Potential Contributions of Training Intensity on Locomotor Performance in Individuals With Chronic Stroke : Journal of Neurologic Physical Therapy

Secondary Logo

Journal Logo

Research Articles

Potential Contributions of Training Intensity on Locomotor Performance in Individuals With Chronic Stroke

Holleran, Carey L. MPT, NCS; Rodriguez, Kelly S. MPT, NCS; Echauz, Anthony DPT; Leech, Kristan A. DPT; Hornby, T. George PT, PhD

Author Information
Journal of Neurologic Physical Therapy 39(2):p 95-102, April 2015. | DOI: 10.1097/NPT.0000000000000077



Many studies have evaluated the efficacy of specific interventions to improve walking ability in individuals poststroke, although the critical training parameters that maximize recovery remain unclear. Previous studies in animal models1,2 and humans with neurologic injury3–6 suggest that the amount of task-specific walking practice may contribute to walking recovery. Specifically, early studies supporting the use of weight-supported treadmill training often compared this intervention to therapies that provided very little walking practice,3,4 with outcomes possibly due to differences in amount of stepping practice. More recent data suggest a relationship between the amount of stepping practice and improvements in locomotor performance.5,6 Additional factors are likely important, however, as suggested by the LEAPS trial,7,8 wherein large amounts of stepping practice did not result in greater improvements in walking function as compared with interventions that provided limited walking practice.

One training parameter that may be important but has received less attention is the intensity of stepping training. Consistent with the field of exercise physiology, training intensity is defined here as the rate of work performed (ie, workload or power) during continuous locomotor tasks and is estimated indirectly through metabolic and cardiorespiratory measures.9 The benefits of high-intensity treadmill training compared with control interventions in individuals with chronic stroke has been demonstrated in multiple studies,10–12 although training intensity may have contributed to results from other studies as well. In previous investigations comparing robotic therapist-assisted treadmill and/or overground training,13,14 potential differences in training intensity15,16 were thought to contribute to the smaller gains attained with robotic-assisted stepping, despite greater or equivalent amounts of stepping practice. Furthermore, in the LEAPS trial,8 average midtraining heart rates (HRs) during locomotor training were 90 beats per minute, which is well below aerobic training zones achieved in other treadmill training studies.5,12 Reduced cardiorespiratory responses during locomotor training indicate decreased metabolic and neuromuscular demands17–19 and may limit the potential gains in walking outcomes.

While both the amount and intensity of stepping practice may influence walking recovery, previous studies comparing high- versus low-intensity interventions do not differentiate between the contributions of these training parameters. More directly, high-intensity treadmill training is often compared with low-intensity interventions that provide limited walking practice.5,8,12 The lack of distinction between stepping amount and intensity is also evident in studies comparing faster versus slower treadmill training interventions,20,21 where greater cardiovascular demands and larger amounts of stepping practice may be achieved at faster speeds with similar training durations. To date, no studies have compared the specific effects of training intensity during locomotor activities while controlling for amount of stepping practice.

The purpose of this preliminary study was to assess the effects of high- versus low-intensity locomotor training on walking outcomes in individuals with chronic hemiparesis poststroke. Using a cross-over repeated-measures design, participants were provided treadmill and overground training, with intensity manipulated by altering selected biomechanical demands of walking while controlling for the amount of stepping practice. Understanding the potential independent contributions of training intensity during locomotor interventions may be important to maximize the efficacy of physical therapy interventions and augment mobility outcomes.


Individuals with chronic (≥6 months) hemiparesis following unilateral supratentorial stroke were recruited. Inclusion criteria consisted of the following: age 18 to 75 years, overground self-selected walking speed less than 0.9 m/s without physical assistance but with assistive devices and bracing below the knee as needed; Mini-Mental State Examination score ≥ 23/30; and medical clearance to participate. Exclusion criteria included severe lower extremity contractures limiting walking performance, significant osteoporosis, cardiovascular or metabolic instability, inability to ambulate independently more than 50 m prior to stroke, previous history of peripheral or central nervous system injury, and inability to adhere to study requirements. Patients could not be concurrently enrolled in physical therapy during training and could not have participated in a research trial for the past 3 months. All training and testing was performed at the Rehabilitation Institute of Chicago. The project was approved by the Northwestern University Institutional Review Board, and all participants provided written informed consent.

