<%BANNER%>

Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2012-12-31.

Permanent Link: http://ufdc.ufl.edu/UFE0042166/00001

Material Information

Title: Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2012-12-31.
Physical Description: Book
Language: english
Creator: Raja, Bhavana
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: Rehabilitation Science -- Dissertations, Academic -- UF
Genre: Rehabilitation Science thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Statement of Responsibility: by Bhavana Raja.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Kautz, Steven A.
Electronic Access: INACCESSIBLE UNTIL 2012-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0042166:00001

Permanent Link: http://ufdc.ufl.edu/UFE0042166/00001

Material Information

Title: Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2012-12-31.
Physical Description: Book
Language: english
Creator: Raja, Bhavana
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: Rehabilitation Science -- Dissertations, Academic -- UF
Genre: Rehabilitation Science thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Statement of Responsibility: by Bhavana Raja.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Kautz, Steven A.
Electronic Access: INACCESSIBLE UNTIL 2012-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0042166:00001


This item has the following downloads:


Full Text

PAGE 1

1 NON PARETIC LEG PERFORMANCE IN HEMIPARETIC WALKING: LOADING ASYMMETRIES, COMPENSATORY MECHANISMS AND RESPONSIVENESS TO LOCOMOTOR TRAINING By BHAVANA RAJA A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010

PAGE 2

2 2010 Bhavana Raja

PAGE 3

3 To my Parents who taught me how to dream and my Husband for helping me fulfill those dreams.

PAGE 4

4 ACKNOWLEDGMENTS I could have never achieved whatever I have today if it was not for the support and cooperation of a lot of people. I owe this all to them.First I would like to express my gratitude to Dr. Steve n Kautz, my advisor and mentor for the Ph D He gave me the freedom to explore on my own and taught me how to express my ideas free of any inhibitions. He led me into approaching a research problem in different ways and getting to the bottom of it. I cannot thank Dr. Kautz enough for guiding me all throughout even when my steps faltered. His mentorship was paramount in providing a well rounded experience. His patience and support has brought me to overcome many crisis situations and eve ntually achieving my final goal. I would like to thank my commi ttee members Dr. Andrea Behrman, Dr. Carolyn n Patten, Dr. Dena Howland and Dr. Lorie Richards for their assistance and feedback throughout the dissertation process. Heartfelt thanks to all members of my laboratory Ryan Knight Kelly Rooney Frank and Dr. C ameron Nott f o r helping me with data analyses, figures and Helen Emery for keeping me on my toes with the administrative tasks I would have otherwise missed. A special t hank s to Dr. Mark Bowden for his cooperation ever since I stepped in the University of Florida. His suggestions and constructive criticism helped me look at my work with a different perspective. He was always available to discuss with me my ideas and to proofread my papers, throwing light on where I needed improvement or re work. I would als o like to thank all my subjects who volunteered to participate in the research studies. None of this research would have been possible without their cooperation. I am extremely grateful to the University of Florida Alumni Association for granting me the Al umni Fellowship that funded my education for four years. I would also like to

PAGE 5

5 thank the faculty and staff in the Rehabilitation Science Doctoral Program and the Department of Physical therapy for their guidance from time to time. I would like to extend my sincere thanks to the fellow RSD students for cheering me up at all ti mes, the people from BRRC who were my support system and home away from home. I would also thanks a ll my friends in the US and India who kept me going in the rough times and walked with me every step of the way. My very special thanks goes to my Pa and Mommy to whom I owe everything I am today. Their unwavering faith and confidence in my abilities is what has shaped out of me the person I am at present. I do not want to miss this opport unity to thank them for showing me the worth of hard work and for letting m e follow my dreams even when they went beyond the boundaries of language and geography. I am extremely thankful to my sister Shruti and brother Siraj for their relentless bullying and efforts to keep m e rejuvenated in some of my most stressful times. I also would like to thank my sisters in law (Anjali, Shobhna and Amita) and brother in law Dr. Ashok Kumar for their support and encouraging ways. My sincere and loving thanks to Vipin my husband whose patience and unwavering love was undeniably the bedrock upon which the last four years of my life have been built. He followed me around the globe. He put his plans and ambitions aside so I could achieve mine. Through this journey I have shared with him times of frustration and fun, moments of tears and laughter. I have always been amazed with his ability to maintain his composure even during my most tumultuous moods. Thank y ou Vipin for your companionship! I would have never been here ha d it not been for your companionship.

PAGE 6

6 Last but not least, thank you dear God for giving me the strength and courage to go through this all the way with determination and sending my way all the amazing people.

