not significantly influence the ankle motion sense threshold (PF: −1% and −2%; DF: 6% and 1% in young and older adults, respectively).SignificanceBased on these results, static stretching should not be performed before ankle activities that require a good balance, precision, and coordination. Concentric contractions could be recommended before activities that challenge our postural stability.     

Consistency and cost of dual-task gait balance measure in healthy adolescents and young adults

Matched control data are commonly used to examine recovery from concussion. Limited data exist, however, examining dual-task gait data consistency collected over time in healthy individuals. The study purposes were to: 1) assess the consistency of single-task and dual-task gait balance control measures, 2) determine the minimal detectable change (MDC) of gait balance control measures, and 3) examine the extent to which age and task complexity affect dual-task walking costs in healthy adolescents and young adults. Twenty-four adolescent (mean age = 15.5 ± 1.1 years) and 21 young adult (mean age = 21.2 ± 4.5 years) healthy participants completed 5 testing sessions across a two-month period, which involved analyses of gait balance control and temporal-distance variables during single-task and dual-task walking conditions in a motion analysis laboratory. Cronbach’s α and MDCs were used to determine the consistency of the gait balance control variables and the smallest amount of change required to distinguish true performance from change due to the performance/measurement variability, respectively. Dual-task costs were evaluated to determine the effect of task complexity and age across time using 3-way ANOVAs. Good to excellent test-retest consistency was found for all single-task and dual-task walking (Cronbach’s α range: 0.764–0.970), with a center-of-mass medial-lateral displacement MDC range of 0.835–0.948 cm. Greater frontal plane dual-task costs were observed during more complex secondary tasks (p < 0.001). The results revealed good-excellent consistency across testing sessions for all variables and indicated dual-task costs are affected by task complexity. Thus, healthy controls can be effective comparators when assessing injured subjects.     

The effects of cognitive versus motor concurrent task on gait in individuals with transtibial amputation,transfemoral amputation and in a healthy control group

BackgroundLower limb amputation causes difficulties in mobility together with motor and sensory loss. Challenging situations such as concurrent tasks cause gait parameters to deteriorate. Understanding the effect of concurrent tasks on gait is important for the rehabilitation of amputees.Research questionAre the effects of concurrent cognitive and motor tasks on gait parameters at fixed speed different in individuals with transtibial amputation, or transfemoral amputation compared to healthy individuals?MethodsThe gait parameters were evaluated of 20 individuals with transtibial amputation, 13 individuals with transfemoral amputation and 20 healthy individuals while walking on a motorized treadmill under single task (ST), cognitive dual task (CDT) and motor dual task (MDT) conditions. The self-selected comfortable velocity, which was determined in the single-task gait, was used in all three walking tests.ResultsST, CDT and MDT gait parameters of individuals with transtibial amputation, transfemoral amputation and healthy individuals were significantly different (p < 0.01). Covariance of step length variability increased in amputees when walking under MDT (p < 0.05). The dual task cost (DTC) for all the gait parameters was similar in all three groups (p > 0.05). The motor DTC of covariance of step length was greater than cognitive DTC (p < 0.05).SignificanceIndividuals with lower limb amputation have the capacity to walk with cognitive and motor tasks without changing velocity on the treadmill, but concurrent motor tasks cause an increase in gait variability. The results of this study suggest that there is an increase in gait variability especially with motor tasks, which may cause a higher risk of falling.Trial number: NCT04392466 (clinicaltrials.gov)     

Head orientation and gait stability in young adults,dancers and older adults

BackgroundControl of body orientation requires head motion detection by the vestibular system and small changes with respect to the gravitational acceleration vector could cause destabilization.Research questionWe aimed to compare the effects of different head orientations on gait stability in young adults, dancers and older adults.MethodsThree groups of 10 subjects were evaluated, the first composed of young adults (aged 18–30 years), the second composed of young healthy dancers under high performance dance training (aged 18–30 years), and the third group composed of community-dwelling older adults (aged 65–80 years). Participants walked on a treadmill at their preferred speed in four distinct head orientation conditions for four minutes each: control (neutral orientation); dynamic yaw (following a target over 45° bilaterally); up (15° neck extension), and down (40° neck flexion). Foot and trunk kinematic data were acquired using a 3D motion capture system and the gait pattern was assessed by basic gait parameters (step length, stride width and corresponding variability) and gait stability (local divergence exponents and margins of stability). Main effects of conditions and groups, as well as their interaction effects, were evaluated by repeated-measures analysis of variance.ResultsInteractions of group and head orientation were found for both step length and stride width variability; main effects of head orientation were found for all evaluated parameters and main effects of group were found for step length and its variability and local divergence exponents in all directions.SignificanceAs expected, the older adults group showed less stable gait (higher local divergence exponent), the shortest step length and greater step length variability. However, contrary to expectation, the dancers were not more stable. The yaw condition was the most challenging for all groups and the down condition seemed to be least challenging.     