Using a repeated-measures crossover design, participants were randomized to receive 12 or fewer sessions over 4 to 5 weeks of either a high- or low-intensity training, followed by a 4-week washout, after which participants received the other training paradigm (Figure 1). The target sample size for this pilot study was 12 participants, with a blocked randomization strategy generated using a randomization code (3 blocks of 4 participants). With no preliminary estimates of effect sizes, a sample size of 12 participants was chosen to provide equivalent numbers of subjects in each 4-participant block.

Figure 1:
Testing and training paradigm of study design.

Participants were assessed before and following each training period, with primary measures of self-selected velocity (SSV), fastest-possible velocity (FV), 6-Minute Walk Test (6MWT), and peak treadmill speed (peak TM speed). Gait speed at SSV and FV was collected using the Gait Mat II (Equitest Inc, Chalfont, Pennsylvania). For the 6MWT, timed walking distance over 6 minutes was performed at SSV with concurrent collection of cardiorespiratory measures using a portable indirect calorimetry system (K4b2, CosMed, Chicago, Illinois). Peak TM speed was determined using a modified graded treadmill test with simultaneous cardiorespiratory data collection. During this test, individuals initially walked at 0.1 m/s for 2 minutes with a safety harness, with speed increased every 2 minutes by 0.1 m/s until the individual reached 85% of age-predicted maximum HR,22 experienced significant gait instability, or requested to stop. Peak TM speed was recorded as the highest speed achieved for 1 or more minutes.

Secondary measures included gait economy (ie, O2 cost per unit distance, ml/kg/m) and HR responses during the 6MWT, and rate of oxygen consumption (

O2, ml/kg/min) at peak TM speed and at TM speeds matched to the highest pretesting speed. For all cardiorespiratory measures, data collected during 2-minute baseline resting (sitting) conditions were subtracted from data collected during testing. Evaluation of O2 cost was calculated as the average

O2 over the 6MWT divided by the average speed walked during the 6MWT or until a rest break. The O2 cost was averaged over the entire duration of the 6MWT, as walking speeds may vary during testing and/or the individual may not be at a steady state of oxygen consumption.23 Heart rate responses were recorded every minute and averaged over 6 minutes. Peak metabolic capacity (

O2 peak) was determined as the average

O2 achieved over the last 30 seconds of walking at the highest treadmill speed. Gait efficiency was estimated by comparing

O2 at the highest matched treadmill speeds from pre- to posttesting (

O2 match). Previous studies have demonstrated significant changes in selected metabolic measures following 4 weeks of high-intensity treadmill walking (ie, 4 weeks).5

The goal of each training session was to perform 30 minutes of treadmill stepping and 10 minutes of overground walking at different training intensities, but with equivalent amounts of stepping practice. Amount of stepping activity per session was measured using accelerometers placed at the ankle (StepWatch, Orthocare, Seattle, Washington). Training HRs were monitored continuously using pulse oximetry to maintain the desired HR range as possible. The goal of high-intensity training was to maintain HRs within 70% to 80% HR reserve as determined by the following formula: targeted% HR reserve = rest HR + [(max HR − rest HR) × (targeted %)]. This HR range has been utilized previously during aerobic training interventions provided to individuals poststroke.6,24 Low-intensity training was targeted as 30% to 40% HR reserve, which represents approximate HR ranges during conventional, clinical physical therapy sessions.25 Targeted HR ranges were reduced by 10 beats per minute for those who were prescribed β-blockers. In individuals unable to attain these ranges, the Rating of Perceived Exertion (RPE) scale was utilized, with targeted ratings of 15 to 17 (“hard” to “very hard”) for high-intensity training and 11 to 13 (below “somewhat hard”) for low-intensity training.

