The Use of Visual Feedback to Improve Temporal Gait Asymmetry Post-Stroke

by

Jessica Noelle Powers

A thesis submitted in conformity with the requirements for the degree of Master of Science

Rehabilitation Sciences Institute University of Toronto

© Copyright by Jessica Noelle Powers 2018

ii

The Use of Visual Feedback to Improve Temporal Gait Asymmetry Post-

Stroke

Jessica Noelle Powers

Master of Science

Rehabilitation Sciences Institute University of Toronto

2018

Abstract

Temporal gait asymmetry (TGA) is common post-stroke and particularly resistant to

conventional therapy. Therefore new rehabilitation approaches are required. One approach

may be visual feedback (FB) during gait training. Optimal parameters for providing visual FB

during motor skill acquisition are well established for healthy adults but remain largely

unknown for motor learning in people with stroke. This study aimed to determine the effect of

visual displays and frequencies for visual FB about TGA during an overground walking practice

session and its effects during a retention session. Participants received feedback at 50% or

100% frequency in one of two display formats (A and B). The largest effect sizes were seen in

the group that received Display B at 100%, and 50%. Individuals that received FB during practice

showed significant changes in TGA at the retention while those individuals who did not receive

feedback showed no changes at retention.

iii

Acknowledgments

First, I would like to thank my thesis supervisor, Dr. Kara Patterson, at the Rehabilitation

Sciences Institute at the University of Toronto. Her encouragement, understanding and

seemingly limitless patience with my learning curve made this journey possible. Thank you to

my co-supervisor, Dr. Dina Brooks, for sharing her optimism and experience. Thank you both for

your support in perusing a career in physiotherapy and for helping make it a reality. Thank you

to my committee members, Dr. Avril Mansfield and Dr. George Mochizuki, for sharing their

unique insights and pushing me to broaden my knowledge. I am also grateful to my lab mates

who were always willing to lend a hand and share an idea.

To my friends and family, to whom I cannot express enough gratitude. For celebrating my

successes, no matter how small, and giving me the courage to tackle daunting challenges. For

lending an ear and talking it out. For finding me when I was lost and reminding me where I was

heading. I am especially grateful to my parents, for their unfaltering confidence in me.

Finally, a special thanks to the participants who so graciously volunteered their time to my

research and whom this work would not be possible without.

iv

Table of Contents

Contents

Acknowledgments........................................................................................................................... iii

Table of Contents ............................................................................................................................ iv

List of Tables ................................................................................................................................... vi

List of Figures ................................................................................................................................. vii

List of Appendices ......................................................................................................................... viii

List of Abbreviations ....................................................................................................................... ix

Overview .....................................................................................................................................1

Literature Review ........................................................................................................................3

2.1 Stroke Overview ...................................................................................................................3

2.1.1 Stroke in Canada ......................................................................................................3

2.2 Course of Recovery ..............................................................................................................4

2.2.1 General Course of Recovery ....................................................................................4

2.2.2 Recovery of Gait .......................................................................................................4

2.3 Gait .......................................................................................................................................5

2.3.1 Normal Gait ..............................................................................................................5

2.3.2 Neural Control of Walking .......................................................................................7

2.3.3 Common Gait Deviations Post-Stroke .....................................................................8

2.4 Motor Learning ................................................................................................................. 17

2.4.1 Motor Learning in Healthy Adults ......................................................................... 20

2.4.2 Motor Learning in Adults with Stroke ................................................................... 25

2.5 Summary ........................................................................................................................... 30

2.6 Study Objectives ............................................................................................................... 31

v

Study: Use of Visual Feedback to Improve Temporal Gait Asymmetry Post-Stroke ............... 32

3.1 Introduction ...................................................................................................................... 32

3.2 Methods ............................................................................................................................ 35

3.2.1 Design .................................................................................................................... 35

3.2.2 Participants ........................................................................................................... 35

3.2.3 Testing Protocol .................................................................................................... 36

3.2.4 Baseline Measurements ....................................................................................... 36

3.2.5 Gait measures ....................................................................................................... 37

3.2.6 Overground walking training with feedback ........................................................ 37

3.2.7 Visual Feedback about TGA .................................................................................. 38

3.2.8 Data Processing ..................................................................................................... 41

3.2.9 Data Analysis ......................................................................................................... 41

3.3 Results ............................................................................................................................... 44

3.4 Discussion.......................................................................................................................... 51

3.5 Conclusion ......................................................................................................................... 55

General Discussion ................................................................................................................... 56

4.1 Limitations......................................................................................................................... 62

4.2 Future directions ............................................................................................................... 64

References .................................................................................................................................... 65

Appendix A .................................................................................................................................... 83

vi

List of Tables

Table 3.1 Participant characteristics ………………………………………………………………………………………. 53

Table 3.2 TGA outcomes at retention ………………………………………………………………………………………57

Table 3.3 Spatiotemporal gait parameters of individuals who showed change in TGA at

retention gait .............................................................................................................................. 58

Table 3.4 Effect size of feedback conditions …………………..……………………………………………………… 59

vii

List of Figures

Figure 3.1 Phases of the gait cycle …………………………………………………………………………………………. 17

Figure 3.2 Visual feedback display A ………………………………………………………………………………………. 49

Figure 3.3 Visual feedback display B ………………………………………………………………………………………. 49

Figures 3.4 Single subject analysis of motor learning curves ………………………………………………….. 55

viii

List of Appendices

Participant instructions for visual feedback display A and B

ix

List of Abbreviations

ADL Activities of Daily Living

CMSA Chedoke McMaster Stroke Assessment

CPG Central Pattern Generator

DLS Double Limb Support

EMG Electromyography

GC Gait Cycle

GRF Ground Reaction Force

HRQoL Health Related Quality of Life

HS Heel Strike

IC Initial Contact

IQR Interquartile Range

KP Knowledge of Performance

KP Knowledge of Results

LE Lower Extremity

MoCA Montreal Cognitive Assessment

NDT Neurodevelopmental Technique

NIHSS National Institute of Health Stroke Scale

RAS Rhythmic Auditory Stimulation

SD Standard Deviation

x

SLS Single Limb Support

TGA Temporal Gait Asymmetry

TO Toe Off

1

Overview

Mobility challenges are common after stroke, with almost 80% of people still experiencing

impaired walking function 3 months post-stroke1. Recovery of walking function is one of the

most commonly stated goals of patients2, however 22% of those who experience walking

impairments do not regain any walking function by the end of rehabilitation3. Many of those

that do regain walking function have continued gait related impairments or deviations such as

decreased velocity4,5, difficulty with balance6, increased energy cost7 and spatiotemporal gait

asymmetry 8. These persistent gait impairments can contribute to a decline in mobility9,10,

increased dependence and reduced health related quality of life (HRQoL)11.

Rehabilitation is often discontinued for patients when they no longer exhibit motor

improvements, or “plateau”, and further improvements are deemed unlikley12. However,

improvements seen in chronic stroke patients with the use of novel interventions13–17 rebuts

the supposed motor recovery plateau. It has been suggested that the plateau is not indicative

of limited capacity for recovery, but rather a neuromuscular adaptation to motor

interventions18. New interventions are required to address persistent gait impairments and

enable patients to overcome the motor recovery plateau. One approach may be to apply the

principles of motor learning to post-stroke gait rehabilitation.

Principles of the use of augmented visual feedback to facilitate motor skill learning were

applied in this study. Feedback is often provided during rehabilitation therapy, however, not in

2

the most effective format19. The purpose of this study was to investigate the effects of visual

feedback on post-stroke gait, specifically, temporal gait asymmetry (TGA). This study employed

a motor learning experimental paradigm that included a single acquisition session where visual

feedback was given during overground walking practice and a retention session 24 hours later

where TGA was assessed during overground walking without feedback.

3

Literature Review

2.1 Stroke Overview

2.1.1 Stroke in Canada

Stroke is the leading cause of disability in Canada, with 50,000 new strokes each year20. There

are 412,000 survivors living with the effects of stroke, and this is expected to rise to between

666,000 and 738,000 by the year 203821 due to Canada’s growing and aging population. Every

individual experiences the effects of stroke differently, which can manifest both physically and

cognitively. Some physical effects of stroke include muscle weakness, spasticity, and loss or

change in sensation causing challenges with movement, mobility, and balance22. Cognitive

effects of stroke can include difficulty with thinking, memory and emotions23,24. Rehabilitation

after stroke is aimed at mitigating physical and cognitive impairments, and increasing

independence and health related quality of life (HRQoL)25.

Despite rehabilitation efforts, many survivors of stroke never fully recover. Approximately 36%

experience significant disability related to stroke impairments that persists 5 years post-stroke26

and more than 40% require ongoing assistance with activities of daily living (ADLs)27. With the

estimated number of individuals experiencing the effects of stroke in Canada expected to rise

substantially over the next years these outcomes are of increasing concern.

4

2.2 Course of Recovery

2.2.1 General Course of Recovery

Recovery after stroke is the re-emergence of structures and functions as they existed previous

to neurologic injury. Recovery is highly individual and is related to a variety of factors such as

site and size of the lesion in the brain, severity of initial motor impairment, and age28–33.

Recovery is generally rapid during the sub-acute phase and tapers to a plateau in the chronic

phase34–38. The largest improvements are seen during the first 3 months following the stroke.

These rapid improvements are seen in all areas including upper extremity function, walking and

ability to perform ADLs39–41. These improvements continue up to 6 months although at a less

significant rate. At around six months, recovery begins to plateau with little if any

improvements seen, especially after one year3,37–39. However, there is evidence from numerous

studies of improved function in the chronic phase of stroke recovery42–45 and the capacity for

continued neuroplasticity has been demonstrated46. There are a variety of interventions

administered at least one year post-onset that have improved daily functioning 47–49, and

motor recovery of the upper and lower extremities (LE)13–16,18.

2.2.2 Recovery of Gait

Impaired walking function is common after stroke, with half of survivors having no walking

function immediately following stroke3and almost 80% still experiencing walking dysfunction

after the first 3 months1. Improvement in walking, as measured by the Barthel index for

5

walking, occurs in the first 11 weeks3, with little change seen after 6 months of onset3,39,41,50.

The speed of gait recovery can be linked to the initial degree of paresis and walking ability (e.g.

independence, velocity, endurance), with faster recovery seen in those with milder paresis and

greater walking ability in the first week post-stroke3,10,51,52. Of those whose walking function

was affected at onset, 64% will regain walking independence3,53, 14% will be able to walk with

assistance, and 22% will not recover walking function3 by the end of rehabilitation. It is

important to note that the studies reporting these outcomes used functional scales, which

measure level of assistance required while walking (e.g. dependent versus independent), and

do not consider other features of gait impacted by stroke, such as velocity, endurance,

symmetry and loading. While it is believed that very little improvement is seen after 6 months,

there is increasing evidence for improving gait velocity and endurance at the chronic stage9,54,55.

2.3 Gait

2.3.1 Normal Gait

Gait is a cyclical and complex sequence of movements that carries an individual from one point

to another by propelling the body forward, while maintaining upright stability 56. Human

locomotion is dependent on stance stability despite constantly changing posture57, propulsion

through muscle and joint activation58,59, shock absorption to minimize impact60, and energy

conservation61, for performance efficiency. These four variables enable locomotion to occur

through alternation of body weight support and limb propulsion between the limbs. A single

6

set of support and propulsion actions of one limb is called a gait cycle(GC)56.

A GC can then be further divided into phases: stance and swing. A GC can be analyzed spatially

or temporally. Spatial characteristics of gait include step length (i.e., the distance traveled by

one limb in a step) and step width (i.e., distance between two feet in a step), but for the

purposes of this thesis, the primary focus will be on temporal factors. Stance is the period in

which one limb maintains contact with the ground to support the body, while swing is the

period in which one limb moves through space in a forward motion to advance position. Swing

begins when the foot is lifted and the toes no longer have any contact with the ground,

commonly referred to as toe-off (TO), and ends when the same foot makes contact again with

the heel, commonly referred to as heel strike (HS) or initial contact (IC). The distance between

HS of one foot to HS of the opposite foot is the step length, while stride length is the distance

from HS of one foot to the next consecutive HS of the same foot. Stance begins at IC and ends

at TO, during which ground contact is maintained56.

The stance phase can be further divided into three parts. At the time of HS, the opposite limb

still has contact with the ground to facilitate the transfer of weight to the other limb. This is a

period where the body is supported by both limbs called initial double limb support (DLS). Once

support has been transferred to the limb that has just initiated floor contact, the opposite limb

enters swing phase, and no longer has contact with the floor. At this point, the body is being

supported by only one limb, and is called single limb support (SLS). When the opposite limb

completes swing phase and HS occurs, this marks the end of SLS and the beginning of DLS56.

7

Figure 3.1 Phases of the gait cycle. Adapted from Bajawa ZH, Wootton RJ, Warfield CA.

Principles and Practice of Pain Medicine, 3rd Edition. (McGraw Hill Education, 2016).62

2.3.2 Neural Control of Walking

The underlying processes controlling human locomotion are believed to be a combination of

central pattern generators (CPG), descending neural pathways and peripheral sensory

feedback63,64. However, much of this knowledge was ascertained from the study of animal

models, leaving the exact nature of human locomotor control debated64–67. The CPG is a

neuronal circuit within the spinal cord that operates independently of descending neural input

able to produce the rhythmical motor patterns through flexor and extensor motor neuron

activation69. While the CPG is able to produce the rhythmical basis of locomotion, supraspinal

and peripheral sensory information is required for creating a flexible motor pattern63,64, capable

8

of spatial and temporal adaptation to extrinsic environmental factors. It is believed that the

importance and role of this descending and peripheral information for the control of

locomotion is greater in humans than the animal models of CPG frequently studied64.

Neural control of walking can be impacted by lesions to various structures of the central

nervous system that disrupt the automaticity and patterning of gait. Injury to the motor cortex,

and descending motor pathways can result in paresis and hypertonicity70, while injury to

somatosensory cortex can interfere with sensory feedback71. These neurologic injures lead to

the gait deviations observed in post-stroke gait.

2.3.3 Common Gait Deviations Post-Stroke

Factors that contribute to gait deviations post-stroke have been characterized using

electromyography (EMG), kinetic, kinematic, and spatiotemporal measures. The most widely

reported spatiotemporal measure of gait after stroke is velocity, which is typically decreased

compared to neurotypical gait72–75. This slower gait is also characterized by decreased cadence

and longer gait cycle74–77. Other spatiotemporal characteristics of gait such as step length, swing

and stance time, and double support time may change after stroke, and can exhibit interlimb

differences77.

The earliest EMG studies of post-stroke gait identified atypical motor performance of the

paretic and nonparetic side, with marked decrease in EMG activation on the paretic side,

increased activation on the nonparetic side, and atypical timing of EMG onset78–80. As is the

case with many characteristics of stroke, there is great interindividual variation in muscle

9

activation patterns, although there are some patterns that are exhibited more often74,80,81.

Knutsson and Richards80 attempted to identify common EMG patterns across individuals,

resulting in three classifications. Type I (spastic gait) demonstrates an early activation of the calf

muscles of the paretic limb during stance phase due to hyperactive stretch reflexes as the knee

is extended at initial contact. Type II (paretic gait) is evidenced by a decrease in muscle

activation of two or more muscle groups of the paretic limb. Type III is characterized by atypical

co-activation of muscle groups, but not an increase or decrease in muscle activity. While these

classifications are limited in profiling the diversity of post-stroke gait82, the general pattern

holds that atypical amplitude and timing are common characteristics of post-stroke EMG

activity74,81,83–85.

Post-stroke gait also exhibits kinetic and kinematic abnormalities, compared to that of a healthy

individual. Slower gait in healthy individuals is marked by smaller joint excursions and force

production86,87 and similar deviations are seen in the typically reduced speed of post-stroke gait

88–90. The most common kinematic deviations of the paretic limb are increased ankle plantar

flexion at initial contact, knee hyperextension at weight acceptance and stance phase,

decreased knee flexion and dorsiflexion through swing phase, hip hiking or circumduction, and

increased external rotation of the hip and ankle74,83,89,91. Furthermore, ground reaction force

(GRF) is significantly reduced in the paretic limb compared to the nonparetic limb92,93, where

the reduced propulsive force of the paretic limb translates to smaller steps lengths of the

nonparetic limb94. While the deviations described are common in post-stroke gait, individuals

may exhibit some or none of these deviations, due to interindividual differences.

10

2.3.3.1 Temporal Gait Asymmetry

Asymmetry is a prominent feature of post-stroke gait. Theoretically, asymmetry can be

expressed using almost any bilateral variable of interest including kinematic variables (e.g. peak

knee flexion angle of the paretic and nonparetic limbs), kinetic variables (e.g. GRF of the paretic

and nonparetic limb during stance) and spatiotemporal variables (e.g. step length). Central to

this thesis is temporal gait asymmetry (TGA).