Generalization of motor module recruitment across standing reactive balance and walking is associated with beam walking performance in young adults

BackgroundRecent studies provide compelling evidence that recruiting a common pool of motor modules across behaviors (i.e., motor module generalization) may facilitate motor performance. In particular, motor module generalization across standing reactive balance and walking is associated with both walking speed and endurance in neurologically impaired populations (e.g., stroke survivors and individual’s with Parkinson’s disease). To test whether this phenomenon is a general neuromuscular strategy associated with well-coordinated walking and not limited to motor impairment, this relationship must be confirmed in neurologically intact adults.Research QuestionIs motor module generalization across standing reactive balance and walking related to walking performance in neurologically intact young adults?MethodsTwo populations of young adults were recruited to capture a wide range of walking performance: professionally-trained ballet dancers (i.e., experts, n = 12) and novices (n = 8). Motor modules (a.k.a. muscle synergies) were extracted from muscles spanning the trunk, hip, knee and ankle during walking and multidirectional perturbations to standing. Motor module generalization was calculated as the number of modules common to these behaviors. Walking performance was assessed using self-selected walking speed and beam-walking proficiency (i.e., distance walked on a narrow beam). Motor module generalization between experts and novices was compared using rank-sum tests and the association between generalization and walking performance was assessed using correlation analyses.ResultsExperts generalized more motor modules across standing reactive balance and walking than novices (p = 0.009). Across all subjects, motor module generalization was moderately associated with increased beam walking proficiency (r = 0.456, p = 0.022) but not walking speed (r = 0.092, p = 0.349).SignificanceSimilar relationships between walking performance and motor module generalization exist in neurologically intact and impaired populations, suggesting that motor module generalization across standing reactive balance and walking may be a general neuromuscular mechanism contributing to the successful control of walking.     

Attentional costs of walking are not affected by variations in lateral balance demands in young and older adults

Increased attentional costs of walking in older adults have been attributed to age-related changes in visuomotor and/or balance control of walking. The present experiment was conducted to examine the hypothesis that attentional costs of walking vary with lateral balance demands during walking in young and older adults. Twenty young and twenty older adults walked on a treadmill at their preferred walking speed under five conditions: unconstrained normal walking, walking on projected visual lines corresponding to either the participant's preferred step width or 50% thereof (i.e. increased balance demand), and walking within low- and high-stiffness lateral stabilization frames (i.e. lower balance demands). Attentional costs were assessed using a probe reaction-time task during these five walking conditions, normalized to baseline performance as obtained during sitting. Both imposed step-width conditions were more attentionally demanding than the three other conditions, in the absence of any other significant differences between conditions. These effects were similar in the two groups. The results indicate that the attentional costs of walking were, in contrast to what has been postulated previously, not influenced by lateral balance demands. The observed difference in attentional costs between normal walking and both visual lines conditions suggests that visuomotor control processes, rather than balance control, strongly affect the attentional costs of walking. A tentative explanation of these results may be that visuomotor control processes are mainly governed by attention-demanding cortical processes, whereas balance is regulated predominantly subcortically.     

The effect of a dual task on gait speed in community dwelling older adults: A systematic review and meta-analysis

Background and purposeReduced walking speed in older adults is associated with adverse health outcomes. This review aims to examine the effect of a cognitive dual-task on the gait speed of community-dwelling older adults with no significant pathology affecting gait.Data sources and study selectionElectronic database searches were performed in, Web of Science, PubMed, SCOPUS, Embase and psychINFO. Eligibility and methodological quality was assessed by two independent reviewers. The effect size on gait speed was measured as the raw mean difference (95% confidence interval) between single and dual-task performance. Pooled estimates of the overall effect were computed using a random effects method and forest plots generated.Data extraction and data synthesis22 studies (27 data sets) with a population of 3728 were reviewed and pooled for meta-analysis. The mean walking speed of participants included in all studies was >1.0 m/s and all studies reported the effect of a cognitive dual-task on gait speed. Sub-analysis examined the effect of type of cognitive task (mental-tracking vs. verbal-fluency). Mean single-task gait speed was 1.21 (0.13) m/s, the addition of a dual-task reduced speed by 0.19 m/s to 1.02 (0.16) m/s (p < 0.00001), both mental-tracking and verbal-fluency tasks resulted in significant reduction in gait speed.Limitations and conclusionThe cross-sectional design of the studies made quality assessment difficult. Despite efforts, high heterogeneity remained, possibly due to participant characteristics and testing protocols. This meta-analysis shows that in community-dwelling older adults, the addition of a dual-task significantly reduces gait speed and may indicate the value of including dual-task walking as part of the standard clinical assessment of older people.     