At the first training session, participants walked on a motorized treadmill with a safety harness without weight support at gradually increasing speeds until HR achieved 30% to 40% HR reserve (ie, ∼35%) for 5 minutes. This was identified as the treadmill training speed for the study duration (both high- and low-intensity interventions). If participants were randomized to the low-intensity training condition first, they continued to walk at this speed without manipulation of other training variables to the extent possible. For participants randomized to the high-intensity training condition first, the workload as estimated by cardiovascular demands (ie, HR) was manipulated at similar training speeds by applying loads or resistance to the trunk and limbs. Application of selected forces will alter muscle activity patterns and power generation, which alters the cardiovascular demands to supply the muscle with necessary metabolic substrates to sustain locomotion. For example, a weighted vest with added mass equal to 10% body weight was worn to increase the metabolic demands of propulsion and weight support during the stance phase of gait.19 If participants walked with the vest but were below their targeted HR range, a 5-lb ankle weight was applied to the paretic limb to increase the demands of leg swing.18 If HR responses were still below the targeted range, posterior-directed resistive forces (Theraband, Akron, Ohio) were applied to the trunk to increase the demands of forward propulsion,17 with the amount of resistance adjusted as necessary to reach the targeted HR range. If HR responses went over the targeted range, resistance and/or load were removed sequentially in the reverse order. All individuals walked with the weighted vest and leg weight during training, although posteriorly directed forces at the trunk were varied to maintain the targeted HR range.

Following completion of the first intervention and the 4-week washout, individuals completed the second intervention with training speeds constrained as close as possible to those achieved during the first training conditions. In one subject, treadmill speeds during the second, low-intensity intervention had to be reduced by 0.5 m/s due to elevated HRs above the desired training zone.

For overground training, distance was controlled by asking participants to walk for 10 minutes at 30% to 40% of their HR reserve on their first visit and the distance was repeated during all subsequent visits (one subject walked ∼70 m per session higher on average during low-intensity training secondary to a documentation error). If enrolled into the high-intensity training, loads or resistance was applied as necessary on the second visit to maintain the targeted HR, with the same distance repeated every session. Individuals who utilized ankle foot orthoses (n = 9) or used an assistive device overground (n = 5) continued to do so throughout all training sessions.

Stepping parameters collected included the total number of sessions, average duration per session, and average steps per session. Parameters of training intensity included ranges of RPE and HR achieved during training, as documented intermittently every 5 to 10 minutes. The average peak HR per session (reported as % of age-predicted maximum, or % HRmax and as % HR reserve), and average peak RPE per session were used for analysis, as these were generally reflective of the training responses during each session.

For statistical analysis, data were analyzed on-protocol versus intent-to-treat to evaluate the specific effects of manipulating training intensity during locomotor interventions. Only data from those who completed more than 6 sessions of both interventions were included. Data are presented as mean ± standard deviation in the text with standard error in the figures. Parametric distributions of the primary outcome were confirmed using the Shapiro-Wilk test, whereas secondary measures were not normally distributed. Baseline comparisons of primary and secondary outcomes prior to high- versus low-intensity interventions were performed using paired t tests or Wilcoxon signed rank test. Statistical analyses of primary outcomes were performed using a mixed-model ANOVA, with main factors of time (pre- vs posttesting), order (first, second), and training intensity (high, low). We were specifically interested in main effects of time, with significant interactions of time × intensity, time × order, and time × intensity × order. Analysis of secondary outcomes was performed by comparing change scores (posttraining − pretraining data) using Wilcoxon signed rank tests, with specific comparisons for intensity (high vs low intensity) and order (first vs second training bouts). Parameters during training were compared between high- and low-intensity interventions using paired t tests or Wilcoxon signed rank as appropriate. Pearson correlation analyses were performed to evaluate potential relationships between primary outcomes following training versus training responses (RPE, HR). Because of the preliminary nature of this investigation, the α value was set to 0.05 and Bonferroni corrections were not performed to minimize type II errors, although the data are discussed in light of the multiple comparisons.