Gait asymmetry has been described as a ‘compensation’ for the motor impairment that enables

gait function95,96. Compensation is considered the emergence of new motor patterns of end

effectors (i.e., body parts such as the leg and foot) to carry out functions executed by different

motor patterns previous to stroke97 Motor impairment of the LE is associated with TGA6,92

suggesting that at least in some cases it is a compensation adopted by the individual in order to

have some walking function. However, motor impairment it does not explain all the variance

observed in TGA post-stroke and there are examples of the presence of TGA in individuals with

higher level of motor recovery. Thus, it is plausible that factors other than motor impairment

contribute to the TGA and this may be particularly true in the specific case of asymmetrical

walkers with milder motor impairment 5. For example, TGA may be directly caused by brain

lesions that impair the ability to produce regular and rhythmic timing of gait events. Therefore,

while TGA may be a compensation for motor impairment of the affected leg in some

individuals, in others it may be a direct result of impairments to the systems involved in

locomotor pattern generation. Therefore, for the purposes of this thesis, TGA will be referred to

11

as a g ‘deviation’, rather than a ‘compensation’, since TGA is a deviation from a typical gait

pattern, that may be a result of compensation for impairment or of impairments directly.

Temporal asymmetry, the specific focus of this thesis, is characterized by unequal amounts of

time spent in specific phases of gait (e.g. swing time, stance time) between right and left limbs.

Temporal asymmetry can be quantified by ratios of the left and right limb values. There are also

established cutoffs to determine whether a ratio value represents asymmetric gait as compared

to healthy older adults98. For example, a swing time ratio (larger value/smaller value) of greater

than 1.06 is considered temporally asymmetric98. Other common temporal deviations include

greater stance time and SLS of the nonparetic limb, and increased overall DLS with terminal DLS

of the paretic limb being greater74–77. One potential explanation for this temporal asymmetry is

motor impairment of the affected limb that leads to difficulty executing swing and,

consequently, increased amount of time spent in the swing phase77,92. Other studies have found

that spasticity of the affected ankle plantarflexors is a primary determinant of TGA, suggesting

that greater hip and knee flexor movements are used during the swing phase to accomplish

foot clearance, consequently increasing the duration of swing of the paretic limb92,99. Another

possible factor causing temporal asymmetry is sensory impairment of the affected limb, which

suggests that better sensation leads to increased confidence in landing the affected foot

terminating swing phase, and therefore shorter swing phase of the affected limb, compared to

those who have more severe sensory impairments92.

12

2.3.3.2 Consequences of Temporal Gait Asymmetry

Temporal gait asymmetries can lead to secondary consequences that can impede mobility and

decrease independence. One consequence may be musculoskeletal injury of the nonparetic

lower limb. The high ground reaction forces repetitively applied through the unaffected limb

has been positively correlated with TGA 68. The nonparetic limb also spends more time in SLS

and less time in swing than the paretic limb, consequently, more time is spent weightbearing

on the nonparetic limb 68,75,100. One possible explanation for this could be a reluctance to bear

weight on the paretic limb, leading to the increased swing of the nonparetic limb and delayed

initiation of foot contact, to reduce the amount of time spent on the paretic limb in SLS,

subsequently, more time is spent in stance on the nonparetic limb than the paretic limb. This

type of repetitive and prolonged loading associated with asymmetric gait has been linked to

joint pain and degeneration. Findings in the unilateral amputee population, who also show

asymmetrical gait by nature of their prosthetic limb, show that spending more time on the

intact limb than the prosthetic limb subjects the intact limb to greater stress that may lead to

joint pain and degeneration of the limb 101,102. Improved gait symmetry may reduce repetitive

loading in the unaffected limb, and decrease the risk of developing musculoskeletal injuries.

Temporal gait asymmetry also has repercussions for the affected limb. Decreased bone

density, or osteoporosis, may occur after prolonged non-use of the affected limb103. Load

bearing of the limb helps to maintain bone density and bone health. The reduced muscle

activity and decreased weight bearing activity of the affected limb leads to loss of bone mineral

density and increased risk of injury 8,103–105 with the risk of hip/femoral neck fracture increasing

13

two fold after stroke106. Reducing immobility of the affected limb and improving symmetrical

weight-bearing during walking could reduce bone mineral density loss and reduce the risk of

injury103,107,108.

Increased energy expenditure, despite decreased capacity for ambulation7,109, is another

consequence of post-stroke gait asymmetry. A study examining the metabolic costs of walking

in healthy individuals indicates that, compared to symmetrical walking, there is a substantial

increase in energy costs when asymmetrical walking is induced using a dual-belt treadmill109.

Similarly, another study on energy expenditure during walking showed that, compared to

healthy controls where gait is automatic, individuals with stroke had a higher oxygen uptake

during walking and therefore a greater demand for energy during asymmetric gait7. This

discrepancy in energy expenditure indicates that the altered gait pattern as a result of a stroke

compromises the efficiency of ambulation 7. Furthermore, a study in the amputee population

with asymmetric gait showed that increased gait symmetry resulted in decreased energy

demands 110. It is reasonable to apply this same concept to individuals with asymmetric gait

due to stroke. To reduce energy inefficiencies due to TGA, new methods of improving gait

symmetry are required.

Furthermore, post-stroke gait asymmetries are associated with challenges with dynamic

balance control111. Balance and gait deficits are linked to post-stroke falls112–114 and loss of

balance while walking is reported as the most common activity when the fall occured91.

Consequences of post-stroke falls can include reduced ambulation, increased sedentary

behavior and decreased community participation115,116. To decrease the risk of these secondary

14

consequences, gait rehabilitation focusing on improved symmetry requires more careful

consideration.

2.3.3.3 Treatment of Gait Disorders Post-Stroke

Gait Training

Post-stroke rehabilitation has a large focus on gait training, with 25% to 45% of physical therapy

time spent practicing gait, especially at the very beginning of rehabilitation117. Common

interventions used to improve walking ability post-stroke are Bobath framework,

neurodevelopmental techniques (NDT), strength training, treadmill training, and intensive

mobility training118. The most commonly used measures of functional ambulation are self-

selected walking speed and endurance, measured by the six minute walk test (6MWTor similar

test), therefore the effectiveness of many gait training programs are evaluated based on these

measures. Bobath and the related yet separate NDT approaches are commonly used in stroke

rehabilitation119–121 and while it does not show superior outcomes compared to other

interventions122–124, it has shown improvements in temporal gait parameters such as SLS

time125. Strength training of the lower extremity (LE) is proven to increase muscle strength of

the limb, which has benefits such as improving bone mineral density, but does not have a direct

effect on improved gait42,45,126. A number of meta-analyses examining the effectiveness of

treadmill training (with or without body weight support) largely agree that there is no

significant difference in gait improvements when compared to or combined with other

physiotherapy interventions (e.g. stretching, strengthening, overground gait training)118,127.

15

However, some studies of treadmill training in people with chronic stroke demonstrated

increased gait velocity after training128,129. Finally, there is evidence suggesting that intensive

mobility training (which includes a combination of aerobic training, functional strength training

and dynamic walking tasks) is a superior approach to improving endurance and walking speed

post-stroke and has additional benefits of improving cardiorespiratory fitness, bone density,

balance confidence and strength 118,130–132.

Common gait training practices, show some improvements in gait ability after stroke, though

the target is often increased velocity and endurance123,133. These interventions are largely

ineffective at improving other temporal gait parameters, with the majority of asymmetric

stroke patients showing no improvement in these measures during rehabilitation134,135.

Although some measures of functional ambulation improved with conventional rehabilitation

strategies, given the potential long-term consequences of persistent temporal asymmetry, a

greater focus should be placed on improving this particular post-stroke gait deviation.

Interventions for TGA

Conventional rehabilitation does not result in improved TGA136, though some experimental

interventions have shown promise for improving TGA. One such intervention is split belt

treadmill training 135,137–139, where patients can practice walking with the belts at different

speeds, to facilitate the movement of the paretic limb. While this treatment has shown to be

effective for short-term improvement in spatiotemporal gait parameters (up to 3 months post-

treatment)137, the equipment is not readily accessible in clinical settings due to high costs140.

16

Rhythmic auditory stimulation (RAS) is another therapeutic approach that shows improvement

in gait asymmetries is rhythmic auditory stimulation(RAS)123,141. This treatment shows improved

asymmetry in spatial characteristics like step length, velocity, as well as temporal variables such

as swing time, although not to the same extent as spatial variables123,141. For example, velocity

and step length improved by 164% and 88%, respectively, while swing symmetry improved by

32% six weeks after RAS training. Although it is important to note that the control group, which

received NDT/Bobath therapy, also showed 16% improvement in swing symmetry at six

weeks123. This finding highlights the view that temporal asymmetry can be changed, however

interventions that specifically target this gait parameter may be required.

Advances in treatment of post-stroke asymmetrical gait are evident123,135, however, they have

limitations: feasibility in a clinical setting, lack of effectiveness for altering temporal

characteristics, compared to spatial characteristics and, does not address underlying factors.

New approaches that address these issues are necessary. One option for improving and

creating new interventions for TGA is through the application of motor learning principles,

specifically augmented feedback. Motor learning principles for individuals with stroke may

differ from those established in healthy adults due to altered sensorimotor capabilities71.

Augmented feedback is currently used during rehabilitation, although parameters for its

application (e.g. format and frequency) remain unclear. Optimized application of feedback

principles during post-stroke gait training, targeted at improving TGA, may enhance gait

outcomes and help reduce the risk of long-term negative consequences.

17

2.4 Motor Learning

Motor learning is defined as “a set of processes associated with practice or experience leading

to relatively permanent changes in capability for skilled movement” 142. Practice can be defined

as repeated attempts to produce a movement that is outside an individual’s current

capability143. During practice, processes and conditions are applied to the task and environment

that alter the interaction the individual has with them, facilitating the search for task solution

and leading to acquisition of a skilled action. The term skilled action has come to mean high

levels of effectiveness and efficiency in movement, encompassing accuracy, consistency,

reliability, and automatic economical execution of movement144. In healthy adults, practice

conditions have been established that optimize acquisition and result in the relatively

permanent ability to produce the skilled action142.

There are theories of motor learning to explain the mechanism of skill acquisition. Schmidt’s

Schema Theory is based on closed loop motor control, in which sensory information about an

ongoing movement is compared with the stored memory of the intended movement that is

used as a reference to detect error and adjust movement accordingly63,145. However, this theory

does not account for the production of novel movements or movements that occur in the

absence of sensory feedback63. Newell’s Ecological Theory proposes that motor learning is

based on a search for optimal strategies for task solution during practice, congruous with the

environmental and task restraints. This search for a task solution is facilitated by complex

perception/cognition/action processes resulting from the interaction between the individual

with the task and environment63,146. Newell’s theory accounts for more factors that are

18

essential to motor learning, though it is still fairly new and yet to be systematically investigated

relative to specific motor skill acquisition examples63.

Other theories explain skill acquisition with respect to the process of learning over time. Fitts

and Posner147 propose a three-stage model for motor learning that focuses on the provision of

cognitive resources and degree of attentional demand of producing a skilled movement.

Initially the new task requires a high attentional focus as the task is investigated and various

strategies are attempted to produce the skilled movement. This ‘cognitive stage’ results in high

variability and large improvements in performance as an effective strategy for producing the

movement is developed. Once the best movement strategy has been selected, the movement is

refined for optimal performance, displaying less variability, smaller improvements and reduced

attentional demand. The final stage of motor learning in this theory, the ‘autonomous stage’,

requires a much lower degree of attention as the automaticity of the task increases. During this

stage, more attention is diverted to external factors that may affect performance such as

scanning the environment for features that could affect the task or preforming another task

simultaneously.

Another three-stage model originally proposed by Bernstein148 and adapted by Vereijken,

Newell and colleagues149, the systems theory, suggests that the central component of motor

learning is the control of the degrees of freedom. The theory suggests that learning a motor

skill begins with a reduction in the degrees of freedom to simplify the task and learn the general

movement, called the ‘novice stage’. The ‘advanced stage’ follows in which more degrees of

freedom are released, resulting in more precision and greater ability to respond to task and

19

environmental demands. In the third stage, the ‘expert stage’, all the degrees of freedom have

been released, increasing the efficiency of execution under a variety of demands by modifying

characteristics of the movement.

A two-stage model of motor skill acquisition has been proposed by Gentile150,151, which

describes the goals of the learner at each stage. During the first stage the learner’s goal is to

understand the task, explore movement strategies and discern characteristics of the

environment that could affect the task performance. During the second stage the learner’s

primary goal is to refine the movement, including consistent and efficient task execution, as

well as the ability to adapt movement to meet changing task and environmental demands.

Fitts and Posner’s model of motor learning stages is particularly suited for rehabilitation in

patient populations as therapy begins early in recovery, which may be equated to the cognitive

stage of the model. Employing the principles of motor learning that are ideally suited to this

stage of the learning process may optimize rehabilitation outcomes. The use of augmented

feedback during the cognitive stage, or early in rehabilitation, is a good illustration of such a

principle152. Frequent provision of feedback during the early stages aids the learner in

understanding the task and creates a reference for correct movement patterns. As the learner

approaches the autonomous stage, feedback should be more precise and less frequent so that

skills can be refined without becoming dependent on the extrinsic source of information about

their movement145,151.

20

2.4.1 Motor Learning in Healthy Adults

Motor learning can be measured directly through neural activity during performance (e.g.

electroencephalography153, functional magnetic resonance imaging154), but is often inferred

from indirect measures of performance, such as movement outcome and movement

characteristics. A common experimental paradigm used to evaluate the effectiveness of the

application of various principles applied to practice is the use of a retention design, which

includes an acquisition and retention phase142,155. During the acquisition phase, the skill is

practiced under specific conditions by manipulating one element of practice to varying degrees

in two or more groups (e.g. control group and experimental group). After a set amount of time,

where the skill is not directly practiced under the outlined conditions, the retention phase

occurs. During retention, the skill is performed under the same conditions by both groups, and

performance is measured. Motor learning is said to have occurred if there is improved

performance during the retention phase. The group that performed better during retention

would indicate the optimal condition to apply to practice that lead to more permanent ability

to produce the skilled action.

Elements of practice can be manipulated to aid in the acquisition of a motor skill. These

elements can include level of practice, practice design, and provision of feedback. Level of

practice refers to the amount of practice, which could be measured in time or number of

repetitions, where the amount of practice required is dependent on the stage of the learner63.

It has been shown that larger improvements are seen early on with lots of practice, then

smaller increments of improvement are observed as there is less and less room for

21

improvement156,157. The rate of improvement at any point of practice is linearly related (on a

log scale) to the amount left to improve142.

Repetition of a skill within practice can be organized to be constant or variable, such that the

same skill is practiced repeatedly or variations thereof, are practiced during a session142,145.

Variable practice leads to greater ability to perform the skill outside the range that it was

practiced. This practice designs may seem to make the acquisition of a skill more challenging,

but leads to better retention performance158,159. When repetitions within practice are varied,

they can be scheduled to optimize learning, by performing repetitions in blocked random order;

this is a form of contextual interference142,160. Using a blocked practice design (low contextual

interference), repetitions of one variation of the skill to be learned are performed, before

moving onto another variation, and so on. A random practice design (high contextual

interference) entails variations of the skill randomly distributed throughout the practice

session. Shea and Morgan161 discovered that high contextual interference leads to poor

performance during practice, but leads to better performance during retention tests. Blocked

practice design also leads to poor transfer of performance in novel contexts, whereas the same

context dependency is not as evident when the skill is practiced with high contextual

interference162–164. Therefore, practice design that includes variations of the skill randomly

distributed throughout practice leads to optimal performance in novel contexts when practice

has concluded.

The manner in which the skill is practiced is also a contributing factor in motor skill acquisition.

The skill may be practiced in its entirety, initiation to completion of the task, every repetition.

22

Conversely, a skill can be practiced in parts, wherein the skill is broken down into segments to

be practiced individually, then combined142. This is most effective when a task can be naturally

divided into parts, such as reaching, grasping and drinking from a cup165,166. Other tasks, where

transition between segments is an integral part of its performance, are ideally practiced in

whole rather than in parts167,168. Gait is a prime example of such a task, in that a patient could

practice bearing weight on a single limb to improve stance stability, however the shifting of

weight from the opposite limb and weight acceptance before achieving SLS are integral to

completing a gait cycle.