Effect of muscle fatigue and physical activity level in motor control of the gait of young adults

The aim of this study was to analyze the effect of muscle fatigue in active and inactive young adults on the kinematic and kinetic parameters of normal gait and obstacle crossing. Twenty male subjects were divided into active (10) and inactive (10), based on self-reported physical activity. Participants performed three trials of two tasks (normal gait and obstacle crossing) before and after a fatigue protocol, consisting of repeated sit-to-stand transfers until the instructed pace could no longer be maintained. MANOVAs were used to compare dependent variables with the following factors: physical activity level, fatigue and task. The endurance time in the fatigue protocol was lower for the inactive group. Changes of gait parameters with fatigue, among which increased step width and increased stride speed were the most consistent, were independent of task and physical activity level. These findings indicate that the kinematic and kinetic parameters of gait are affected by muscle fatigue irrespective of the physical activity level of the subjects and type of gait. Inactive individuals used a slightly different strategy than active individuals when crossing an obstacle, independently of muscle fatigue.     

Effects of manual task complexity on gait parameters in school-aged children and adults

This study examined the dual-task interference effects of complexity (simple vs. complex), type of task (carrying a pitcher vs. tray), and age (young adults vs. 7–10 year old children) on temporal-spatial and variability measures of gait. All participants first walked on the GAITRite^® walkway without any concurrent task, followed by four dual-task gait conditions. The group of children had a more variable step length and step time than adults across all walking conditions. They also slowed down, took fewer, smaller steps and spent more time in double limb support than adults in the complex dual task conditions. Gait in healthy young adults and school aged children was relatively unaffected by concurrent performance of simple versions of the manual tasks. Our overall analysis suggests that dual-task gait in school aged children is still developing and has not yet reached adult capacity. This study also highlights the critical role of task demand and complexity in dual-task interference.     

Effects of manual task complexity on gait parameters in school-aged children and adults

This study examined the dual-task interference effects of complexity (simple vs. complex), type of task (carrying a pitcher vs. tray), and age (young adults vs. 7–10 year old children) on temporal-spatial and variability measures of gait. All participants first walked on the GAITRite^® walkway without any concurrent task, followed by four dual-task gait conditions. The group of children had a more variable step length and step time than adults across all walking conditions. They also slowed down, took fewer, smaller steps and spent more time in double limb support than adults in the complex dual task conditions. Gait in healthy young adults and school aged children was relatively unaffected by concurrent performance of simple versions of the manual tasks. Our overall analysis suggests that dual-task gait in school aged children is still developing and has not yet reached adult capacity. This study also highlights the critical role of task demand and complexity in dual-task interference.     

Enhanced gait variability index and cognitive performance in Asian adults: Results from the Yishun Study

BackgroundAlthough gait variability has been linked to cognitive decline among older adults, the lack of a comprehensive composite gait variability score has dampened the application of gait variability.Research questionDoes the enhanced gait variability index (EGVI) - a composite score gait variability index - provide differential and useful information on cognitive decline in community-dwelling adults from that using gait speed?MethodsHealthy community-dwelling adults (n = 311) aged 21–90 were individually administered the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS). Habitual gait speed and spatiotemporal parameters were measured using a 6 m instrumented walkway system. The EGVI for each participant was calculated from five spatiotemporal parameters - step length(cm), step time(s), stance time(s), single support time(s) and stride velocity(cm/s). Linear regression models, controlling for age, gender, and education, were built to examine the independent effects of EGVI or gait speed on global cognition and individual domains.ResultsMultiple regression revealed that gait speed contributed significantly to the performance of the domain “Attention” (p = 0.04) whereas EGVI contributed significantly for the performance of the domain “Visuospatial” (p = 0.04) and “Delayed Memory” (p = 0.02).SignificanceEGVI provides differential and useful information from using gait speed alone. The EGVI may offer a solution to measure or track GV changes in relation to cognitive changes.     