Fourteen of 17 individuals who were consented for the study fulfilled inclusion criteria and were randomized, although only 12 individuals were able to complete the study. One individual who dropped could not attend further sessions during low-intensity training and another individual's resting and training HRs were elevated well above the targeted ranges during the first 2 sessions; therefore, training was not feasible. No adverse events were noted in any subjects during either training condition. Of the 12 who completed training, 6 had performed another treadmill training intervention more than 3 months prior to enrollment. Demographic and clinical characteristics of those who completed training are provided (Table 1), with no differences between participants who initially received high- versus low-intensity training (all P > 0.10).

Table 1:
Demographics and Baseline Characteristics

During training sessions, TM training speeds were 0.11 ± 0.18 m/s slower than SSV, 0.12 ± 0.21 m/s slower than average speed during 6MWT, and 0.23 ± 0.26 m/s slower than FV, with nearly identical speeds at low- versus high-intensity conditions. Number of training sessions, average duration/session, and average steps/session were not different between high- versus low-intensity conditions (Table 2). During high-intensity training, peak HR per session averaged across all sessions indicated that the target training intensity (ie, within ±5% of the desired HR range) was not met in 6 of 12 participants (4/5 with lower HR responses on β-blockers, 1 higher than target HR range). For low intensity, we were unable to achieve the desired HR range in 5 of 12 participants (3/4 with lower responses on β-blockers, 1 higher than HR range). Nonetheless, significant differences in HR responses were observed between high- versus low-intensity interventions (average differences = 16 ± 24% HRmax or 30 ± 16% HR reserve, both P < 0.0001). Average peak RPE achieved was also significantly different between training conditions (average differences = 5.7 ± 2.0; P < 0.0001).

Table 2:
Training Parameters During High- Versus Low- Intensity Training; P From Paired Comparisons Are Provided

Primary and secondary outcomes following high- and low-intensity training are detailed in Table 3, with no differences in pretraining measures prior to either condition. The mixed-model ANOVA performed for the primary measures revealed a significant main effect of time (pre- vs posttesting) for all outcomes except SSV (P = 0.07). A significant interaction was observed between Time × Intensity for 6MWT (P < 0.01), with nonsignificant Time × Intensity interactions for other variables (closest to significance was peak TM speed at P = 0.08; Figure 2). A significant interaction of Time × Order was not observed for any variable (closest was FV at P = 0.06). A significant 3-way interaction of Time × Intensity × Order was observed for 6MWT (P = 0.02) and peak TM speed (P < 0.01). For these outcomes, post hoc comparisons indicated that the largest changes were observed in participants who initially received high-intensity training. For example, changes in 6MWT averaged 64 ± 57 m following high-intensity training performed first, whereas average changes for other training conditions were < 20 m (high second: 16 ± 17 m; low first: 13 ± 19; low second: 0.3 ± 22 m). Similarly, improvements in peak TM speed were greatest when high-intensity training was performed first (mean increase = 0.23 ± 0.12 m/s), with smaller changes in the other conditions (high second: 0.05 ± 0.08 m/s; low first: 0.13 ± 0.08 m/s; low second: 0.05 ± 0.08 m/s). Individual 6MWT and peak TM speed data at each testing period are provided in Appendix 1.