During training of a skill, instruction about the movement are often provided to guide the

learner’s performance and can be delivered as explicit or implicit information142,160. Where

explicit instructions inform the learner of how to carry out the movement with rules and

knowledge about the movement itself, implicit instructions do not declare such rules, allowing

the learner to make discoveries about the movement outcome independently169. In healthy

learners, copious evidence supports that providing explicit information about the movement,

requiring conscious awareness of performance, actually degrades learning169–171. Conversely,

implicit learning allows the learner to discover correct movement patterns and use sensory

information to guide movement with less attentional demand, rather than focusing on specific

movement goals which may hinder automatic control of movement172. Similarly, attentional

focus (internal versus external) is an aspect of practice that has a similar effect, where external

foci (focus on movement outcome) is superior in facilitating skill acquisition, compared to

internal foci (focus on body movements)173,174. While there are numerous studies proposing

that the optimal focus of attention is dependent on the stage of the learner173,175–177, a review

23

by Wulf178, provides strong evidence that external attentional focus is superior, regardless of

learning stage179–181. It may be surmised that similar to the effects of explicit knowledge,

internal foci demand too much conscious awareness, interfering with the automaticity of the

movement. Gallwey & Kriegel182 provide a rather simplistic yet appropriate explanation for this

effect: “doing rather than thinking about doing”.

Feedback During Motor Learning

Feedback during motor learning is generally classified into two types: intrinsic and extrinsic

feedback142,160,183. Intrinsic feedback, also called task-intrinsic feedback, refers to any source of

information inherent to the movement and environment: visual, auditory, tactile and

proprioception142,184. Extrinsic, or augmented feedback, is often utilized during practice to

enhance the learner’s intrinsic feedback and provide motivation to continue working toward

the movement goal185–187. The broad use of the term augmented feedback can be used to

describe any source of information provided to the learner originating from an external source

and will be referred to in the remainder of this section, as simply feedback. While feedback is

not always necessary and can even hinder motor skill acquisition, it is often used to enhance

motor learning160 presented in appropriate format (e.g. type, frequency, content, mode).

There are two main categories of feedback, knowledge of results (KR) and knowledge of

performance (KP). KR is information about the outcome of the movement, whereas KP is

information about the movement characteristics150, both of which have advantages depending

on the goal of the skill160. As KR is a focus of this thesis, further discussion of feedback will be in

24

reference to KR. The content of KR can be information about correct aspects of performance or

errors; which is superior is still largely debated and may be dependent upon the role of the

feedback (e.g. direct external attention, provide motivation) 160. KR can take a variety of forms;

for example, the information about the movement outcome can be qualitative (e.g.

right/wrong, long/short), quantitative (e.g. numerically, graphical representation) or a

combination thereof (e.g. error occurrence and the magnitude of error); the latter of which

provides more guidance in how to correct the movement in future trials142.

Feedback can be provided in differing amounts and various schedules, which can impact the

effectiveness in improving performance and retention of a motor skill. Feedback can be

provided in real time during the repetition (i.e., concurrent feedback), or it can be provided

once the repetition is complete, called terminal feedback183. There is a vast body of literature

supporting the use of terminal feedback and its benefits. Concurrent feedback can be effective

under certain learning conditions where intrinsic feedback is not always readily available;

however, more often, concurrent feedback has a negative effect on learning. Concurrent

feedback can give rise to dependency upon external information, such that practice

performance improves, but retention of the skill when the feedback is no longer available is

degreaded188–190. It is proposed that this is due to a ‘blocking’ effect that the augmented

feedback has on the learner’s inherent feedback, whereby the skill is not truly learned, as

determined by retained ability to perform at a later period. A similar effect is seen when

evaluating frequency of feedback (e.g. how often the feedback is provided)191–193. It is largely

accepted that less than 100% feedback is best for skill retention183. The same type of

25

dependency on external information and lack of retained skilled performance can be seen

when feedback is provided after every trial.

It is clear that many variables of practice can be manipulated in order to optimize skill motor

learning, such as schedule, type, organization and use of feedback. Generally speaking, in

healthy adults, practice sessions that are variable with high contextual interference are ideal for

skill retention and transferability161. Whether repetitions of the task are practiced in parts, or as

a whole, is dependent upon the nature of the task. To reduce constrained motor control and

reduce attentional demand, implicit instruction and an external focus of attention should be

used. Based on the literature reviewed, it is apparent that augmented feedback in the form of

knowledge of results is beneficial for skill acquisition, with less feedback (<100%) showing

better retention performance. The field of motor learning is vast and very well studied,

although it remains unclear whether these same principles of practice can be effectively applied

to neurological populations such as people with stroke.

2.4.2 Motor Learning in Adults with Stroke

The field of motor learning is concerned with the acquisition of movement skills, but can be

defined in the context of recovery and rehabilitation as the reacquisition of movement skills63.

Rehabilitation is “fundamentally a process of relearning how to move”194. Due to altered

neurological structures and communication as a result of stroke, the principles of motor

learning applied to neurotypical learners may not be the most effective approach for skill

reacquisition in this population. In recent years, various aspects of practice (or therapy

26

sessions) during post-stroke motor re-learning have received increased attention19,195. Many of

these studies focus on motor tasks of the UE196–198, rather than the LE199,200, though some

studies investigating motor learning principles applied to gait will be reviewed in this section.

Optimal scheduling of practice for people with chronic hemiparesis may not be congruent to

that of healthy adults, in that variations of a task randomly distributed throughout practice may

only improve some aspects of motor performance. A study by Hanlon201 compared the effects

of random and blocked practice of a reach and grasp task among people with unilateral stroke.

The random practice group completed 10 trials per session, where the primary task was

alternated with three variations, creating high contextual interference. The blocked practice

group performed 10 trials of the primary task in blocks of 5 trials (low contextual interference),

with a five-minute rest between blocks. Unlike studies of practice schedule with neurotypical

individuals, the practice type did not have an effect on task performance during acquisition.

There was no difference in the number of trials required to meet success criterion (completion

of the task three consecutive times), nor the time required to successfully complete the task;

although the random practice group subjectively reported that it was more difficult to perform

the task. At retention trials, the random practice group required fewer trials to meet criterion,

however the blocked practice group improved in this respect from the first to second retention

trial. Interestingly, time to complete successful trials was not different between the groups at

retention trials. These finding suggest that random practice providing high contextual

interference leads to superior retention of some aspects of performance (e.g. number of trials

required to complete task) of an upper extremity task after stroke but not others (e.g. time to

complete task).

27

Principles regarding type of task training (whole vs part) are consistent with those of

neurotypical adults, in that training type is dependent on the task being trained. Motor learning

principles developed from work with neurotypical adults has shown that part task training is

ideal for tasks that can naturally be divided into parts166, while whole task practice is better for

tasks in which transition segments are imperative to the performance of the task168.

Hochstenbach and Mulder202 suggest that due to the motor, perceptual, cognitive and

behavioral effects of stroke, total movement patterns are remembered better than isolated

actions. There is limited evidence that part-task training aids in learning a serial task (e.g. sit to

stand)203, but is not as effective for learning a continuous task (e.g. walking)204, suggesting that

part/whole task training principles align with those developed in neurotypical adults. The

influence of implicit and explicit information during motor learning is also consistent with that

of healthy adults. Boyd and colleagues198,205 have done extensive work examining the effect of

explicit information on implicit learning of a motor-sequence of finger tapping. They have

confirmed that explicit information interferes with motor learning after stroke198 similar to

neurotypical adults171. Further, the magnitude of capacity for implicit learning is greater in

those with mild strokes, compared to those with moderate strokes195. It is again important to

note that these studies used an upper extremity task. In regard to focus of attention during task

performance, findings in the stroke population are again consistent with healthy adults, in that

internal focus of attention is superior. Studies examining the effect of attention foci on an

upper extremity task concluded that external focus of attention was superior206,207, which was

also found to be the case in a study evaluating the automaticity of a LE stepping task208.

28

Feedback

Augmented feedback during motor learning is important for individuals with neurologic injury

like stroke, since intrinsic feedback may be limited due to disruption of neural networks71. This

means that intrinsic feedback about movement (e.g. proprioception and tactility) may be

unreliable and/or may be interpreted incorrectly71. Feedback is often provided during therapy

sessions in the form of verbal, visual or manual guidance209, however the content of which is

usually motivational and reinforcing, rather than providing information about a movement19. It

has been suggested that both KR (knowledge of results) and KP (knowledge of performance)

are effective for improving hemiparetic arm motor performance, although KP may have more

lasting benefits and lead to better movement quality197,210,211. The optimal mode of feedback

remains unknown, however it is apparent that people with stroke are able to use a variety of

feedback sources (e.g. visual212 or auditory211). Very few studies have investigated optimal

parameters for feedback scheduling after stroke and the evidence available is inconclusive. A

study that used terminal visual feedback about an upper extremity timed tracing task at a

frequency of 100% or a relative frequency of 67%, found that feedback condition did not affect

accuracy or consistency of the task at retention when feedback was no longer provided 212.

Similarly, a study using a grip force production task providing feedback about actual force

production at either 100% or 50% frequencies, found no significant difference between groups

at no feedback retention213. It is apparent that there still remains a large gap in the literature

regarding the provision of augmented feedback during the relearning of an upper extremity

motor task.

29

While some principles for feedback of the upper extremity remain ambiguous, the provision of

augmented feedback for LE motor tasks has received considerably less attention, especially a

bilateral task that was previously automatic such as gait. Stanton and colleagues199 conducted a

review examining the use of biofeedback (visual, auditory, tactile), and the effect it has on

lower limb activities such as standing up, weight shifting and walking. Feedback provided

included information about weight distribution between limbs, muscle activity, spatiotemporal

gait parameters, and joint angles which was provided via visual, auditory or tactile mediums.

They found that biofeedback had a moderate effect on these LE activities immediately after

training, when compared with conventional therapy, suggesting that this type of feedback is

superior to conventional therapist feedback (e.g. verbal/socio-affective feedback)199 in the

short-term. However, the long-term effects of this intervention are unclear due to limited data

and biofeedback is not commonly used in stroke rehabilitation214. There is a scarcity of

information about the use of terminal feedback of any kind during gait training. This is

concerning since the interventions described above are not easily implemented in the clinical

setting. Previous work by Patterson and colleagues indicates that concurrent visual feedback

can result in immediate changes in TGA during treadmill training. The visual feedback provided

was a symmetry plot that presented information about the individual’s TGA as a ratio of right

over left stance time, where the goal was to move the symmetry plot into the target range in

the center of the display. Four of six participants who received feedback improved their

symmetry in response to this visual feedback. However, concurrent feedback typically requires

special equipment that can measure movement while it is occurring and customized programs

that can generate feedback based on measures in real time. Furthermore, while the use of

30

concurrent feedback has shown promise in improving some aspects of LE activities, it may also

lead to dependency and the blocking effect described in healthy adults, limiting capacity for

true learning to occur188. Thus there is a need to investigate the effects of terminal augmented

feedback for gait training.

2.5 Summary

With the incidence of stroke in Canada on the rise, and less than ideal stroke rehabilitation

outcomes, it is important to investigate effective rehabilitation strategies. Better post-stroke

rehabilitation outcomes will lead to decreased dependence and increased HRQoL. Persisting

gait deviations, such a TGA, can affect mobility and can have long-term negative consequences

like musculoskeletal injury, bone density loss, challenges with balance and inefficient energy

expenditure. TGA often persists after conventional rehabilitation therapy, and current

experimental interventions, which have shown some promise for improving TGA, are often not

feasible in the clinical setting. Clearly new approaches for treating TGA are needed, one of

which may be to incorporate the principles of motor learning in overground gait training.

While motor learning is well studied, and principles are firmly established for skill acquisition in

healthy adults, motor learning as it applies to individuals post-stroke is still being investigated.

This is especially true for motor learning of a bilateral lower extremity task which was

previously automatic, such as walking. The principles that could be applied to practice of a LE

task to enhance skill acquisition remains largely unknown. Further knowledge of how these

principles apply to motor learning of the LE after stroke can aid the design of gait training

programs that effectively improve TGA.

31

2.6 Study Objectives

This study aims to investigate the use of visual feedback provided to individuals with stroke

during an overground walking practice session to improve TGA at a retention session. The initial

primary objective of this study was to investigate the effect of different frequencies of

augmented visual feedback on temporal gait asymmetry. Based on preliminary data using the

first visual feedback display it was determined that a second objective was necessary; to

investigate the effect of different visual displays of augmented feedback on temporal gait

asymmetry.

32

Study: Use of Visual Feedback to Improve Temporal Gait Asymmetry Post-Stroke

3.1 Introduction

Walking is frequently affected by stroke and improved walking function is a commonly stated

rehabilitation goal2. Post-stroke gait features a number of spatiotemporal deviations including

decreased velocity, increased double support time and asymmetries between the limbs74,98.

Although some improvements are made with rehabilitation, gait is still considerably slower and

more asymmetric at discharge compared to healthy older adults6.

One deviation in particular, temporal gait asymmetry (TGA) occurs in 55% of individuals with

stroke and is described as spending unequal amounts of time on each leg during walking6. If left

untreated, TGA may have long-term negative consequences such as challenges with balance

control111, inefficient energy expenditure109, loss of bone mineral density in the paretic limb103,

and risk of musculoskeletal injury in the nonparetic limb102. Unfortunately, TGA does not

respond to conventional rehabilitation and there is some evidence that TGA may worsen over

time5,98,100.

Some experimental interventions show promise for improving TGA, including split belt treadmill

training215and rhythmic auditory stimulation141. While improvements in TGA have been made

with split belt treadmill training, this type of intervention is not always feasible in a clinical

setting due to high equipment costs215. Rhythmic auditory stimulation has also shown

therapeutic promise, although greater improvements are seen in other gait characteristics such

33

as step length and velocity compared to improvements in TGA141. While these experimental

interventions show some improvement in temporal symmetry, they are either not readily

accessible in the clinical setting, or not as effective for improving TGA compared to spatial

symmetry.

Overground gait training is commonly used in clinical practice and research216 and has the

benefit of repeated practice under conditions that closely approximate normal walking,

including terrain, sensory cues, weight bearing, and velocity closer to normal walking speed.

Conventional overground gait training post-stroke improves velocity and endurance216,217, but

practical and effective overground gait training strategies are required to address temporal

asymmetry in post-stroke gait.

One potential approach to improve the effectiveness of overground gait training for TGA is to

apply concepts of motor learning, defined as the relatively permanent capability to produce a

skilled movement as a result of practice or experience142. In the context of post-stroke

rehabilitation, motor recovery can be viewed as the process of relearning motor skills such as

symmetrical gait194. These parallels between motor recovery and motor (re)learning may

facilitate improvement in walking function – especially in producing more symmetrical gait.

One motor learning principle of particular interest to post-stroke gait rehabilitation is

augmented feedback, defined as any information that is not immediately available to the

learner from the body’s own movement 63. Augmented feedback is of specific interest since

sensorimotor impairments caused by stroke may impede the feedback inherent to the learner’s

movement71. The retention test is a common paradigm used to evaluate the effects of motor

34

learning concepts such as augmented feedback provided during practice of a motor task. At the

retention test, the motor task that was practiced is evaluated after a period without practice

and any improvements in performance are considered evidence of motor learning218. Studies

using such retention tests have demonstrated that augmented feedback can improve motor

learning in healthy adults191,219. Furthermore, feedback can be given to the learner during a

practice session at different intervals but in general, less frequent feedback (e.g. feedback after

every 3rd attempt) is typically better for skill retention220.

Some motor learning principles have been investigated for upper limb recovery post-stoke221

however, the optimal conditions for the delivery of augmented feedback particularly for the re-

learning of a bilateral lower limb activity like symmetrical walking has not been examined.

Some motor learning principles have been investigated for upper limb recovery post-stoke221

however, the optimal conditions for the delivery of augmented feedback particularly for the re-

learning of a bilateral lower limb activity like symmetrical walking has not been examined.

Therefore, the overall aim of this study is to investigate the use of visual feedback provided to

individuals with stroke during an overground walking practice session to improve TGA at a

retention session. The initial primary objective of this study was to investigate the effect of

different frequencies of augmented visual feedback on temporal gait asymmetry. Based on

preliminary data using the first visual feedback display it was determined that a second

objective was necessary; to investigate the effect of different visual displays of augmented

feedback on temporal gait asymmetry.

35

3.2 Methods

3.2.1 Design

This study used a cross-sectional observational design that employed an acquisition and

retention protocol142 to test the effect of 2 different visual feedback displays providing

information about the amount of time spent on each leg during walking and 3 frequencies of

visual feedback on TGA. Participants were randomly allocated into one of three feedback

frequencies: 100% feedback (after every trial), ~50% feedback (after every second trial), or 0%

feedback. Participants assigned to feedback groups (i.e., 50% or 100%) received feedback in the

form of Display A or Display B. Therefore, there were five feedback condition groups: no

feedback, 50% Display A (50A), 100% Display A (100A), 50% Display B (50B) and 100% Display B

(100B).