The effect of exercise on gait and balance in patients with chronic fatigue syndrome

This study investigated anecdotal reports of gait and balance abnormalities in subjects with Chronic Fatigue Syndrome (CFS) by examining the effects of a light exercise test on postural sway and various gait parameters. Tests were performed on 11 CFS patients and 11 age- and sex-matched sedentary controls. Results demonstrated that postural sway was not significantly different in both groups before or after the exercise test. There were, however, significant differences in gait parameters between the two groups confirming anecdotal evidence, but these differences were not exacerbated by the exercise test. Heart rate responses demonstrated that both groups were exercising at similar loads, although this was perceived to be higher by the CFS group.     

Attentional demands of cued walking in healthy young and elderly adults

Among a wide range of negative consequences stemming from excess mass in children, recent studies suggest an impairment of postural control, including basic capabilities such as static and dynamic balance. Such impairment may be compounded when additional tasks are performed, such as carrying localized loads as occurs among children using a backpack. To investigate this, postural control was measured among 77 overweight and obese children (6-11 years old), and an equal number of normal-weight children matched for gender, age, and height. Testing was conducted at school, in which center of pressure (COP) time series during quiet standing were obtained in the presence and absence of each student's backpack. A traditional postural control measure derived from COP (mean velocity) did not indicate significant differences between overweight and normal-weight children, regardless of backpack presence. In contrast, a complexity index (derived from multiscale entropy) suggested the existence of different postural strategies and reduced balance capabilities among overweight children, whose consequences need to be further clarified.     

The effect of physical exhaustion on gait stability in young and older individuals

Fatigue directly affects key features of the sensorimotor system which disorganizes voluntary control of movement accuracy. Local dynamic stability of walking is considered a sensitive measure for neuromuscular performance. To gain greater insight in the role of fatigue in motor behaviour in older and young adults during walking, the current experiment analyses gait patterns of healthy young but maximal fatigued individuals and gait patterns in submaximal fatigued older adults.Ten young and 18 older subjects performed a bicycle incremental exercise test on a cycle ergometer. In young subjects, the incremental test was performed until total physical exhaustion. In older subjects, the test was performed until submaximal fatigue. Prior to and after the test, the participants walked for 2.5 min on a treadmill. Based on linear acceleration data of the trunk, local dynamic stability was assessed. Student's t-test was used to check if differences are statistically significant.In young individuals, we found a significant decrease in the finite-time maximal Lyapunov exponents between unfatigued walking and maximal fatigued walking. In older participants, significant increases in the finite-time maximal Lyapunov between unfatigued walking and submaximal fatigued walking were observed.The results indicate that (1) young and sporty subjects become more stable after having passed a maximum cardiopulmonary exercise test on a cycle ergometer while (2) older individuals walk less locally stable in a submaximal fatigued condition. Older cohorts might show a higher fall risk when they are physically fatigued.     

The metabolic cost of walking balance control and adaptation in young adults

BackgroundOur aim was to quantify the role of metabolic energy cost in governing neuromuscular adaptation to prolonged exposure to optical flow walking balance perturbations in young adults.Research questionWe hypothesized that metabolic cost would increase at the onset of balance perturbations in a manner consistent with wider and shorter steps and increased step-to-step variability. We also hypothesized that metabolic cost would decrease with prolonged exposure in a manner consistent with a return of step width and step length to values seen during normal, unperturbed walking.MethodsHealthy young adults (n = 18) walked on a treadmill while viewing a virtual hallway. Optical flow balance perturbations were introduced over a 10-minute interval during a 20-minute walking bout while measuring step kinematics and metabolic energy cost. For all outcome measures, we computed average values during the following four time periods of interest: Pre (minutes 3–5), Early Perturbation (minutes 5–7), Late Perturbation (minutes 13–15), and Post (minutes 18–20). A repeated-measures ANOVA tested for main effects of time, following by post-hoc pairwise comparisons.ResultsWith the onset of perturbations, participants walked with 3% shorter, 17% wider, and 53–73% more variable steps. These changes were accompanied by a significant 12% increase in net metabolic power compared to walking normally. With prolonged exposure to perturbations, step width and step length tended toward values seen during normal, unperturbed walking – changes accompanied by a 5% reduction in metabolic power (p-values≤0.05).SignificanceOur study reveals that the adoption of generalized anticipatory control at the onset of optical flow balance perturbations and the subsequent shift to task-specific reactive control following prolonged exposure have meaningful metabolic consequences. Moreover, our findings suggest that metabolic energy cost may shape the strategies we use to adapt walking balance in response to perturbations.