Figure 2:
Differences in primary locomotor outcomes of (A) 6-Minute Walk Test (6MWT) and (B) peak treadmill speed (peak TM speed); dark lines indicate high-intensity training, dashed lines indicate low intensity, square symbols denote high-intensity training performed first, low intensity second, triangles denote low intensity first, high intensity second
Table 3:
Pretraining and Posttraining Data for High- Versus Low-Intensity Traininga

For secondary measures, there were no differences in O2 cost or average HR responses during 6MWT, nor in

O2 peak and

O2 match during treadmill testing analyzed together or separately for intensity or for order (smallest P = 0.07 for

O2 peak favoring low-intensity training). There also were no differences between change scores of all measures between high- and low-intensity training (smallest P = 0.15 for

O2 match favoring high-intensity training; Table 3). Interestingly, average HR responses during the 6MWT approached or were within the high-intensity training zone for many participants, with an average % HRmax of 63 ± 12% (101 ± 19 beats per minute) across all tests, with no differences before or following training. The HR responses averaged across both pretesting 6MWTs were related to differences from training speeds to SSV, FV, or average speed during the 6MWT (all r > 0.60; P < 0.01).

Further correlation analyses revealed potential relationships between differences in selected walking outcomes with training responses during high- and low-intensity training. Specific differences were calculated as change scores between % HRmax, % HR reserve, and RPE during high- versus low-intensity (eg, average % HRmax during high intensity – average % HRmax during low intensity) versus differences in improvements in 6MWT and peak TM speed following high- and low-intensity training (eg, Δ6MWT following high intensity − Δ6MWT following low intensity). Significant moderate correlations were observed between differences in Δ6MWT following high- versus low-intensity training and differences in % HRmax (r = 0.60, P = 0.04; Figure 3A) and % HR reserve (r = 0.62, P = 0.02). There was no significant correlation between differences in Δ6MWT and RPE (r = 0.52, P = 0.08; Figure 3B). Moderate but nonsignificant correlations were observed between differences in Δ peak TM speed and parameters of training intensity (ie, both % HRmax and % HR reserve: r = 0.54, P = 0.07; RPE: r = 0.48, P = 0.11).

Figure 3:
Correlations between differences in changes in 6MWT during high- versus low-intensity training versus (A) differences in average peak % HRmax and (B) differences in average peak Rating of Perceived Exertion during high- and low-intensity training.


The primary findings of this preliminary study suggest that high-intensity locomotor training may improve selected walking outcomes to a greater extent than low-intensity training in individuals poststroke. Significant differences in 6MWT were observed despite equivalent amounts of stepping practice during training sessions. Furthermore, a significant Time × Intensity × Order interaction was observed for both 6MWT and peak TM speed, with the largest improvements in individuals who initially received high-intensity training. Specific changes in 6MWT were above small minimally clinically important differences following high-intensity training across all participants, but above substantial minimally clinically important difference only for those who received high-intensity training first.26 Moderate correlations between differences in Δ6MWT and differences in training intensity (HR) during each training condition were also significant.

Limitations of this preliminary investigation include the lack of blinded assessors, small sample size, and the use of multiple comparisons. Utilizing this strategy to manipulate training intensity, we maintained α = 0.05 to minimize the potential for type II errors (ie, number of false negatives), with no previous data to estimate sample and effect sizes for the primary outcomes. Calculating multiple comparisons may, of course, introduce significant findings by chance and increase the probability of committing type I errors. Nonetheless, the data described here provide a preliminary estimate of the comparative efficacy of high- versus low-intensity interventions.

The findings of significant contributions of order of training interventions deserve further discussion and suggest that testing or carry-over effects could contribute to the present findings. For testing effects, 50% of the training population had been exposed previously to the testing conditions, such that the expected effect of repeated testing might be small. Furthermore, order was only a significant factor as an interaction with intensity and time. As such, a carry-over effect may explain the larger initial changes in locomotor outcomes, as participants averaged approximately 2500 steps per session in either training intervention. These large doses of stepping training represent greater levels of daily physical activity in participants with chronic stroke with similar gait speeds5 and may contribute to these observed order effects, although only in the high-intensity condition.