3.2.2 Participants

Individuals with first occurrence of stroke >6 months since onset who exhibited temporally

asymmetric gait (swing ratio >1.0698) were enrolled in this study. Participants were recruited

from the local community and screened for eligibility prior to enrollment. All participants were

over 18 years of age and able to ambulate independently with or without a walking aid.

Exclusion criteria for participation were: visual neglect, inability to comprehend and/or follow

multi step instructions, history of multiple strokes, other neurological or musculoskeletal

conditions that affected gait. This study was approved by the University Health Network

Research Ethics Board and all participants provided written informed consent.

36

3.2.3 Testing Protocol

Participants underwent a baseline assessment to characterize clinical presentation and gait.

Following the baseline assessment, participants were randomly assigned to a feedback

condition (NoFB, 50A, 100A, 50B, 100B). Participants then practiced the overground walking

task during the acquisition session (up to 25 trials), and returned to complete the retention test

(10 trials).

3.2.4 Baseline Measurements

Clinical measures were recorded to characterize type and severity of impairments. Motor

impairment of the leg and foot was measured using the Chedoke-McMaster Stroke Assessment

(CMSA), which has established validity and reliability222. CMSA is measured using a 7-point scale

with lower scores indicating more severe impairment. The National Institute of Health Stroke

Scale (NIHSS) is a valid and reliable scale for measuring neurologic impairment, with higher

scores indicating more severe impairment223. Plantar cutaneous sensation was measured using

a monofilament examination (Touch-Test Sensory Evaluator, North Coast Medical, Inc., Morgan

Hill, CA) on the heel and base of the fifth metatarsal of both feet. The monofilament was

pressed against the target area until the filament bent, then the participant was asked whether

they could feel it. Sham trials were randomly included throughout the testing, in which the

monofilament did not touch the foot, but the participant was asked if they could feel it.

Cutaneous sensation was considered intact if the participant responded correctly to 2 out of 3

trials. This process was repeated for each monofilament of decreasing thickness, until the

37

participant was no longer able to correctly identify 2 out of 3 trials, marking the individual’s

threshold for plantar cutaneous sensation. Each monofilament thickness was marked by a

score, where a higher score (i.e., thicker monofilament) indicated a reduced cutaneous

sensation. The Montreal Cognitive Assessment (MoCA) is a screening tool which covers eight

cognitive domains224 and was used to characterize level of cognitive impairment. The MoCA is

both sensitive and reliable for detecting mild cognitive impairment in stroke225.

3.2.5 Gait measures

Spatiotemporal parameters of gait were measured using a pressure sensitive mat. The Zeno

Walkway (ProtoKinetics, Havertown, PA, USA) is 488 cm long and 61 cm wide mat with 18432

1.27 cm x 1.27 cm embedded sensors. The scan rate was 120 Hz. The pressure sensitive mat

recorded footfall events, which were processed using PKMAS software to calculate spatial-

temporal parameters. Similar technology, such as GAITRite, has been used to measure spatial-

temporal parameters of gait among persons with stroke which demonstrates good test-retest

reliability226,227. Walking trials were 30 seconds of steady gait (i.e., constant mean velocity),

which typically included 2-3 passes over the pressure sensitive mat, at the participant’s

preferred gait velocity.

3.2.6 Overground walking training with feedback

Participants walked across the pressure sensitive mat during the training session, which

recorded spatiotemporal parameters used to generate the feedback provided.

38

3.2.6.1 Acquisition Session

During the acquisition session participants practiced overground walking across the pressure

sensitive mat in trials of 30 seconds. The goal was to complete 25 trials during the acquisition

session but this was adjusted according to participant tolerance. Rest breaks were provided

when necessary. During each trial, the pressure sensitive mat recorded footfall events that

were used to generate feedback. Feedback was provided to the participant at the end of the 30

second walking trials in accordance with the assigned frequency of feedback.

3.2.6.2 Retention Session

Participants returned for the retention session 24 hours after the acquisition session. This

session consisted of 10 30-second walking trials over the pressure sensitive mat. During this

session, no feedback was provided to any participants and temporal gait parameters were

recorded.

3.2.7 Visual Feedback about TGA

During the acquisition session, visual feedback was provided to participants on a tablet (Galaxy

Tab E, Samsung, Suwon, South Korea) in one of two formats: Display A (Figure 3.1) or Display B

(Figure 3.2). The visual feedback was displayed using Excel for Android (Microsoft, Redmond,

WA, USA). Display A was generated using the right and left swing times (seconds) collected by

the pressure sensitive mat on the previous walking trial. The swing time values were entered

into an Excel spreadsheet with a formula that calculated the TGA ratio (equation here). This

TGA ratio was then plotted in Excel using the scatter plot function which was then displayed to

39

the participant. Similarly, Display B was generated using the left and right swing times (percent

of gait cycle) collected by the pressure sensitive mat on the previous trial. These values were

entered into an Excel spreadsheet, which then generated a bar graph of the time spent on each

limb and was displayed to the participant. The full script for the instructions are included in

Appendix A but are briefly described in Figure 3.1 and Figure 3.2.

40

Figure 3.2 Display A - a visual analog scale with a marker (solid black line) indicating

amount of time spent on one leg compared to the other during walking. The

participant was instructed to move the marker closer to the center of the display, and

into the target zone (shaded area), by spending more even amounts of time on each

leg during the subsequent trials.

Figure 3.3 Display B - a bar graph with two bars: time spent on left side, time spent on

right side. The participant was instructed to make the left and right bars even height

by spending more even amounts of time on each leg during the subsequent trials.

41

3.2.8 Data Processing

PKMAS software was used to automatically identify footfalls and impressions made by walking

aids on the pressure sensitive mat. Each pass on the mat was then edited manually to remove

any partial footfalls. After this processing, PKMAS software automatically calculated

spatiotemporal parameters. This processing was completed following the baseline clinical gait

assessment and after each walking trial that required feedback during the acquisition session.

After the retention session, all remaining trials were processed and calculated variables were

exported to an excel file for further statistical analysis and to create individual motor learning

curves. Swing time symmetry was calculated by dividing the right and left swing times (sec.),

with the greater value as the numerator98. Other variables of interest included step length,

velocity, cadence, swing and stance time as percent of gait cycle and initial and terminal

support as percent of gait cycle.

3.2.9 Data Analysis

Descriptive Statistical Analysis

Statistical analysis was carried out with R Project 3.4.1 software (Vienna, Austria)228 and

Microsoft Excel 2016 software. Mean and standard deviation (SD) was generated for each

feedback group for demographic variables and median and interquartile range (IQR) were

generated for clinical variables (CMSA leg scores, MoCA scores, NIHSS, scores). One-way

ANOVA and Kruskal-Wallis tests were used for normally and non-normally distributed variables,

42

respectively, to compare demographic and baseline variables between groups. An alpha level of

0.05 was used as the threshold for statistical significance for all analyses.

Motor learning Curves and Single Subject Analysis

A motor learning curve for each participant was produced by plotting the TGA (left y-axis) and

velocity (right y-axis) of each trial (x-axis) during acquisition and retention sessions. Single

subject analysis, specifically the 2 standard deviation band method, was used to determine TGA

outcome at retention for each participant. The motor learning paradigm closely parallels the A-

B-A single subject study design with a behaviour (in this case TGA) measured during a baseline

period, and intervention period (acquisition) and then during withdrawl of the intervention

(retention)229. The 2SD band method is appropriate for single subject analysis when baseline

data is stable and therefore sensitive to variability across phases230, with 2 consecutive data

points outside 2SD indicating a significant change in performance across phases231. The 2SD

band method was selected for analysis in this study because baseline TGA was expected to be

reasonably stable as participants had not yet received treatment, and the multiple measures at

retention were suitable for assessing change. The mean and 2SD of baseline TGA, as measured

during clinical assessment, was calculated and used to create a band to which TGA during each

trial of the retention session was compared. If the majority of retention trials (i.e. 6/10) had a

TGA below 2 SD from the baseline TGA, the individual was classified as exhibiting a significant

change in TGA. Participants that had <6 retention trials below 2SD band were categorized as ’no

change’.

43

Comparison of Spatiotemporal Gait Parameters at Baseline and Retention

In order to determine what aspects of gait changed to accomplish greater symmetry, paired t-

tests were used to look for differences in gait parameters between baseline and retention in

participants that were classified as having ‘changed’ symmetry. Measures of interest in this

study were swing time (% of gait cycle), stance time (% of gait cycle), initial double support time

(% of gait cycle), terminal double support time (% of gait cycle), step length (cm), of the paretic

and nonparetic limbs, along with cadence (steps/min) and velocity (m/s). These spatiotemporal

variables commonly deviate from normative values in asymmetrical post-stroke gait6,232, and

were therefore of interest to determine how they change in order to achieve improved TGA.

These parameters were used in previous work to investigate change in gait symmetry within an

individual and indicated that some of these parameters changed as symmetry improved with

increased walking speed233.

Effect Size

Effect size (Cohen’s d) was calculated to determine the effectiveness of each feedback

condition (50A, 100A, 50B, 100B) compared to the control group based on mean change in TGA

from baseline to retention.

44

3.3 Results

Participants

Sixteen participants (5 females, 11 males) with a mean (SD) age of 65.2(6.6) years, who were

6.2(5.4) years post-stroke participated. All participants exhibited TGA at clinical assessment

with a mean TGA ratio of 1.32 (0.17) and performed walking trials using a walking aid as

required (8 participants used an aid). Demographic data for individual participants and group

means are presented in Table 1. The groups were not statistically significantly different on age

(F(4,15) = 0.13, p = 0.97), time post-stroke (H(4,15) = 4.93, p = 0.29), CMSA score of the leg

(H(4,15) = 2.78, p = 0.60), MoCA (F(4,15) = 0.13, p = 0.97), NIHSS score (H(4,15) = 1.55, p = 0.82),

velocity (F(4,15) = 0.86, p = 0.52).

45

Tab

le 3

.1 P

arti

cip

ant

char

acte

rist

ics

Gro

up

P

arti

cip

ant

ID

B

asel

ine

TGA

A

ge

(yea

rs)

Sex

(M/F

) Ti

me

Po

st-

Stro

ke

(yea

rs)

CM

SA

(leg

) M

oC

A

NIH

SS

Gai

t ve

loci

ty

(m/s

)

Co

ntr

ol

1 1.

47

75

M

3.9

6

27

0

1.04

(n

=4)

2 1.

32

60

M

2.9

5

20

3

0.93

3 1.

24

56

M

5.3

7

29

1

0.78

4 1.

39

61

M

2.2

5

28

2

0.85

G

rou

p m

ean

(sd

) 1.

36(0

.1)

63.0

(8.3

)

3.6(

1.3)

5.

5(1.

8)

27.5

(7.0

) 1.

5(2.

5)

0.90

(0.1

)

50

A

5 1.

45

83

M

6.7

4

28

0

0.63

(n

=2)

6 1.

25

66

M

2.6

5

25

2

0.91

G

rou

p m

ean

(sd

) 1.

35(0

.1)

74.5

(12

.02

)

4.6(

2.9)

4.

5(1.

0)

26.5

(3.0

) 1.

0(2.

0)

0.77

(0.2

)

100A

7

1.30

44

F

6.6

6

30

1

0.91

(n

=4)

8 1.

12

60

M

1.6

6

28

1

0.83

9 1.

15

81

F 8.

7

6

20

1 0.

61

10

1.

70

59

M

5.2

2

23

8

0.61

G

rou

p m

ean

(sd

) 1.

31(0

.3)

61(7

.0)

5.

5(3.

0)

6.0(

3.0)

25

.5(8

.8)

1.0(

5.3)

0.

74(0

.2)

50

B

11

1.17

69

M

1.

5

5

27

1 0.

89

(n=3

) 12

1.

24

58

M

0.8

6

27

2

0.84

13

1.34

56

F

4.5

4

25

4

0.79

G

rou

p m

ean

(sd

) 1.

25(0

.1)

61(7

.0)

2.

2(2.

0)

5.0(

2.0)

27

(2.0

) 2.

0(3.

0)

0.81

(0.1

)

100B

14

1.

62

68

F 23

.9

6

29

2 0.

92

(n=3

) 15

1.

28

68

M

2.4

5

22

4

0.78

16

1.09

64

F

21.0

) 5

23

1

0.79

G

rou

p m

ean

(sd

) 1.

33(0

.3)

66.7

(1.9

)

15.8

(11

.7)

5.3(

0.6)

24

.5(3

.8)

2.3(

1.6)

0.

83(0

.08)

Sam

ple

Mea

n (

sd)

1.32

(0.1

) 64

.2(6

.6)

6.

2(5.

44)

5.

0(1.

0)

27.0

(5.0

) 1.

5(1.

8)

0.81

(0.1

)

46

Motor Learning Curves and Single Subject Analysis

Figures 3.3-1 to 3.3-16 represent performance and retention of each participant. Half of

participants that received feedback (6/12) during practice showed improved symmetry at

retention. Severity of TGA was categorized as mild-moderate (≤1.5) or severe (>1.5), based on a

similar categorization used in a previous study examining TGA6. Fourteen participants had mild

temporal asymmetry, and two had severe asymmetry. Of those that showed improved TGA at

retention 4 had mild-moderate asymmetry and 2 had severe asymmetry.

47

1.

.

2.

.

3.

.

4.

.

5.

.

6.

.

7.

.

8.

.

9.

.

10.

.

48

11. 12.

13. 14.

.

15.

.

16.

.

Figure 3.4 Single subject analysis of TGA ratio (black line + circle marker), compared to

2SD (shaded band) of baseline asymmetry (white dashed line within shaded band),

and velocity (m/s) (grey lines + square marker) across trials for acquisition and 24 hour

retention session. Participants 1-4 did not receive feedback and participants 5-16

received feedback.

49

Participants classified as ‘change’ (i.e., exhibiting improved symmetry at retention) and ‘no

change’ with the 2SD band method are summarized in Table 3.2.

Table 3.2 TGA outcome at retention

Group Change No Change Total

Control 0 4 4 50A 1 1 2 100A 1 3 4 50B 1 2 3 100B 3 0 3

Total 6 10 16

Comparison of Spatiotemporal Parameters of Baseline Gait to Retention Gait

Mean values for spatiotemporal parameters of participants who showed change in symmetry

(i.e., improved TGA) at retention (n=6) are summarized in Table 3.3. Paired t-test of each

parameter (baseline-retention) indicated that the ‘change’ group exhibited significant changes

in swing time of the nonparetic limbs (t(5)=-3.99,p= 0.010) and stance time nonparetic limbs

(t(5)=3.99,p=0.010).

50

Table 3.3 Spatiotemporal parameters of individuals who showed change in TGA at retention gait (n=6)

Parameters Baseline Retention

Velocity (m/sec.)* 0.51 (0.16) 0.52 (0.1) Cadence (steps/min.)* 77.1 (12.9) 75.9 (11.2)

Paretic Limb

Swing Time (%)* 31.9 (4.9) 31.5 (4.0) Stance Time (%)* 68.1 (4.9) 68.5 (4.0)

Initial Double Support Time (%)* 18.6 (5.0) 17.5 (4.3) Terminal Double Support Time (%)* 25.1 (5.9) 24.1 (6.7)

Step Length (cm)* 39.2 (9.0) 40.3 (6.4)

Nonparetic Limb Swing Time (%)* 23.8 (5.6) 26.8 (6.1)

Stance Time (%)* 76.2 (5.6) 73.2 (6.1) Initial Double Support Time (%)* 25.3 (6.1) 24.0 (6.7)

Terminal Double Support Time (%)* 18.7 (4.9) 17.5 (4.3) Step Length (cm)* 40.1 (9.1) 42.6 (7.5)

Note. Values are mean (SD) * p = < .05

51

Effect Size

All FB conditions had at least one participant demonstrate a change in TGA at retention.

However, Cohen’s d values for Display B were larger compared to Display A for both 50% and

100% feedback frequencies. Feedback condition 100B had the largest effect size and a mean

difference of change in TGA at least twice as large as all other feedback conditions.