PAGE 7

7 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ .......... 11 LIST OF FIGURES ................................ ................................ ................................ ........ 12 LIST OF ABBREVIATIONS ................................ ................................ ........................... 14 ABSTRACT ................................ ................................ ................................ ................... 15 CHAPTER 1 LITERATURE REVIEW ................................ ................................ .......................... 18 Introductory Statements ................................ ................................ .......................... 18 Stroke: Definition, Incidence ................................ ................................ ............. 20 Pathophysiology of Stroke ................................ ................................ ................ 20 Characteristics of Hemiparetic Gait ................................ ................................ ........ 21 Biomechanical Characteristics of Hemiparetic Gait ................................ ................ 21 Kinematic Characteristics ................................ ................................ ................. 22 Kinetic Characteristics ................................ ................................ ...................... 23 Ground reaction forces (GRF) ................................ ................................ .... 24 Joint moments and power ................................ ................................ .......... 27 Electromyography ................................ ................................ ............................ 29 Interlimb Coordination and Need for Studying the Non paretic Leg ........................ 30 Evidence in Animal Literature ................................ ................................ ........... 30 Evidence in Human Literature ................................ ................................ .......... 31 Evidence in upper extremity literature ................................ ........................ 31 Evidence in lower extremity ................................ ................................ ....... 32 Evidence of the Role of Non Paretic Leg in Walking ................................ ........ 34 Post stroke Rehabilitation ................................ ................................ ....................... 34 Traditional Gait Rehabilitation Therapies ................................ ......................... 34 Activity Based Gait Rehabilitation Therapies ................................ .................... 36 Evidence of neuroplasticity following therapy from animal literature .......... 36 Evidence from human literature ................................ ................................ 38 Relevance of the Reviewed Literature to the Studie s of the Dissertation ............... 41 2 RATIONALE FOR VARIABLE SELECTION AND ANALYSIS ................................ 47 Background ................................ ................................ ................................ ............. 47 Hemiparesis: Unilateral vs. Bilateral ................................ ................................ 47 Evidence of Bilateral Involvement in Upper Extremity ................................ ...... 48 Causes of Bilateral Limb Involvement Following a Unilateral Lesi on ................ 48 Goals of Studies of the Dissertation ................................ ................................ ........ 49

PAGE 8

8 Variables Used ................................ ................................ ................................ ....... 50 Study 1 ................................ ................................ ................................ ............. 50 Measure of limb loading and unloading ................................ ..................... 50 Timing of limb loading and limb unloading ................................ ................. 51 Pattern of limb loading and limb unloading ................................ ................ 53 Kinematic parameters ................................ ................................ ................ 54 Study 2 ................................ ................................ ................................ ............. 55 EMG timing variable ................................ ................................ ................... 56 EMG magnitude ................................ ................................ ......................... 60 Work Done ................................ ................................ ................................ 62 Definitions of Imp ortant Terms ................................ ................................ ......... 63 Compensation ................................ ................................ ............................ 63 Impairment ................................ ................................ ................................ 64 Analysis Plan ................................ ................................ ................................ .... 64 Hypotheses ................................ ................................ ................................ ...... 66 Approach to Interpretations ................................ ................................ .............. 73 Limitations of the analysis ................................ ................................ ................ 74 3 MAGNITUDE AND RATE OF LIMB LOADING AND UNLOADING DURING HEMIPARETIC GAIT IS ASYMMETRIC AND RELATED TO LEG KINEMATICS. ................................ ................................ ................................ ......... 78 Background ................................ ................................ ................................ ............. 78 Methods ................................ ................................ ................................ .................. 80 Participants ................................ ................................ ................................ ....... 80 Experimental Protocol ................................ ................................ ...................... 81 Study Variables ................................ ................................ ................................ 82 Magnitude of loading/unloading ................................ ................................ 82 Timing of loading/unloading ................................ ................................ ....... 82 Pattern of limb loading/unloading ................................ ............................... 82 Kinematics ................................ ................................ ................................ 83 Statistical Analys is ................................ ................................ ............................ 83 Results ................................ ................................ ................................ .................... 83 Magnitude of Limb Loading ................................ ................................ .............. 84 Magnitude of Limb Unloading ................................ ................................ ........... 84 Timing and Pattern of Limb Loading and Unloading ................................ ......... 85 Relationship between Medial Lateral Ground Reaction Forces and Limb Loading and Unloading ................................ ................................ ................. 87 Relationship bet ween Knee Angle and Limb Loading and Unloading .............. 87 Relationship between Medial Lateral Leg Angle and Limb Loading and Unloading ................................ ................................ ................................ ...... 88 Discussion ................................ ................................ ................................ .............. 88 4 COMPENSATION AND IMPAIRMENT PATTERNS OF THE NON PARETIC LEG IN HEMIPARETIC GAIT. ................................ ................................ ................ 97 Background ................................ ................................ ................................ ............. 97