Despite significant changes in peak TM speed and 6MWT, there were no significant contributions of order or intensity on SSV and FV. These findings were not surprising considering previous data from other studies detailing the efficacy of high-intensity aerobic treadmill training versus control interventions.5,10–12 The limited gains in SSV and FV may also be due to the lack of focus on faster walking practice, as training speeds were controlled to minimize differences in amount of stepping practice and differences in neural activation strategies associated with altered spatiotemporal parameters at differing velocities.27 Specifically, training at higher speeds but for shorter durations may have resulted in equivalent amounts of stepping practice and distances as compared with training at lower speeds for longer durations. However, faster training could result in altered spatiotemporal patterns that could influence gait performance,28,29 as demonstrated previously.20,21 Of course, application of loads and resistance during walking also alters neural activation strategies,17,18 and their contribution to the present data is unclear. Specifically, the added masses and forces applied to the trunk and legs during training to increase neuromuscular and cardiovascular demands can also elicit adaptive strategies utilized by the nervous system to improve walking function. Separating these mechanisms will be difficult if not impossible, as any perturbation to alter neuromuscular demands should also alter cardiovascular demands during continuous locomotor tasks.

Potential mechanisms underlying the changes in locomotor performance following high-intensity training are not entirely clear and may be multifaceted. Previous data suggest that high-intensity training in individuals poststroke can alter both cardiovascular and metabolic responses to enhance locomotor performance, although changes are often inconsistent. Namely, improved gait efficiency may contribute to improvements in O2 cost but limit improvements in

O2 peak, as demonstrated following a previous 4-week training paradigm.5 This phenomenon appears to contribute here, where

O2 peak following high-intensity training did not improve, possibly due in part to nonsignificant decreases in gait efficiency (lower

O2 at matched TM speed).

Beyond metabolic responses, greater walking improvements following high- versus low-intensity training may be due to neural factors, although their contributions are not often discussed in the rehabilitation literature. For example, previous data have demonstrated improved activation of subcortical and cortical networks following high-intensity stepping training in individuals poststroke.11,30 Such plastic changes are likely driven by the elevated neural output necessary to sustain muscular activity required during high-intensity locomotor training. Increased neural activity could promote short-term changes in functional synaptic connectivity (ie, long-term potentiation31) and contribute to sustained changes through activity-dependent production and release of neural growth factors (ie, neurotrophins) to support synaptic and neural functions.32,33 There is, for example, a large body of literature in healthy humans that suggests that exercise intensity strongly influences expression of specific growth factors (eg, brain-derived neurotrophic factors; see reviews34,35). Whether training intensity enhances growth factor production, which in turn contributes to sustained improvements in locomotor performance in humans with neurologic injury, is not clear, but is a potentially fruitful area of research.

While further research is warranted, this preliminary investigation suggests that the intensity of locomotor interventions may contribute significantly to locomotor improvements. Importantly, the amount and specificity of stepping practice were very similar between training conditions, although this strategy may have also limited the magnitude of differences between conditions. In the clinical setting, selective manipulation of only 1 of these training parameters is unnecessary, whereas consideration of all these factors (ie, specificity, amount, and intensity) may be important to maximize locomotor outcomes. Specifically, training intensity can be manipulated by increasing treadmill or overground walking speeds; such training can augment cardiovascular and neuromuscular demands and increase the amount of task-specific practice performed within training sessions. To further challenge patients to achieve higher training intensities, patients can perform walking activities at progressively increasing treadmill speeds or with applied loads/forces to maximize the benefits of the interventions. Further research should be directed toward understanding which dosage parameters and their progression influence locomotor recovery in this and other patient populations.


This study details one of the first attempts to delineate the effects of intensity versus amount of locomotor practice, revealing significantly greater improvements following high- versus low-intensity training in individuals poststroke in selected measures of locomotor function. In combination with previous data, providing high-intensity stepping practice in the clinical setting can facilitate gains in specific motor outcomes but should be considered as one of several variables related to dosage that may influence walking function.