Table 3.4 Effect size (Cohen’s d)

Group Mean Difference Pooled SD Effect Size

50A (n=2) 0.05 0.09 0.54 100A (n=4) -0.04 0.11 -0.37

50B (n=3) 0.07 0.08 0.86 100B (n=3) 0.14 0.12 1.23

3.4 Discussion

Providing augmented feedback during a single session of overground walking can significantly

change TGA in some individuals with post-stroke TGA. In contrast, participants who performed

overground walking with no feedback exhibited no improvement in TGA. With respect to

feedback display and frequency, both displays and frequencies induced change in TGA in at

least one participant and 3 out of 4 display-frequency combinations had at least a moderate

effect size (e.g. d>0.5). However, Display B with 100% feedback was associated with a change in

TGA in 100% of participants that received it and it also had the largest effect size of all the

feedback display-frequency combinations. Thus, the effect of augmented feedback on TGA may

52

be influenced by the format in which information about TGA is presented and the frequency at

which it is given.

Both frequencies of feedback for Display B had a large effect on the change in TGA compared to

those in the control group. Display A had a moderate effect on TGA when feedback was

provided for 50% of trials. Meanwhile, receiving Display A for 100% of trials had a small

negative effect, wherein mean change in TGA for the group was worse compared to the control

group. It is suspected that Display A was misinterpreted by participants and therefore did not

have a great effect on TGA. During the acquisition session multiple participants verbalized

finding Display A difficult to understand. Other participants described the way they were

interpreting the feedback at some point during or after the acquisition session, which was

different from its intended meaning. In these cases participants’ statements indicated their

interpretation of feedback as indicating their position spatially on the mat (i.e., walking in the

middle of the mat versus to one side or the other) rather that the intended meaning which was

the temporal symmetry pattern. For example, one participant said, “I was trying to walk closer

to the middle of the mat, to move the line on the display”. This was somewhat surprising since

previous work by Patterson220 used a similar visual display during a single treadmill training

session which was associated with reduced temporal asymmetry and no such misinterpretation

was observed. However, in this case the feedback was provided concurrently which may have

facilitated the correct interpretation of feedback display. In previous post-stroke motor

learning studies, a variety of visual information is presented as augmented feedback, however a

study by Sackley and Lincoln used a display similar to Display B of the current study, to evaluate

its effect on weight bearing symmetry during quiet standing234. Two vertical bars represented

53

the weight distribution of each limb on force plates, moving up and down in accordance with

weight bearing of the left and right side, where the goal was to bring the bars within 5% of each

other. While this visual feedback intervention resulted in improved weight bearing symmetry

and reduced postural sway, the feedback was concurrent, and the population was individuals

with subacute stroke234. The design of visual displays has not been well studied, although

abstract visualizations such as the plots used to design Display A (target area/shaded band) and

Display B (bars representing distribution) has proven to be effective for motor learning in

healthy adults and individuals with stroke220,234,235. Future work could investigate visual

feedback design factors that enhance interpretation and hence effectiveness of visual feedback,

as the current state of the knowledge is very limited.

Previous work on the use of visual feedback for motor learning in healthy adults suggests that

the stage of the learner and the complexity of the task dictate the frequency at which feedback

should be provided. In the early phases of learning, frequent feedback aids the learner in

understanding the task and exploring movement patterns145,193. It is also suggested that

frequent feedback is optimal for the learning of complex motor tasks236. It is possible that the

100% frequency schedule resulted in more participants with changed TGA, compared to those

that received feedback at 50% frequency, due to the learning stage of the participants and the

nature of the task. Participants were in the early stages of learning the novel bilateral

coordination task and therefore more frequent feedback may have facilitated the search for a

movement pattern that resulted in greater symmetry more successfully than those who

received less feedback.

54

It is important to note that TGA with visual feedback did not appear to be accompanied by a

change in velocity at retention. This is positive since it implies that altering one component of

the gait pattern (i.e. temporal symmetry) does not necessarily result in a negative trade-off on

other aspects of gait (i.e. forward progression). Furthermore, reduced TGA at retention seemed

to be achieved by an increase in nonparetic swing time indicating that participants spent more

time in SLS of the paretic limb in order to make time spent on each limb more equal. This is an

encouraging outcome since the intended effect of the feedback was to improve TGA through

increased use of the paretic limb which was recommended by previous work on the use of

feedback to improve TGA220. Increased time spent in SLS on the paretic limb is likely associated

with increased weight bearing which is beneficial for maintaining bone mineral density and

reducing risk of hip fracture after stroke103. Similarly, decreased overall stance time of the

nonparetic limb means reduced weight bearing on the nonparetic limb which lessens the

impact that repetitive and prolonged loading can have on the joints and reduce the risk of

musculoskeletal injury102. Improved symmetry of swing and stance time of the nonparetic limb

may also lead to more efficient energyexpenditure109. The impact of continued overground gait

training with visual feedback on TGA and these secondary negative consequences are

important issues to investigate further.

Improved symmetry was not associated with velocity or severity of TGA at baseline. One

limitation of the 2SD band method is that the threshold used to judge change in TGA is not the

same across individuals. Those with greater variability (i.e. larger SD) at baseline had a much

larger 2SD band to which TGA at retention was compared to those with less variability at

55

baseline. The range of SD for the participants that were classified as change in TGA was 0.01 –

0.07 whereas the SD range for the group that did not change in TGA was 0.02 – 0.09.

3.5 Conclusion

The main finding of this study is that TGA changes significantly in some individuals post-stroke

after receiving a single session of overground walking with augmented visual feedback.

However, the effect may be influenced by the format in which the FB is provided. It also

appears that both 50% and 100% FB results in improved TGA at retention but 100% has a bigger

effect size. Results from this study, will inform longer-term TGA gait training studies, using

display format B and providing feedback for 100% of trials. An overground gait training program

that improves post-stroke TGA may reduce the risk of long-term negative consequences

associated with TGA, increase mobility, decrease dependence and improve HRQoL.

56

General Discussion

Conventional gait training improves some aspects of post-stroke gait such as velocity and

endurance123, however, more effective interventions for treating post-stroke gait asymmetry

are required. Previous work has shown improvement in spatial variables with experimental

interventions, but do not improve temporal asymmetry to the same extent141,215. As TGA can

lead to long-term negative consequences68,103,109, it is important that effective interventions

that target TGA are developed. The current work investigates one such approach, using

augmented feedback and the principles of motor learning to improve TGA.

Responsiveness of TGA to intervention

Temporally asymmetric gait has been considered resistant to intervention237, and some

consider it a neuromuscular compensation necessary for optimizing walking function238.

However, the current work demonstrates that changes can be made to temporal parameters of

asymmetric gait after just one practice session and that these changes do not compromise

other functional measures of gait such as velocity. All participants included in the study were in

the chronic stage of stroke recovery, when improvements in walking are said to be limited3, and

yet half of participants who received feedback retained changes to their walking pattern 24

hours later in response to the feedback provided. The findings of this study suggest that more

temporally symmetric gait can be achieved in chronic stroke in response to interventions

targeted at changing temporal gait parameters and does not compromise other features of gait

considered important for functional ambulation.

57

It is possible that responsiveness of TGA to visual feedback was impacted by individual

underlying sensorimotor and cognitive impairments. Although there were no differences in

CMSA, NIHSS, MoCA and plantar cutaneous scores between groups, the individualized nature of

impairments among people with stroke causes no surprise that individuals of varying degrees of

motor function and neurologic impairment were able to change TGA in response to the

feedback. It is reasonable to surmise that those with greater somatosensory impairment would

receive less reliable intrinsic feedback during practice about the result of the movement, and

more challenges in executing the movement71. Nonetheless, there appears to be no trend in

severity of impairment and responsiveness of TGA to visual feedback. Cognitive impairment has

also been linked to degraded motor control and learning239, yet the median MoCA scores of

those who had changed symmetry (median (IQR), 25.5(5.5)) at retention is remarkably similar

to those that did not respond to feedback (26(4.8)). However, the range of the level of

impairments is limited due to the small sample size and a future study with a larger sample size

may better detect the effect of different impairments on the use of visual feedback to improve

TGA in people with stroke.

Another potential factor that may impact responsiveness of TGA to feedback is lesion location.

Lesion location was self-reported by 8 participants to the best of their knowledge (3 basal

ganglia, central sulcus, internal capsule, pons, frontal lobe, middle cerebral artery, posterior

base) and was unknown for the other 8 participants. It is long known that the cerebellum plays

important roles in motor control and learning154. There is evidence from fMRI studies of

sequential motor skills that the cerebellum is particularly active during the early phase of

58

learning240, when the movement is explored and adapted147. A lesion to the cerebellum could

impede motor relearning and interfere with the coordination of movement. It is possible that a

lack of responsiveness to the intervention, especially during the early stages of relearning the

task, is due to injury to the cerebellum. However, no participants in the sample reported lesion

to the cerebellum. Another lesion site that may impede relearning of temporally symmetric

gait, is the putamen. There is evidence that damage to the putamen is associated with the

presence of TGA in the chronic stage241. A lesion to this location may cause TGA to be

particularly resistant to interventions, at least over one practice session. Interestingly, 3 of the 6

participants who did not respond to feedback, reported basal ganglia as lesion location.

However, imaging studies of brain activation patterns during motor tasks indicate that, despite

injury to these areas, motor learning is possible through recruitment of secondary motor areas

such as supplementary and premotor areas242,243.

Display Format

The current work demonstrated that augmented feedback in the form of visual information

about temporal asymmetry can improve TGA after one overground practice session. The study

specifically investigated the effectiveness of different visual displays and the optimal frequency

at which it should be provided during practice. The results suggest that the format the

information is presented in may play a role in its effectiveness. Both displays presented

information about the amount of time spent on one leg compared to the other during a 30

second walking trial, but Display B resulted in twice as many participants exhibiting changed

TGA at the retention session compared to Display A. It seems likely that this is due, at least in

59

part, to misinterpretation of the information presented in Display A by some participants, and

statements to that effect were made by multiple participants. They expressed directly, or

implied through various comments that they were using the information presented in Display A

to adjust where they were walking on the mat (e.g. closer to the middle of the mat), rather than

adjusting the time spent on each leg during the trial. This is despite repeated verbal instructions

about how to interpret the display with respect to timing of events in the gait cycle. It is

possible that participants perceived Display A to be a literal representation of the mat, with the

red band indicating where on the mat they were aiming to walk and the black line representing

their actual location on the mat. Participants viewed the feedback while standing or sitting at

the end of the mat looking down the length of it, which may have also contributed to Display A

being perceived as a representation of the width of the mat. Furthermore, the goal of the task

was to move the marker across the display, which is inherently a spatial task, and perhaps

another possible reason for misinterpretation of the information provided in Display A.

Meanwhile, Display B seemed to be more intuitive and participants expressed no uncertainty

about the meaning of the information presented. It is possible that Display B was more easily

understood since each bar was a direct measure of time was spent on each leg, and it was more

apparent which one was larger. It may be assumed that those that improved TGA, understood

that the display was presenting temporal information about their walking, since improved TGA

was achieved by decreasing stance time of the nonparetic limb; the intended effect of the

feedback.

60

Feedback Frequency

One visual display appeared more effective than the other, and the same is true of the

frequency at which it was provided, again with double the amount of participants showing

improved TGA receiving feedback after every trial (100%), opposed to receiving feedback after

every other trial (~50%). In addition, Display B given at 100% frequency also had the largest

effect size. This is opposite to feedback frequency principles established for healthy adults

where too much feedback can degrade motor learning191, although the larger amount of

feedback may be more beneficial for people with stroke since it supplements intrinsic feedback

that may be otherwise compromised71. It is also possible that the greater need for feedback is

only present in the early stages in learning the task, then, similar to healthy adults, less

feedback may be recommended as the learner has a good understanding of the task and the

new focus is to refine the movement pattern147. Future work examining the effect of feedback

frequency in different stages of relearning could optimize gait outcomes after stroke.

Implications for Rehabilitation

This type of intervention using augmented feedback shows promise for gait training targeting

temporal asymmetry in people with chronic stroke, although it is not necessarily feasible in a

clinical setting at this stage largely due to equipment costs. Some pieces of equipment used,

such as a tablet and laptop, are often common-place in clinics, although the pressure sensitive

mat used to record spatiotemporal gait parameters used to generate the feedback is quite

costly. Many clinics may not have funds to support this type of equipment, although they are

61

being adopted in some settings associated with large research and teaching hospitals in large

urban settings. An alternative method for recording spatiotemporal gait parameters that is

more cost effective (e.g. accelerometers/tablet application), could make this type of

intervention feasible for implementing in therapy.

However, some findings from the current work may be relevant to clinical practice once tested

further in a larger clinical trial, such as the frequency of feedback to provide during gait training

after stroke. The results from this work suggest that more feedback (e.g. 100%) during the early

stages of relearning a task may better facilitate the process, specifically in gait training for

temporal gait asymmetry. Therapists using feedback (e.g. visual or verbal feedback) during gait

training targeting temporal asymmetry, may better facilitate the learning process by providing

feedback after every repetition. This study indicates that one practice session leads to short-

term retention of changes to symmetry, however, future work could investigate the impact of

more practice sessions and longer follow up periods to determine the long-term retention of

these changes. Furthermore, this study evidenced that interpretation of feedback can change

from learner to learner and may not be interpreted by the learner in the intended way. It is

suggested that therapists confirm that patients understand and interpret the information

represented in the visual display in the intended way. Providing concise instructions that

outline the visual information may have a large impact on whether the feedback is effective

during therapy. Further work investigating the interpretation and comparison of visual displays

could inform the design a display that effectively conveys visual information for relearning a

walking task in people with chronic stroke.

62

4.1 Limitations

One limitation of this study is the small sample size and uneven groups sizes, leading to poor

generalizability. Part of the challenge in enrolling participants, was finding participants who fit

the inclusion criteria, especially temporally asymmetric gait. Twenty-four people with a first

occurrence of stroke were screened, of which seventeen people exhibited temporally

asymmetric gait. One participant with TGA was unable to complete the study due to the time

sensitive nature of the study visits; acquisition and retention sessions required travel to

Toronto Rehabilitation Institute on two consecutive days. Many other potential participants

were contacted, but declined participation; indicating that they would be unable to meet the

specific time demands of the study visits. Small group sizes were due to a protocol change in

which Display B was added as a feedback condition, when it became evident that Display A was

ineffective, likely due to misinterpretation of the information presented. Although the

statistical analysis lacks power, single subject analysis enables conclusions about the effects of a

condition based on the response of a single participant to be drawn. Analysis by group in this

type of size sample is also made possible by the effect size calculation, that takes into account

the small sample size. It is important to note that the small sample size of the current study

limits the conclusions that can be drawn. The findings from this study will inform the next of

several steps towards an eventual large clinical trial study. The next step is to test Display B

provided at 100% frequency compared to a control group receiving no feedback, to support the

format and frequency of visual feedback that should be used to improve post-stroke TGA.

63

Another limitation of the study is the baseline TGA measure that was used to determine change

in symmetry at retention. Many participants showed great variability of TGA during the 30

second walking trial, which is common in post-stroke gait244. However, this variability created

different standards for improvement for each individual using the 2SD band method. Since

improved TGA was defined by the majority of retention trials having a TGA ratio less than 2SD

of baseline, those that showed more variation at baseline were required to improve by a larger

margin. Meanwhile, it would be more likely for those that showed little variation at baseline, to

improve TGA. Currently, there is no criteria for selecting the most appropriate analysis method

for single subject research designs230. However, the 2SD band method was used in this work

because it assumes stable baseline data (i.e., TGA prior to intervention), is sensitive to changes

across phases (i.e., baseline, acquisition, retention) and the multiple measures at retention lent

itself well to assessing change (i.e., ≥6 trials below 2SD).

The retention test paradigm used in this study consisted of one practice session (25 trials) and

one retention session (10 trials) 24 hours later. Many post-stroke motor learning studies

employ a similar protocol of one acquisition and one retention session180,212, however multiple

practice sessions are also common245,246 and are closer to the reality of clinical practice in stroke

rehabilitation. The use of one practice session may be seen as a limitation, though there is

evidence that one overground gait training session can induce short-term changes to

spatiotemporal parameters in people with stroke247. The shorter testing protocol used in the

current study showed changes in TGA over a single session and allowed for proof of concept to

inform a larger gait training study. This suggests that further training using these methods could

leader to greater changes in TGA.

64

4.2 Future directions

The findings from the current work demonstrate that visual FB can lead to improvements in

TGA after one training session and provides insight for the development of future gait training

programs. A larger randomized controlled trial using Display B to provide visual information

about TGA at a frequency of 100% is recommended to investigate the optimal frequency at

which FB should be provided. A future clinical trial should be designed with frequency and

intensity of training comparable to that used in other studies on gait rehabilitation with chronic

stroke, with multiple follow up sessions. For example, training may take place over a 4 week

period, 2-3 session per week, each session consisting of 25 trials with follow-ups at 1 and 3

months post training 43,200,248. Follow-up will aid in determining longer-term retention of

improvements to TGA following gait training with visual FB. A gait training program for

improving TGA that employs the principles of motor learning, as they apply to individuals with

stroke, may help reduce long-term negative consequences, decrease dependence and increase

HRQoL.