PAGE 9

9 Methods ................................ ................................ ................................ .................. 99 Participants ................................ ................................ ................................ ....... 99 Procedures ................................ ................................ ................................ ..... 100 Subject preparation ................................ ................................ .................. 100 Data Collection ................................ ................................ ......................... 101 Data Recording and Processing ................................ .............................. 101 Calculations of Study Variables ................................ ................................ ...... 102 EMG Variables ................................ ................................ ......................... 103 Kinetic and kinematic variables ................................ ................................ 103 Statistical Analyses ................................ ................................ ........................ 103 Results ................................ ................................ ................................ .................. 104 Magnitude of EMG Act ivity ................................ ................................ ............. 104 Timing of EMG Activity ................................ ................................ ................... 105 Correlation Analysis ................................ ................................ ....................... 106 Discussion ................................ ................................ ................................ ............ 107 Muscle Activity over the Gait Cycle ................................ ................................ 108 Muscle Activity during Different Regions of the Gait Cycle ............................. 109 Implications of Analysis ................................ ................................ .................. 113 Cl inical Relevance ................................ ................................ .......................... 114 5 RESPONSE OF NON PARETIC LEG TO LOCOMOTOR TRAINING POST STROKE ................................ ................................ ................................ ............... 126 Background ................................ ................................ ................................ ........... 126 Methods ................................ ................................ ................................ ................ 128 Subject Demographics ................................ ................................ ................... 128 Subject Classification ................................ ................................ ..................... 129 Intervention ................................ ................................ ................................ ..... 129 Biomechanical Data Collection ................................ ................................ ....... 131 Data Recording and Processing ................................ ................................ ..... 132 Calculations of Study Variables ................................ ................................ ...... 133 EMG variables ................................ ................................ ......................... 133 Statistical Analyses ................................ ................................ ........................ 134 Results ................................ ................................ ................................ .................. 134 Magnitude of EMG Act ivity in the Entire Gait Cycle ................................ ........ 135 Magnitude EMG Activity in Different Regions of the Gait Cycle ..................... 135 Percentage Duration of Muscle Activity in the Entire Gait Cycle (Appendix B, table B 1) ................................ ................................ ................................ 137 Percentage Duration of Muscle Activity in Different Regions of the Gait Cycle ................................ ................................ ................................ ........... 137 Correlational Analysis ................................ ................................ ..................... 139 Discussion ................................ ................................ ................................ ............ 140 Muscle Activity over the Entire Gait Cycle ................................ ...................... 141 Muscle Activity in Different Regions of the Gait Cycle ................................ .... 142 6 CONCLUSION ................................ ................................ ................................ ...... 152

PAGE 10

10 Future Directions ................................ ................................ ................................ .. 155 Summary ................................ ................................ ................................ .............. 156 APPENDIX A MAGNITUDE OF INTEGRATED EMG ACTIVITY OVER THE ENTIRE GAIT CYCLE ................................ ................................ ................................ .................. 158 B TABLES FOR PERCENTAGE DURATION OF ACTIVITY BEFORE AND AFTER LOCOMOTOR TRAINING ................................ ................................ ....... 166 LIST OF REFERENCES ................................ ................................ ............................. 170 BIOGRAPH ICAL SKETCH ................................ ................................ .......................... 185

PAGE 11

11 LIST OF TABLES Table page 2 1 List of loading and unloading details of hemiparetic individuals ......................... 92 B 1 Represents the percentage duration of activity pre and post LT over the entire gait cycle in indiv iduals walking at different speeds.. .............................. 166 B 2 Represents percentage duration of activity of each muscle in different regions of the gait cycle in indi viduals walking at slow speed pre LT.. ............. 166 B 3 Represents percentage duration of activity of each muscle in different regions of the gait cycle in individuals walking at moderate speed pre LT.. ..... 167 B 4 Represents percentage duration of activity of each muscle in different regions of the gait cycle in individuals wa lking at slow speed post LT. ............. 167 B 5 Represents percentage duration of activity of each muscle in dif ferent regions of gait cycle in individuals wal king at slow speed pre LT and moderate speed post LT. ................................ ................................ ............... 167 B 6 Represents percentage duration of activity of each mus cle in different regions of gait cycle in individu als walking at moderate speed pre post LT. .. 168 B 7 Represents percentage duration of activity of each muscle in different regions of the gait cycle in individuals walking at fast speed post LT. ............. 168 B 8 Represents percentage duration of activity of each muscle in different regions of the gait cycle in healthy control individuals walking at slow speed 168 B 9 Represents percentage duration of activity of each mu scle in different regions of gait cycle in healthy control individuals walking at moderate speed 169 B 10 Represents percentage duration of activity of each muscle in different regions of the gait cycle in healthy control individ uals walking at fast speed ... 169