Funding for this study was provided by NIDRR: H133B031127 to TGH and a James Brown IV Fellowship for RIC Allied Health Staff to KS.


1. Cha J, Heng C, Reinkensmeyer DJ, Roy RR, Edgerton VR, De Leon RD. Locomotor ability in spinal rats is dependent on the amount of activity imposed on the hindlimbs during treadmill training. J Neurotrauma. 2007;24:1000–1012.
2. De Leon RD, Hodgson JA, Roy RR, Edgerton VR. Locomotor capacity attributable to step training versus spontaneous recovery after spinalization in adult cats. J Neurophysiol. 1998;79:1329–1340.
3. Hesse S, Bertelt C, Jahnke MT, et al. Treadmill training with partial body weight support compared with physiotherapy in nonambulatory hemiparetic patients. Stroke. 1995;26:976–981.
4. Wernig A, Muller S, Nanassy A, Cagol E. Laufband therapy based on “rules of spinal locomotion” is effective in spinal cord injured persons. Eur J Neurosci. 1995;7:823–829.
5. Moore JL, Roth EJ, Killian C, Hornby TG. Locomotor training improves daily stepping activity and gait efficiency in individuals poststroke who have reached a “plateau” in recovery. Stroke. 2010;41:129–135.
6. Holleran CL, Straube DS, Kinnaird CR, Leddy AL, Hennessy PW, Hornby TG. Feasibility and potential efficacy of high intensity stepping training in variable contexts in subacute and chronic stroke. Neurorehabil Neural Repair. 2013.
7. Duncan PW, Sullivan KJ, Behrman AL, et al. Protocol for the Locomotor Experience Applied Post-stroke (LEAPS) trial: a randomized controlled trial. BMC Neurol. 2007;28(7):643–51.
8. Duncan PW, Sullivan KJ, Behrman AL, et al. Body-weight–supported treadmill rehabilitation after stroke. N Engl J Med. 2011;364:2026–2036.
9. Katch VL, McArdle WD, Katch FI. Essentials of Exercise Physiology. Baltimore, MD: Lippincott Williams & Williams; 2011.
10. Globas C, Becker C, Cerny J, et al. Chronic stroke survivors benefit from high-intensity aerobic treadmill exercise: a randomized control trial. Neurorehabil Neural Repair. 2012;26:85–95.
11. Luft AR, Macko RF, Forrester LW, et al. Treadmill exercise activates subcortical neural networks and improves walking after stroke: a randomized controlled trial. Stroke. 2008;39:3341–3350.
12. Macko RF, Ivey FM, Forrester LW, et al. Treadmill exercise rehabilitation improves ambulatory function and cardiovascular fitness in patients with chronic stroke: a randomized, controlled trial. Stroke. 2005;36:2206–2211.
13. Hidler J, Nichols D, Pelliccio M, et al. Multicenter randomized clinical trial evaluating the effectiveness of the Lokomat in subacute stroke. Neurorehabil Neural Repair. 2009;23:5–13.
14. Hornby TG, Campbell DD, Kahn JH, Demott T, Moore JL, Roth HR. Enhanced gait-related improvements after therapist- versus robotic-assisted locomotor training in subjects with chronic stroke: a randomized controlled study. Stroke. 2008;39:1786–1792.
15. Israel JF, Campbell DD, Kahn JH, Hornby TG. Metabolic costs and muscle activity patterns during robotic- and therapist-assisted treadmill walking in individuals with incomplete spinal cord injury. Phys Ther. 2006;86:1466–1478.
16. Hornby TG, Kinnaird CR, Holleran CL, Rafferty MR, Rodriguez KS, Cain JB. Kinematic, muscular, and metabolic responses during exoskeletal-, elliptical-, or therapist-assisted stepping in people with incomplete spinal cord injury. Phys Ther. 2012;92:1278–1291.