65

References

1. Algurén, B., Lundgren-Nilsson, Å. & Sunnerhagen, K. S. Functioning of stroke survivors – A validation of the ICF core set for stroke in Sweden. Disabil. Rehabil. 32, 551–559 (2010).

2. Bohannon, R. W., Andrews, A. W. & Smith, M. B. Rehabilitation goals of patients with hemiplegia. Int. J. Rehabil. Res. 11, 181–184 (1988).

3. Jorgensen, H. S., Nakayama, H., Raaschou, H. O. & Olsen, T. S. Recovery of walking function in stroke patients: The Copenhagen study. Arch. Phys. Med. Rehabil. 76, 27–32 (1995).

4. Turnbull, G. & Wall, J. Long-term changes in hemiplegic gait. Gait Posture 3, 258–261 (1995).

5. Patterson, K. K., Gage, W. H., Brooks, D., Black, S. E. & McIlroy, W. E. Changes in Gait Symmetry and Velocity After Stroke: A Cross-Sectional Study From Weeks to Years After Stroke. Neurorehabil. Neural Repair 24, 783–790 (2010).

6. Patterson, K. K. et al. Gait Asymmetry in Community-Ambulating Stroke Survivors. Arch. Phys. Med. Rehabil. 89, 304–310 (2008).

7. Platts, M. M., Rafferty, D. & Paul, L. Metabolic Cost of Overground Gait in Younger Stroke Patients and Healthy Controls: Med. Sci. Sports Exerc. 38, 1041–1046 (2006).

8. Jørgensen, L., Jacobsen, B. K., Wilsgaard, T. & Magnus, J. H. Walking after Stroke: Does It Matter? Changes in Bone Mineral Density Within the First 12 Months after Stroke. A Longitudinal Study. Osteoporos. Int. 11, 381–387 (2000).

9. Wade, D. T., Collen, F. M., Robb, G. F. & Warlow, C. P. Physiotherapy intervention late after stroke and mobility. BMJ 304, 609–613 (1992).

10. Wade, D., Wood, V., Heller, A., Maggs, J. & Langton Hever, R. Walking after stroke. Measurement and recovery over the first 3 months. Scand. J. Rehabil. Med. 19, 25–30 (1987).

11. Schmid, A. et al. Improvements in Speed-Based Gait Classifications Are Meaningful. Stroke 38, 2096–2100 (2007).

12. Jørgensen, H. S. et al. Outcome and time course of recovery in stroke. Part I: Outcome. The Copenhagen Stroke Study. Arch. Phys. Med. Rehabil. 76, 399–405 (1995).

13. Smith, G. V., Silver, K. H. C., Goldberg, A. P. & Macko, R. F. “Task-Oriented” Exercise Improves Hamstring Strength and Spastic Reflexes in Chronic Stroke Patients. Stroke 30, 2112–2118 (1999).

66

14. Whitall, J., Waller, S. M., Silver, K. H. C. & Macko, R. F. Repetitive Bilateral Arm Training With Rhythmic Auditory Cueing Improves Motor Function in Chronic Hemiparetic Stroke. Stroke 31, 2390–2395 (2000).

15. Page SJ. Imagery improves upper extremity motor function in chronic stroke patients: a pilot study. Occup. Ther. J. Res. 20, 200–215 (2000).

16. Page SJ, Sisto S, Levine P & McGrath RE. Efficacy of modified constraint-induced movement therapy in chronic stroke: a single-blinded randomized controlled trial. Arch. Phys. Med. Rehabil. 85, 14–174 (2004).

17. Dias, D. et al. Can we improve gait skills in chronic hemiplegics? A randomised control trial with gait trainer. Eur. Medicophysica 43, 499–504 (2007).

18. Page, S. J., Gater, D. R. & Bach-y-Rita, P. Reconsidering the motor recovery plateau in stroke rehabilitation. Arch. Phys. Med. Rehabil. 85, 1377–1381 (2004).

19. Talvitie, U. Socio-affective characteristics and properties of extrinsic feedback in physiotherapy. Physiother. Res. Int. 5, 173–189 (2000).

20. Heart & Stroke Foundation. Together against a rising tide: Advancing stroke systems of care. (2014).

21. Krueger, H. et al. Prevalence of Individuals Experiencing the Effects of Stroke in Canada. Stroke 46, 2226–2231 (2015).

22. Langhorne, P., Coupar, F. & Pollock, A. Motor recovery after stroke: a systematic review. Lancet Neurol. 8, 741–754 (2009).

23. Tatemichi, T. K. et al. Cognitive impairment after stroke: frequency, patterns, and relationship to functional abilities. J. Neurol. Neurosurg. Psychiatry 57, 202–207 (1994).

24. Hackett, M. L. & Pickles, K. Part I: frequency of depression after stroke: an updated systematic review and meta-analysis of observational studies. Int. J. Stroke 9, 1017–1025 (2014).

25. Langhorne, P., Bernhardt, J. & Kwakkel, G. Stroke rehabilitation. Lancet 377, 1693–1702 (2011).

26. Hankey, G. J., Jamrozik, K., Broadhurst, R. J., Forbes, S. & Anderson, C. S. Long-Term Disability After First-Ever Stroke and Related Prognostic Factors in the Perth Community Stroke Study, 1989–1990. Stroke 33, 1034–1040 (2002).

27. Hackett, M. L., Duncan, J. R., Anderson, C. S., Broad, J. B. & Bonita, R. Health-Related Quality of Life Among Long-Term Survivors of Stroke: Results From the Auckland Stroke Study, 1991–1992. Stroke 31, 440–447 (2000).

67

28. Prabhakaran, S. et al. Inter-individual Variability in the Capacity for Motor Recovery After Ischemic Stroke. Neurorehabil. Neural Repair 22, 64–71 (2008).

29. Masiero, S., Avesani, R., Armani, M., Verena, P. & Ermani, M. Predictive factors for ambulation in stroke patients in the rehabilitation setting: A multivariate analysis. Clin. Neurol. Neurosurg. 109, 763–769 (2007).

30. Bagg, S., Pombo, A. P. & Hopman, W. Effect of age on functional outcome after stroke rehabilitation. Stroke 33, 179–185 (2002).

31. Nakayama, H., Jørgensen, H. S., Raaschou, H. O. & Olsen, T. S. The influence of age on stroke outcome. The Copenhagen Stroke Study. Stroke 25, 808–813 (1994).

32. Hendricks, H. T., Zwarts, M. J., Plat, E. F. & van Limbeek, J. Systematic review for the early prediction of motor and functional outcome after stroke by using motor-evoked potentials. Arch. Phys. Med. Rehabil. 83, 1303–1308 (2002).

33. Chen, C.-L., Tang, F.-T., Chen, H.-C., Chung, C.-Y. & Wong, M.-K. Brain lesion size and location: Effects on motor recovery and functional outcome in stroke patients. Arch. Phys. Med. Rehabil. 81, 447–452 (2000).

34. Biernaskie, J., Chernenko, G. & Corbett, D. Efficacy of Rehabilitative Experience Declines with Time after Focal Ischemic Brain Injury. J. Neurosci. 24, 1245–1254 (2004).

35. Carmichael, S. T. et al. Growth-associated gene expression after stroke: evidence for a growth-promoting region in peri-infarct cortex. Exp. Neurol. 193, 291–311 (2005).

36. Jørgensen, H. S. et al. Outcome and time course of recovery in stroke. Part II: Time course of recovery. The Copenhagen Stroke Study. Arch. Phys. Med. Rehabil. 76, 406–412 (1995).

37. Kelly-Hayes, M. et al. Time course of functional recovery after stroke: the Framingham study. J. Neurol. Rehabil. 3, 65–70 (1989).

38. Verheyden, G. et al. Time Course of Trunk, Arm, Leg, and Functional Recovery After Ischemic Stroke. Neurorehabil. Neural Repair 22, 173–179 (2008).

39. Skilbeck, C. E., Wade, D. T., Hewer, R. L. & Wood, V. A. Recovery after stroke. J. Neurol. Neurosurg. Psychiatry 46, 5–8 (1983).

40. Kinsella, G. & Ford, B. Acute recovery from patterns in stroke patients: neuropsychological factors. Med. J. Aust. 2, 663–666 (1980).

41. Andrews, K., Brocklehurst, J. C., Richards, B. & Laycock, P. J. The rate of recovery from stroke-and its measurement. Int. Rehabil. Med. 3, 155–161 (1981).

68

42. Flansbjer, U.-B., Miller, M., Downham, D. & Lexell, J. Progressive Resistance Training After Stroke: Effects on Muscle Strength, Muscle Tone, Gait Performance and Perceived Participation. J. Rehabil. Med. 40, 42–48 (2008).

43. Ada, L., Dean, C. M., Hall, J. M., Bampton, J. & Crompton, S. A treadmill and overground walking program improves walking in persons residing in the community after stroke: a placebo-controlled, randomized trial. Arch. Phys. Med. Rehabil. 84, 1486–1491 (2003).

44. Teixeira-Salmela, L. F., Olney, S. J., Nadeau, S. & Brouwer, B. Muscle strengthening and physical conditioning to reduce impairment and disability in chronic stroke survivors. Arch. Phys. Med. Rehabil. 80, 1211–1218 (1999).

45. Ouellette, M. M. et al. High-Intensity Resistance Training Improves Muscle Strength, Self-Reported Function, and Disability in Long-Term Stroke Survivors. Stroke 35, 1404–1409 (2004).

46. Murphy, T. H. & Corbett, D. Plasticity during stroke recovery: from synapse to behaviour. Nat. Rev. Neurosci. 10, 861–872 (2009).

47. Tangeman, P., Banaitis, D. & Williams, A. Rehabilitaion of chronic stroke patients: changes in functional performance. Arch. Phys. Med. Rehabil. 71, 876–880 (1990).

48. van der Lee, J. H. et al. Forced use of the upper extremity in chronic stroke patients. (Vrije Universiteit Amsterdam, The Netherlands, 2001).

49. Kunkel, A. et al. Constrint-induced movement therapy for moto recovery in chronic stroke patients. Arch. Phys. Med. Rehabil. 80, 624–628 (1999).

50. Mayo, N. E., Korner-Bitensky, N. A. & Becker, R. M. Recovery time of independent function post-stroke. J. Phys. Med. 70, 5–12 (1991).

51. Olsen, T. S. Arm and leg paresis as outcome predictors in stroke rehabilitation. Stroke 21, 247–251 (1990).

52. Friedman, P. J. Gait recovery after hemiplegic stroke. Int. Disabil. Stud. 12, 119–122 (1990).

53. Wade, D. T., Wood, V. A. & Hewer, R. L. Recovery after stroke–the first 3 months. J. Neurol. Neurosurg. Psychiatry 48, 7–13 (1985).

54. Stuart, M. et al. Community-Based Adaptive Physical Activity Program for Chronic Stroke: Feasibility, Safety, and Efficacy of the Empoli Model. Neurorehabil. Neural Repair 23, 726–734 (2009).

69

55. Pang, M. Y. C., Eng, J. J., Dawson, A. S., McKay, H. A. & Harris, J. E. A Community-based Fitness and Mobility Exercise (FAME) Program for Older Adults with Chronic Stroke: a Randomized Controlled Trial. J. Am. Geriatr. Soc. 53, 1667–1674 (2005).

56. Perry, J. & Burnfield, J. M. Gait Analysis: Normal and Pathological Function. (SLACK Incorporated, 2010).

57. Garcia, M., Chatterjee, A., Ruina, A., Coleman, M. & others. The simplest walking model: stability, complexity, and scaling. J Biomech Eng Trans ASME 120, 281–288 (1998).

58. Carlsöö, S. The initiation of walking. Cells Tissues Organs 65, 1–9 (1966).

59. Nissan, M. & Whittle, M. W. Initiation of gait in normal subjects: a preliminary study. J. Biomed. Eng. 12, 165–171 (1990).

60. Murray, M. & Clarkson, B. The vertical pathways of the foot during level walking. I. Range of variability in normal men. Phys. Ther. 46, 585–9 (1966).

61. Waters, R. L. & Mulroy, S. The energy expenditure of normal and pathologic gait. Gait Posture 9, 207–231 (1999).

62. Bajawa, Z., Wootton, R. & Warfield, C. Principles and Practice of Pain Medicine. (McGraw Hill Education, 2016).

63. Shumway-Cook, A. & Woolacott, M. Motor control: Translating research into clinical practice. (Lippincott Williams & Wilkins, 2012).

64. Nielsen, J. B. How we walk: central control of muscle activity during human walking. The Neuroscientist 9, 195–204 (2003).

65. Brown, T. G. The Intrinsic Factors in the Act of Progression in the Mammal. Proc. R. Soc. Lond. Ser. B Contain. Pap. Biol. Character 84, 308–319 (1911).

66. Brown, T. G. The Factors in Rhythmic Activity of the Nervous System. Proc. R. Soc. Lond. Ser. B Contain. Pap. Biol. Character 85, 278–289 (1912).

67. Duysens, J. & Van de Crommert, H. W. Neural control of locomotion; Part 1: The central pattern generator from cats to humans. Gait Posture 7, 131–141 (1998).

68. Kim, C. M. & Eng, J. J. Symmetry in vertical ground reaction force is accompanied by symmetry in temporal but not distance variables of gait in persons with stroke. Gait Posture 18, 23–28 (2003).

69. Dimitrijevic, M. R., Gerasimenko, Y. & Pinter, M. M. Evidence for a Spinal Central Pattern Generator in Humansa. Ann. N. Y. Acad. Sci. 860, 360–376 (1998).

70

70. Corbetta, M. et al. Common Behavioral Clusters and Subcortical Anatomy in Stroke. Neuron 85, 927–941 (2015).

71. Connell, L., Lincoln, N. & Radford, K. Somatosensory impairment after stroke: frequency of different deficits and their recovery. Clin. Rehabil. 22, 758–767 (2008).

72. Bohannon, R. Gait performance of hemiparetic stroke patients: Selected variables. Arch. Phys. Med. Rehabil. 64, 777–781 (1987).

73. Perry, J., Garrett, M., Gronley, J. K. & Mulroy, S. J. Classification of Walking Handicap in the Stroke Population. Stroke 26, 982–989 (1995).

74. Olney, S. J. & Richards, C. Hemiparetic gait following stroke. Part I: Characteristics. Gait Posture 4, 136–148 (1996).

75. von Schroeder, H. P., Coutts, R. D., Lyden, P. D., Billings Jr., E. & Nickel, V. L. Gait parameters following stroke: A practical assessment. J. Rehabil. Res. Dev. 32, 25–31 (1995).

76. Olney, S. J., Griffin, M. P. & McBride, I. D. Temporal, kinematic, and kinetic variables related to gait speed in subjects with hemiplegia: A regression approach. Phys. Ther. 74, 872–885 (1994).

77. Brandstater, M. E., De, H. B., Gowland, C. & Clark, B. M. Hemiplegic gait: analysis of temporal variables. Arch. Phys. Med. Rehabil. 64, 583–587 (1983).

78. Marks, M. & Hirschberg, G. G. Analysis of the Hemiplegic Gait. Ann. N. Y. Acad. Sci. 74, 59–77 (1958).

79. Hirschberg, G. G. & Nathanson, M. Electromyographic recording of muscular activity in normal and spastic gaits. Arch. Phys. Med. Rehabil. 33, 217–225 (1952).

80. Knutsson, E. & Richards, C. Different types of disturbed motor control in gait of hemiparetic patients. Brain 102, 405–430 (1979).

81. Den Otter, A. R., Geurts, A. C. H., Mulder, T. & Duysens, J. Abnormalities in the temporal patterning of lower extremity muscle activity in hemiparetic gait. Gait Posture 25, 342–352 (2007).

82. Dietz, V. Proprioception and locomotor disorders. Nat. Rev. Neurosci. 3, 781–790 (2002).

83. Lamontagne, A., Stephenson, J. L. & Fung, J. Physiological evaluation of gait disturbances post stroke. Clin. Neurophysiol. 118, 717–729 (2007).

84. Lamontagne, A., Richards, C. L. & Malouin, F. Coactivation during gait as an adaptive behavior after stroke. J. Electromyogr. Kinesiol. 10, 407–415 (2000).

71

85. Mulroy, S., Gronley, J., Weiss, W., Newsam, C. & Perry, J. Use of cluster analysis for gait pattern classification of patients in the early and late recovery phases following stroke. Gait Posture 18, 114–125 (2003).

86. Murray, M. P., Mollinger, L. A., Gardner, G. M. & Sepic, S. B. Kinematic and EMG patterns during slow, free, and fast walking. J. Orthop. Res. 2, 272–280 (1984).