17. Gottschall J, Kram R. Energy cost and muscular activity required for propulsion during walking. J Appl Physiol. 2003;94:1766–1772.
18. Gottschall J, Kram R. Energy cost and muscular activity required for leg swing during walking. J Appl Physiol. 2005;99:23–30.
19. Grabowski A, Farley CT, Kram R. Independent metabolic costs of supporting body weight and accelerating body mass during walking. J Appl Physiol. 2005;98:579–583.
20. Sullivan K, Knowlton BJ, Dobkin BH. Step training with body weight support: effect of treadmill speed and practice paradigms on poststroke locomotor recovery. Arch Phys Med Rehabil. 2002;83:683–691.
21. Pohl M, Mehrholz J, Ritschel C, Ruckriem S. Speed-dependent treadmill training in ambulatory hemiparetic stroke patients: a randomized controlled trial. Stroke. 2002;33:553–558.
22. Tanaka H, Monahan KD, Seals DR. Age-predicted maximal heart rate revisited. J Am Coll Cardiol. 2001;37:153–156.
23. Eng JJ, Chu KS, Dawson AS, Kim CM, Hepburn KE. Functional walk tests in individuals with stroke: relation to perceived exertion and myocardial exertion. Stroke. 2002;33:756–761.
24. Pang MY, Eng JJ, Dawson AS, McKay HA, Harris JE. A community-based fitness and mobility exercise program for older adults with chronic stroke: a randomized, controlled trial. J Am Geriatr Soc. 2005;53:1667–1674.
25. MacKay-Lyons MJ, Makrides L. Cardiovascular stress during a contemporary stroke rehabilitation program: is the intensity adequate to induce a training effect? Arch Phys Med Rehabil. 2002;83:1378–1383.
26. Perera S, Mody SH, Woodman RC, Studenski SA. Meaningful change and responsiveness in common physical performance measures in older adults. J Am Geriatr Soc. 2006;54:743–749.
27. Hesse S, Konrad M, Uhlenbrock D. Treadmill walking with partial body weight support versus floor walking in hemiparetic subjects. Arch Phys Med Rehabil. 1999;80:421–427.
28. Silver KH, Macko RF, Forrester LW, Goldberg AP, Smith GV. Effects of aerobic treadmill training on gait velocity, cadence, and gait symmetry in chronic hemiparetic stroke: a preliminary report. Neurorehabil Neural Repair. 2000;14:65–71.
29. Vasudevan EV, Bastian AJ. Split-belt treadmill adaptation shows different functional networks for fast and slow human walking. J Neurophysiol. 2010;103:183–191.
30. Enzinger C, Dawes H, Johansen-Berg H, et al. Brain activity changes associated with treadmill training after stroke. Stroke. 2009;40:2460–2467.
31. Nudo RJ, Plautz EJ, Frost SB. Role of adaptive plasticity in recovery of function after damage to motor cortex. Muscle Nerve. 2001;24:1000–1019.
32. Mendell LM, Munson JB, Arvanian VL. Neurotrophins and synaptic plasticity in the mammalian spinal cord. J Physiol. 2001;533:91–97.
33. Arvanian V. Role of neurotrophins in spinal plasticity and locomotion. Curr Pharm Des. 2013;19:4509–4516.
34. Huang T, Larsen KT, Ried-Larsen M, Moller NC, Andersen LB. The effects of physical activity and exercise on brain-derived neurotrophic factor in healthy humans: a review. Scand J Med Sci Sports. 2014;24:1–10.
35. Knaepen K, Goekint M, Heyman EM, Meeusen R. Neuroplasticity–exercise–induced response of peripheral brain-derived neurotrophic factor: a systematic review of experimental studies in human subjects. Sports Med. 2010;40:765–801.
Appendix 1:
Individual Pre- and Posttraining Measures for 6MWT and Peak TM Speeds Appendix

aerobic training; stepping; rehabilitation

Supplemental Digital Content

© 2015 Neurology Section, APTA