87. Murray, M. P. P. D., Kory, R. C. M. D., Clarkson, B. H. L. P. T. & Sepic, S. B. B. S. Comparison of free and fast speed walking patterns of normal men. J. Phys. Med. 45, 8–24 (1966).

88. Chen, C.-L. et al. Gait Performance with Compensatory Adaptations in Stroke Patients with Different Degrees of Motor Recovery: Am. J. Phys. Med. Rehabil. 82, 925–935 (2003).

89. Kim, C. M. & Eng, J. J. Magnitude and pattern of 3D kinematic and kinetic gait profiles in persons with stroke: relationship to walking speed. Gait Posture 20, 140–146 (2004).

90. Bowden, M. G., Balasubramanian, C. K., Neptune, R. R. & Kautz, S. A. Anterior-Posterior Ground Reaction Forces as a Measure of Paretic Leg Contribution in Hemiparetic Walking. Stroke 37, 872–876 (2006).

91. Beyaert, C., Vasa, R. & Frykberg, G. E. Gait post-stroke: Pathophysiology and rehabilitation strategies. Neurophysiol. Clin. Neurophysiol. 45, 335–355 (2015).

92. Hsu, A.-L., Tang, P.-F. & Jan, M.-H. Analysis of impairments influencing gait velocity and asymmetry of hemiplegic patients after mild to moderate stroke11No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the authors(s) or upon any organization with which the author(s) is/are associated. Arch. Phys. Med. Rehabil. 84, 1185–1193 (2003).

93. Chen, C. et al. Ground reaction force patterns in stroke patients with various degrees of motor recovery determined by plantar dynamic analysis. Chang Gung Med. J. 30, 62 (2007).

94. Balasubramanian, C. K., Bowden, M. G., Neptune, R. R. & Kautz, S. A. Relationship Between Step Length Asymmetry and Walking Performance in Subjects With Chronic Hemiparesis. Arch. Phys. Med. Rehabil. 88, 43–49 (2007).

95. Boehm, W. L. & Gruben, K. G. Post-Stroke Walking Behaviors Consistent with Altered Ground Reaction Force Direction Control Advise New Approaches to Research and Therapy. Transl. Stroke Res. 7, 3–11 (2016).

96. Lauziere, S., Betschart, M. & Nadeau, S. Understanding Spatial and Temporal Gait Asymmetries in Individuals Post Stroke. Int. J. Phys. Med. Rehabil. 02, (2014).

97. Levin, M. F., Kleim, J. A. & Wolf, S. L. What Do Motor “Recovery” and “Compensation” Mean in Patients Following Stroke? Neurorehabil. Neural Repair 23, 313–319 (2009).

72

98. Patterson, K. K., Gage, W. H., Brooks, D., Black, S. E. & McIlroy, W. E. Evaluation of gait symmetry after stroke: A comparison of current methods and recommendations for standardization. Gait Posture 31, 241–246 (2010).

99. Hesse, S., Jahnke, M. T., Luecke, D. & Mauritz, K. H. Short-term electrical stimulation enhances the effectiveness of Botulinum toxin in the treatment of lower limb spasticity in hemiparetic patients. Neurosci. Lett. 201, 37–40 (1995).

100. Wall, J. C. & Turnbull, G. I. Gait asymmetries in residual hemiplegia. Arch. Phys. Med. Rehabil. 67, 550–553 (1986).

101. Nolan, L. et al. Adjustments in gait symmetry with walking speed in trans-femoral and trans-tibial amputees. Gait Posture 17, 142–151 (2003).

102. Norvell, D. C. et al. The prevalence of knee pain and symptomatic knee osteoarthritis among veteran traumatic amputees and nonamputees. Arch. Phys. Med. Rehabil. 86, 487–493 (2005).

103. Carda, S., Cisari, C., Invernizzi, M. & Bevilacqua, M. Osteoporosis after Stroke: A Review of the Causes and Potential Treatments. Cerebrovasc. Dis. 28, 191–200 (2009).

104. Jørgensen, L., Crabtree, N. J., Reeve, J. & Jacobsen, B. K. Ambulatory level and asymmetrical weight bearing after stroke affects bone loss in the upper and lower part of the femoral neck differently: bone adaptation after decreased mechanical loading. Bone 27, 701–707 (2000).

105. Dennis, M. & McDowell, M. Fractures after stroke: Frequency, types, and association. Stroke 33, 728–734 (2002).

106. Pouwels, S. et al. Risk of Hip/Femur Fracture After Stroke: A Population-Based Case-Control Study. Stroke 40, 3281–3285 (2009).

107. Minaire, P. Immobilization osteoporosis: a review. Clin. Rheumatol. 8, 95–103 (1989).

108. Worthen, L. C. et al. Key characteristics of walking correlate with bone density in individuals with chronic stroke. J. Rehabil. Res. Dev. 42, 761 (2005).

109. Ellis, R. G., Howard, K. C. & Kram, R. The metabolic and mechanical costs of step time asymmetry in walking. Proc. R. Soc. B Biol. Sci. 280, 20122784–20122784 (2013).

110. Davis, B. et al. Real time viual feedback diminishes energy consumption of amputee subjects during treadmill locomotion. J. Prosthet. Orthot. 16, 49–54 (2004).

111. Lewek, M. D., Wutzke, C. J. & Zinder, S. M. The Relationship Between Spatiotemporal Gait Asymmetry and Balance in Individuals with Chronic Stroke. J. Appl. Biomech. 30, 31–36 (2014).

73

112. Vivian Weerdesteyn PhD, P. T., de Niet MSc, M., van Duijnhoven MSc, H. J. & Geurts, A. C. Falls in individuals with stroke. J. Rehabil. Res. Dev. 45, 1195 (2008).

113. Forster, A. & Young, J. Incidence And Consequences Of Falls Due To Stroke: A Systematic Inquiry. BMJ 311, 83–86 (1995).

114. Harris, J. E., Eng, J. J., Marigold, D. S., Tokuno, C. D. & Louis, C. L. Relationship of Balance and Mobility to Fall Incidence in People With Chronic Stroke. Phys. Ther. Wash. 85, 150–8 (2005).

115. Michael, K. M., Allen, J. K. & Macko, R. F. Reduced Ambulatory Activity After Stroke: The Role of Balance, Gait, and Cardiovascular Fitness. Arch. Phys. Med. Rehabil. 86, 1552–1556 (2005).

116. Batchelor, F. A., Mackintosh, S. F., Said, C. M. & Hill, K. D. Falls after Stroke. Int. J. Stroke 7, 482–490 (2012).

117. Latham, N. K. et al. Physical Therapy During Stroke Rehabilitation for People With Different Walking Abilities. Arch. Phys. Med. Rehabil. 86, 41–50 (2005).

118. Eng, J. J. & Tang, P.-F. Gait training strategies to optimize walking ability in people with stroke: a synthesis of the evidence. Expert Rev. Neurother. 7, 1417–1436 (2007).

119. Graham, J. V., Eustace, C., Brock, K., Swain, E. & Irwin-Carruthers, S. The Bobath Concept in Contemporary Clinical Practice. Top. Stroke Rehabil. 16, 57–68 (2009).

120. Lennon, S. Physiotherapy practice in stroke rehabilitation: a survey. Disabil. Rehabil. 25, 455–461 (2003).

121. Natarajan, P. Current clinical practices in stroke rehabilitation: Regional pilot survey. J. Rehabil. Res. Dev. 45, 841–850 (2008).

122. Vliet, P. M. van, Lincoln, N. B. & Foxall, A. Comparison of Bobath based and movement science based treatment for stroke: a randomised controlled trial. J. Neurol. Neurosurg. Psychiatry 76, 503–508 (2005).

123. Thaut, M. H. et al. Rhythmic Auditor y Stimulation Improves Gait More Than NDT/Bobath Training in Near-Ambulatory Patients Early Poststroke: A Single-Blind, Randomized Trial. Neurorehabil. Neural Repair 21, 455–459 (2007).

124. Langhammer, B. & Stanghelle, J. K. Bobath or Motor Relearning Programme? A comparison of two different approaches of physiotherapy in stroke rehabilitation: a randomized controlled study. Clin. Rehabil. 14, 361–369 (2000).

125. Lennon, S., Ashburn, A. & Baxter, D. Gait outcome following outpatient physiotherapy based on the Bobath concept in people post stroke. Disabil. Rehabil. 28, 873–881 (2006).

74

126. Flansbjer, U.-B., Lexell, J. & Brogårdh, C. Long-Term Benefits of Progressive Resistance Training in Chronic Stroke: A 4-year Follow-Up. J. Rehabil. Med. 44, 218–221 (2012).

127. Moseley, A. M., Stark, A., Cameron, I. D. & Pollock, A. Treadmill training and body weight support for walking after stroke. in Cochrane Database of Systematic Reviews (ed. The Cochrane Collaboration) (John Wiley & Sons, Ltd, 2005). doi:10.1002/14651858.CD002840.pub2

128. Combs-Miller, S. A. et al. Body weight-supported treadmill training vs. overground walking training for persons with chronic stroke: a pilot randomized controlled trial. Clin. Rehabil. 28, 873–884 (2014).

129. Macko, R. F. et al. Treadmill Exercise Rehabilitation Improves Ambulatory Function and Cardiovascular Fitness in Patients With Chronic Stroke: A Randomized, Controlled Trial. Stroke 36, 2206–2211 (2005).

130. Outermans, J. C., van Peppen, R. P., Wittink, H., Takken, T. & Kwakkel, G. Effects of a high-intensity task-oriented training on gait performance early after stroke: a pilot study. Clin. Rehabil. 24, 979–987 (2010).

131. Chu, K. S. et al. Water-based exercise for cardiovascular fitness in people with chronic stroke: a randomized controlled trial11No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the author(s) or upon any organization with which the author(s) is/are associated. Arch. Phys. Med. Rehabil. 85, 870–874 (2004).

132. Blennerhassett, J. & Dite, W. Additional task-related practice improves mobility and upper limb function early after stroke: A randomised controlled trial. Aust. J. Physiother. 50, 219–224 (2004).

133. Yen, C.-L., Wang, R.-Y., Liao, K.-K., Huang, C.-C. & Yang, Y.-R. Gait Training—Induced Change in Corticomotor Excitability in Patients With Chronic Stroke. Neurorehabil. Neural Repair 22, 22–30 (2008).

134. Reisman, D. S., Wityk, R., Silver, K. & Bastian, A. J. Split-Belt Treadmill Adaptation Transfers to Overground Walking in Persons Poststroke. Neurorehabil. Neural Repair 23, 735–744 (2009).

135. Reisman, D. S., McLean, H., Keller, J., Danks, K. A. & Bastian, A. J. Repeated split-belt treadmill training improves poststroke step length asymmetry. Neurorehabil. Neural Repair 27, 460–468 (2013).

136. Patterson, K. K. et al. Longitudinal Changes in Poststroke Spatiotemporal Gait Asymmetry Over Inpatient Rehabilitation. Neurorehabil. Neural Repair 29, 153–162 (2015).

137. Malone, L. A. & Bastian, A. J. Spatial and temporal asymmetries in gait predict split-belt adaptation behavior in stroke. Neurorehabil. Neural Repair 28, 230–240 (2014).

75

138. Reisman DS, Bastian AJ & Morton SM. Neurophysiologic and rehabilitation insights from the split-belt and other locomotor adaptation paradigms. Phys. Ther. 90, 187–195 (2010).

139. Reisman, D. S., Wityk, R., Silver, K. & Bastian, A. J. Locomotor adaptation on a split-belt treadmill can improve walking symmetry post-stroke. Brain 130, 1861–1872 (2007).

140. Patterson, K. K. Clinician’s Commentary. Physiother. Can. 61, 161–162 (2009).

141. Thaut, M. H. & Abiru, M. Rhythmic Auditory Stimulation in Rehabilitation of Movement Disorders: A Review Of Current Research. Music Percept. 27, 263–269 (2010).

142. Schmidt, R. A. & Lee, T. D. Motor control and learning: A behavioural emphasis. (Human Kinetics, 2011).

143. Schmidt, R. A. Motor learning principles for physical therapy. Contemp. Manag. Mot. Control Probl. Proc. II STEPConference June 1991 Alex. 49–63 (1991).

144. Guthrie, E. R. The psychology of learning. (Harper & Row, 1952).

145. Schmidt, R. A. A schema theory of discrete motor skill learning. Psychol. Rev. 82, 225 (1975).

146. Newell, K. M. Motor skill acquisition. Annu. Rev. Psychol. 42, 213–237 (1991).

147. Fitts, P. M. & Posner, M. I. Human performance. (Brooks/Cole, 1967).

148. Bernstein, N. The coordination and regulation of movement. (Pergamon, 1967).

149. Vereijken, B., Emmerik, R. E. A. van, Whiting, H. T. A. & Newell, K. M. Free(z)ing Degrees of Freedom in Skill Acquisition. J. Mot. Behav. 24, 133–142 (1992).

150. Gentile, A. M. Movement organization and delayed alternation behavior of monkeys following selective ablation of frontal cortex. Acta Neurobiol Exp 32, 277–304 (1972).

151. Gentile, A. Skill acquisition: action movement, and nueromotor processes. in Movement science: foundations for physical therapy in rehabilitation (eds. Carr, J. H., Shepherd, R. B. & Gordon, J.) (Aspen Publishers, 1987).

152. Poole, J. L. Application of Motor Learning Principles in Occupational Therapy. Am. J. Occup. Ther. 45, 531–537 (1991).

153. Lange, R. K., Braun, C. & Godde, B. Coordinate processing during the left-to-right hand transfer investigated by EEG. Exp. Brain Res. 168, 547–556 (2006).

154. Doya, K. Complementary roles of basal ganglia and cerebellum in learning and motor control. Curr. Opin. Neurobiol. 10, 732–739 (2000).

76

155. Thorndike, E. L. The Law of Effect. Am. J. Psychol. 39, 212–222 (1927).

156. Anderson, A. Newell, A. Rosenbloom, PS. Feedback during training and retention of motor skills. in Cognitive skills and their acquisition. 1–55 (Earlbaum, 1981).

157. Melton, A. Fitts, PM. Perceptual motor skill learning. in Categories of human learning. (Academic Press, 1964).

158. Shea, C. H. & Kohl, R. M. Specificity and Variability of Practice. Res. Q. Exerc. Sport 61, 169–177 (1990).

159. Shoenfelt, E. L., Snyder, L. A., Maue, A. E., McDowell, C. P. & Woolard, C. D. Comparison of Constant and Variable Practice Conditions on Free-Throw Shooting. Percept. Mot. Skills 94, 1113–1123 (2002).

160. Magill, R. & Anderson, D. Motor Learning and Control: Concepts and Application. (McGraw-Hill, 2014).

161. Shea, J. B. & Morgan, R. J. Effects of contextual interference and age on acquisition, retention, and transfer of motor skill. J. Exp. Psychol. [Hum. Learn.] 5, 179–187 (1979).

162. Shea, C. H. & Kohl, R. M. Specificity and Variability of Practice. Res. Q. Exerc. Sport 61, 169–177 (1990).

163. Wu, W. F. W. et al. Contextual interference and augmented feedback: Is there an additive effect for motor learning? Hum. Mov. Sci. 30, 1092–1101 (2011).

164. Magill, R. A. & Hall, K. G. A review of the contextual interference effect in motor skill acquisition. Hum. Mov. Sci. 9, 241–289 (1990).

165. Wightman, D. C. & Lintern, G. Part-Task Training for Tracking and Manual Control. Hum. Factors J. Hum. Factors Ergon. Soc. 27, 267–283 (1985).

166. Mané, A. M., Adams, J. A. & Donchin, E. Adaptive and part-whole training in the acquisition of a complex perceptual-motor skill. Acta Psychol. (Amst.) 71, 179–196 (1989).

167. Briggs, G. E. & Brogden, W. J. The effect of component practice on performance of a lever-positioning skill. J. Exp. Psychol. 48, 375–380 (1954).

168. Naylor, J. & Briggs, G. Effects of task complexity and task organization on the relative efficiency of part and whole training methods. J. Exp. Psychol. 65, 217–24 (1963).

169. Green, T. D. & Flowers, J. H. Implicit versus Explicit Learning Processes in a Probabilistic, Continuous Fine-Motor Catching Task. J. Mot. Behav. 23, 293–300 (1991).

77

170. Masters, R. S. W. Knowledge, knerves and know-how: The role of explicit versus implicit knowledge in the breakdown of a complex motor skill under pressure. Br. J. Psychol. 83, 343–358 (1992).

171. Wulf, G. & Weigelt, C. Instructions about Physical Principles in Learning a Complex Motor Skill: To Tell or Not to Tell…. Res. Q. Exerc. Sport 68, 362–367 (1997).

172. Magill, R. A. Knowledge is more than we can talk about: Implicit learning in motor skill acquisition. Res. Q. Exerc. Sport 69, 104–110 (1998).

173. Wulf, G., HoB, M. & Prinz, W. Instructions for motor learning: Differential effects of internal versus external focus of attention. J. Mot. Behav. 30, 169–179 (1998).

174. Wulf, G. & Prinz, W. Directing attention to movement effects enhances learning: A review. Psychon. Bull. Rev. 8, 648–660 (2001).

175. Beilock, S. L., Bertenthal, B. I., McCoy, A. M. & Carr, T. H. Haste does not always make waste: Expertise, direction of attention, and speed versus accuracy in performing sensorimotor skills. Psychon. Bull. Rev. 11, 373–379 (2004).

176. Koedijker, J. M. et al. Attention and time constraints in perceptual-motor learning and performance: Instruction, analogy, and skill level. Conscious. Cogn. 20, 245–256 (2011).

177. Castaneda, B. & Gray, R. Effects of focus of attention on baseball batting performance in players of differing skill levels. J. Sport Exerc. Psychol. 29, 60–77 (2007).

178. Wulf, G. Attentional focus and motor learning: a review of 15 years. Int. Rev. Sport Exerc. Psychol. 6, 77–104 (2013).

179. Wulf, G. & Su, J. An external focus of attention enhances golf shot accuracy in beginners and experts. Res. Q. Exerc. Sport 78, 384–389 (2007).

180. Wulf, G., Lauterbach, B. & Toole, T. The Learning Advantages of an External Focus of Attention in Golf. Res. Q. Exerc. Sport 70, 120–126 (1999).

181. Marchant, D. C., Clough, P. J., Crawshaw, M. & Levy, A. Novice motor skill performance and task experience is influenced by attentional focusing instructions and instruction preferences. Int. J. Sport Exerc. Psychol. 7, 488–502 (2009).

182. Gallwey, T. & Kriegel, B. Inner Skiing. (Random House, 1977).

183. Schmidt, R. A. & Lee, T. D. Motor learning and performance: From principles to application. (Human Kinetics, 2014).

184. Anderson, D. I., Magill, R. A. & Sekiya, H. Motor Learning as a Function of KR Schedule and Characteristics of Task-Intrinsic Feedback. J. Mot. Behav. 33, 59–66 (2001).

78

185. Chiviacowsky, S. & Wulf, G. Feedback after good trials enhances learning. Res. Q. Exerc. Sport 78, 40–47 (2007).

186. Wulf, G., Shea, C. & Lewthwaite, R. Motor skill learning and performance: a review of influential factors: Motor skill learning and performance. Med. Educ. 44, 75–84 (2010).

187. Hodges, N. J. & Williams, A. M. Wulf G, Shea CH. Understanding the role of augmented feedback: the good, the bad, and the ugly. in Skill acquisition in sport: Research, theory and practice. 417 (Routledge, 2004).

188. Schmidt, R. A. & Wulf, G. Continuous Concurrent Feedback Degrades Skill Learning: Implications for Training and Simulation. Hum. Factors J. Hum. Factors Ergon. Soc. 39, 509–525 (1997).

189. Sigrist, R., Rauter, G., Riener, R. & Wolf, P. Terminal feedback outperforms concurrent visual, auditory, and haptic feedback in learning a complex rowing-type task. J. Mot. Behav. 45, 455–472 (2013).

190. Park, J.-H., Shea, C. H. & Wright, D. L. Reduced-Frequency Concurrent and Terminal Feedback: A Test of the Guidance Hypothesis. J. Mot. Behav. 32, 287–296 (2000).

191. Winstein, C. J. & Schmidt, R. A. Reduced frequency of knowledge of results enhances motor skill learning. J. Exp. Psychol. Learn. Mem. Cogn. 16, 677 (1990).

192. Wulf, G., Schmidt, R. A. & Deubel, H. Reduced feedback frequency enhances generalized motor program learning but not parameterization learning. J. Exp. Psychol. Learn. Mem. Cogn. 19, 1134 (1993).

193. Guadagnoli, M. A. & Lee, T. D. Challenge point: a framework for conceptualizing the effects of various practice conditions in motor learning. J. Mot. Behav. 36, 212–224 (2004).

194. Carr, J. H. & Shepherd, R. B. Movement Science: Foundations for Physical Therapy in Rehabilitation. (Aspen, 1987).

195. Boyd, L. A., Quaney, B. M., Pohl, P. S. & Winstein, C. J. Learning Implicitly: Effects of Task and Severity After Stroke. Neurorehabil. Neural Repair 21, 444–454 (2007).

196. Chen, J. L., Fujii, S. & Schlaug, G. The use of augmented auditory feedback to improve arm reaching in stroke: a case series. Disabil. Rehabil. 38, 1115–1124 (2016).

197. Cirstea CM, Ptito A & Levin MF. Feedback and cognition in arm motor skill reacquisition after stroke. Stroke 00392499 37, 1237–1242 (2006).

198. Boyd, L. A. & Winstein, C. J. Explicit information interferes with the implicit motor learning of both continuous and discrete movement tasks after stroke. J. Neurol. Phys. Ther. 30, 46–57 (2006).

79

199. Stanton, R., Ada, L., Dean, C. M. & Preston, E. Biofeedback improves performance in lower limb activities more than usual therapy in people following stroke: a systematic review. J. Physiother. 63, 11–16 (2017).

200. Jonsdottir, J. et al. Task-Oriented Biofeedback to Improve Gait in Individuals With Chronic Stroke: Motor Learning Approach. Neurorehabil. Neural Repair 24, 478–485 (2010).

201. Hanlon, R. E. Moto learning following unilateral stroke. Arch. Phys. Med. Rehabil. 77, 811–815 (1996).

202. Hochstenbach, J. & Mulder, T. Neuropsychology and the relearning of motor skills following stroke. Int. J. Rehabil. Res. 22, 11–19 (1999).

203. Dean, C. M. & Shepherd, R. B. Task-Related Training Improves Performance of Seated Reaching Tasks After Stroke: A Randomized Controlled Trial. Stroke 28, 722–728 (1997).

204. Winstein, C. J., Gerdner, E. R., McNeal, D. R., Barto, P. S. & Nicholson, D. E. Standing balance trainer: Effect on balance and locomotion in hemiparetic adults. Arch. Phys. Med. Rehabil. 755–762 (1989).

205. Boyd, L. A. & Winstein, C. J. Impact of explicit information on implicit motor-sequence learning following middle cerebral artery stroke. Phys. Ther. 83, 976–989 (2003).

206. Durham, K. F. et al. Attentional focus of feedback for improving performance of reach-to-grasp after stroke: a randomised crossover study. Physiotherapy 100, 108–115 (2014).

207. Fasoli, S. E., Trombly, C. A., Tickle-Degnen, L. & Verfaellie, M. H. Effect of instructions on functional reach in persons with and without cerebrovascular accident. Am. J. Occup. Ther. 56, 380–390 (2002).

208. Kal, E. C. et al. Stay Focused! The Effects of Internal and External Focus of Attention on Movement Automaticity in Patients with Stroke. PLOS ONE 10, e0136917 (2015).

209. Talvitie, U. Guidance strategies and motor modelling in physiotherapy. Physiother. Theory Pract. 12, 49–60 (1996).

210. Cirstea, M. C. & Levin, M. F. Improvement of Arm Movement Patterns and Endpoint Control Depends on Type of Feedback During Practice in Stroke Survivors. Neurorehabil. Neural Repair 21, 398–411 (2007).

211. Maulucci, R. A. & Eckhouse, R. H. Retraining reaching in chronic stroke with real-time auditory feedback. NeuroRehabilitation 16, 171–182 (2001).

212. Winstein, C. J., Merians, A. S. & Sullivan, K. J. Motor learning after unilateral brain damage. Neuropsychologia 37, 975–987 (1999).

80

213. Zavieh, M. K., Esfandiarpour, F., Rahmatipour, M. & Akbarzadeh, A. The effect of knowledge of results frequency on motor learning in individuals with stroke. J. Rehabil. Med. 3, 34–41 (2015).

214. Stanton, R., Ada, L., Dean, C. M. & Preston, E. Feedback Received While Practicing Everyday Activities During Rehabilitation After Stroke: An Observational Study: Feedback in Stroke Rehabilitation. Physiother. Res. Int. 20, 166–173 (2015).

215. Betschart, M., McFayden, B. J. & Nadeau, S. Repeated split-belt treadmill walking improved gait ability in individuals with chronic stroke: A pilot study. Physiother. Theory Pract. 0, 1–10 (2017).

216. States, R. A., Pappas, E. & Salem, Y. Overground Physical Therapy Gait Training for Chronic Stroke Patients With Mobility Deficits. Stroke 40, e627–e628 (2009).

217. Park, I., Lee, Y., Moon, B. & Sim, S. A comparison of the effects of overground gait training and treadmill gait training according to stroke patients’ gait velocity. J. Phys. Ther. Sci. 25, 379–382 (2013).

218. Adams, J. A. A Closed-Loop Theory of Motor Learning. J. Mot. Behav. 3, 111–150 (1971).

219. Young, D. E. & Schmidt, R. A. Augmented Kinematic Feedback for Motor Learning. J. Mot. Behav. 24, 261–273 (1992).

220. Patterson, K. K. Gait asymmetry post-stroke. (Department of Rehabilitation Science, University of Toronto, 2010).

221. Winstein, C. J. & Wolf, S. L. Task-oriented training to promote upper extremity recovery. Stein J Harvey RL Macko RF Winstein CJ Zorowitz Rd Stroke Recovery Rehabil. N. Y. Demos Med. 267–290 (2008).

222. Gowland, C. et al. Measuring physical impairment and disability with the Chedoke-McMaster Stroke Assessment. Stroke 24, 58–63 (1993).

223. Brott, T. et al. Measurements of acute cerebral infarction: a clinical examination scale. Stroke 20, 864–870 (1989).

224. Nasreddine, Z. S. et al. The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment. J. Am. Geriatr. Soc. 53, 695–699 (2005).

225. Dong, Y. et al. The Montreal Cognitive Assessment (MoCA) is superior to the Mini-Mental State Examination (MMSE) for the detection of vascular cognitive impairment after acute stroke. J. Neurol. Sci. 299, 15–18 (2010).

226. Kuys, S. S., Brauer, S. G. & Ada, L. Test-retest reliability of the GAITRite system in people with stroke undergoing rehabilitation. Disabil. Rehabil. 33, 1848–1853 (2011).

81

227. Cho, K. H., Lee, H. J. & Lee, W. H. Test–retest reliability of the GAITRite walkway system for the spatio-temporal gait parameters while dual-tasking in post-stroke patients. Disabil. Rehabil. 37, 512–516 (2015).

228. R Core Team. R: A Language and Environment for Statistical Computing. (R Foundation for Statistical Computing, 2017).

229. Portney, L. & Watkins, M. Single-Subject Designs. in Foundations of Clinical Research: Applications to Practice 223–264 (Prentice Hall Health, 2000).

230. Nourbakhsh, M. R. & Ottenbacher, K. J. The statistical analysis of single-subject data: a comparative examination. Phys. Ther. 74, 768–776 (1994).

231. Gottman, J. & Leiblum, S. How to do psychotherapy and how to evaluate it: A manual for beginners. (Holt, Rinehart & Winston, 1974).

232. Titianova, E. B. & Tarkka, I. M. Asymmetry in walking performance and postural sway in patients with chronic unilateral cerebral infarction. J. Rehabil. Res. Dev. 32, 236 (1995).

233. Crosby, L. et al. Is faster always better? A description of how temporal gait asymmetry changes with increased walking speed following stroke. in (2017).

234. Sackley, C. M. & Lincoln, N. B. Single blind randomized controlled trial of visual feedback after stroke: Effects on stance symmetry and function. Disabil. Rehabil. 19, 536–546 (1997).

235. Lee, M., Moseley, A. & Refshauge, K. Effect of Feedback on Learning a Vertebral Joint Mobilization Skill. Phys. Ther. 70, 97–102 (1990).

236. Kovacs, A. J. & Shea, C. H. The learning of 90° continuous relative phase with and without Lissajous feedback: External and internally generated bimanual coordination. Acta Psychol. (Amst.) 136, 311–320 (2011).

237. Patterson, S. L. Effect of treadmill exercise training on spatial and temporal gait parameters in subjects with chronic stroke: A preliminary report. J. Rehabil. Res. Dev. 45, 221–228 (2008).

238. Denotter, A., Geurts, A., Mulder, T. & Duysens, J. Gait recovery is not associated with changes in the temporal patterning of muscle activity during treadmill walking in patients with post-stroke hemiparesis. Clin. Neurophysiol. 117, 4–15 (2006).

239. Wu, Q., Chan, J. S. Y. & Yan, J. H. Mild cognitive impairment affects motor control and skill learning. Rev. Neurosci. 27, 197–217 (2016).

240. Ungerleider, L. Imaging Brain Plasticity during Motor Skill Learning. Neurobiol. Learn. Mem. 78, 553–564 (2002).

82

241. Alexander, L. D. et al. Association Between Gait Asymmetry and Brain Lesion Location in Stroke Patients. Stroke 40, 537–544 (2009).

242. Ward, N. S. & Cohen, L. G. Mechanisms underlying recovery of motor function after stroke. Arch. Neurol. 61, 1844–1848 (2004).

243. Ward, N. S. Neural correlates of motor recovery after stroke: a longitudinal fMRI study. Brain 126, 2476–2496 (2003).

244. Balasubramanian, C. K., Neptune, R. R. & Kautz, S. A. Variability in spatiotemporal step characteristics and its relationship to walking performance post-stroke. Gait Posture 29, 408–414 (2009).

245. Boyd, L. A., Vidoni, E. D. & Wessel, B. D. Motor learning after stroke: Is skill acquisition a prerequisite for contralesional neuroplastic change? Neurosci. Lett. 482, 21–25 (2010).

246. Borich, M. R., Brown, K. E. & Boyd, L. A. Motor Skill Learning Is Associated With Diffusion Characteristics of White Matter in Individuals With Chronic Stroke. J. Neurol. Phys. Ther. 38, 151–160 (2014).

247. Bonnyaud, C. et al. Does a Single Gait Training Session Performed Either Overground or on a Treadmill Induce Specific Short-Term Effects on Gait Parameters in Patients with Hemiparesis? A Randomized Controlled Study. Top. Stroke Rehabil. 20, 509–518 (2013).

248. Lewek, M. D., Feasel, J., Wentz, E., Brooks Jr, F. P. & Whitton, M. C. Use of visual and proprioceptive feedback to improve gait speed and spatiotemporal symmetry following chronic stroke: a case series. Phys. Ther. 92, 748–756 (2012).

83

Appendix A

General Instructions for Acquisition Session

Today you will be performing a series walking tasks. You will be asked to walk across the grey

mat on the floor, turn around at the marker on other side and walk back across the mat, if time

permits. You will be asked to perform this task in blocks of 30 seconds, up to 25 trials. You will

have a rest after each trial. Does this sound okay to you?

Display A

After every trial (for 100% condition, or every other trial for 50% condition), a tablet will be

shown to you which will show you the amount of time you spend on each leg when you were

walking. For example, if you spend equal amounts of time on both legs, then the marker will be

displayed in the center. If you spend more time on your right leg while walking, then the marker

will be on the right side of the screen. The more time you spent on the right leg the further to

the right the marker will be. If you spend only slightly more time on your right leg then the

marker will be closer to the center. The same is true for the left side.

This red band is the target area for your walking. A marker inside this area indicates an

improvement in your walking pattern, compared to when we first began. So after you get the

feedback, your goal is to try to walk so that the marker moves into the red band which is the

target area. The closer the marker is to the center of the display, the better. Would you mind

explaining it to me in your own words to make sure that we both understand?

84

Display B

After each trial, a tablet will be shown to you which will show you the amount of time you

spent on each leg while you were walking. Each bar is showing the amount of time spent on

your right and left legs. Or you can think of it as the amount of time spent with your right and

left foot on the ground.

The dark blue bar on the left represents the amount of time spent on your left leg (or foot)

while walking, and the light blue bar on the right side represents the amount of time spent on

your right leg (or foot) while walking. If the bar on the left is taller than the bar on the right,

then this means that you are spending more time on your left leg while you were walking. If the

right bar is much taller than the left bar, then you are spending a lot more time on your right

leg (or foot). The same is true if the right bar is taller.

Your goal is to make the right and left bars equal height, by evening out the amount of time

spent on each leg while walking. Would you mind explaining it to me in your own words to

make sure that we both understand?