PAGE 12

12 LIST OF FIGURES Figure page 1 1 Represents different phases of the gait cycle. ................................ .................... 43 1 2 Represents hip, knee and ankle joint angles subject walking.. ........................... 44 1 3 Represents the ground reaction forces ................................ ............................... 45 1 4 Represents power profiles in hip, knee and ankle subject walking at self selected speed. ................................ ................................ ................................ .. 46 2 1 Represents the vertical GRF in a healthy control subject. ................................ .. 75 2 2 Represents the calculation of integrated EMG activity.. ................................ ..... 76 2 3 Represents the steps of k means analysis. s. ................................ .................... 77 3 1 Represents characteristics of one subject in each speed group.. ....................... 93 3 2 Represents average characteristics of all subjects in each speed group.. ......... 94 3 3 Repres ents individuals with different loading and unloading patterns against their self selected walking speed. t. ................................ ................................ .... 95 3 4 Represents the cross over point (circled in black) during paretic loading and unloading during walking. ................................ ................................ ................... 96 4 1 Represents the different regions of the gait cycle. ................................ ............ 115 4 2 Represents the calculation of integrated EMG activity.. ................................ ... 116 4 3 Represents the steps of k means analysis. ................................ ...................... 117 4 4 Integrated EMG activity for the entire gait cycle in the different groups is plotted to represent the magnitude of muscle activity. ................................ ...... 118 4 5 Represents intergrated EMG activity in different regions of the gait cycle in individuals in slow speed group. ................................ ................................ ....... 119 4 6 Represents intergrated EMG activity in different regions of the gait cycle in individuals in moderate speed group. ................................ ............................... 120 4 7 Represents intergrated EMG activity in different regions of the gait cycle in individuals in fast speed group. ................................ ................................ ........ 121 4 8 Represents percentage duration of EMG activity in different speed groups (slow, moderate and fast speed).. ................................ ................................ .... 122

PAGE 13

13 4 9 Represents percentage duration of EMG activity in different regions of the gait cycle in individuals in slow speed group. ................................ ................... 123 4 10 Represents percentage duration of EMG activity in different regions of the gait cycle in individuals in moderate speed group.. ................................ .......... 124 4 11 Represents percentage duration of EMG activity in different regions of the gait cycle in individuals in fast speed group.. ................................ .................... 125 5 1 Represents the different regions of the gait cycle.. ................................ ........... 146 5 2 Represents the calculation of integrated EMG activity.. ................................ ... 146 5 3 Represents the steps of k means analysis. ................................ ...................... 147 5 4 Represents difference in muscle activity in individuals walking at slow speed (0 <0.4m/s) before and after locomotor training. ................................ .............. 148 5 5 Represents difference in muscle activity in individuals wa lking at slow speed before and moderate spee d after locomotor training. ................................ ....... 149 5 6 Represents difference in muscle activity in individuals walking at moderate speed (0.4 0.8m/s ) before and after locomotor training. ................................ .. 150 5 7 Represents difference in muscle activity in individuals walking at moderate speed (0.4 0.8m/s) before and fast speed (> 0.8m/s) ................................ ..... 151 A 1 Integrated EMG activity in individuals walking at slow speed pre LT. ............... 158 A 2 Integrated EMG activity in individuals walking at moderate speed pre LT. ...... 159 A 3 Represents integrated EMG activity in different regions of the gait cycle in individuals walking at slow speed pre LT. ................................ ....................... 160 A 4 Represents integrated EMG activity in different regions of the gait cycle in individuals walking at slow speed pre and post LT. ................................ ......... 161 A 5 Represents integrated EMG activity in different regions of the gait cycle in individuals walking at slow speed pre and moderate speed post LT. ............. 162 A 6 Represents integrated EMG activity in different regions of the gait cycle in individuals walking at moderate speed pre LT. ................................ ................ 163 A 7 Represents integrated EMG activity in different regions of the gait cycle in individuals walking at moderate speed pre and post LT. ................................ 164 A 8 Represents integrated EMG activity in different regions of the gait cycle in individuals walking at moderate speed pre LT and at fast speed post LT.. ..... 165

PAGE 14

14 LIST OF ABBREVIATION S BWS Body Weight Support BWSTT Body weigh t supported treadmill training CNS Central Nervous System CPG Central Pattern Generator EMG Electromyography GM Gluteus Medius GRF Ground Reaction Forces LL/LU Limb loading and unloading LH Lateral Hamstrings LT Locomotor Training MG Medial Gastrocnemius MH Medial Hamstrings NDT Neuro Development Therapy PNF Propioceptive Neuromuscular Facilitation RF Rectus Femoris SCI Spinal Cord Injury SO Soleus TA Tibiablis Anterior VM Vastus Medialis

PAGE 15

15 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy NON PARETIC LEG PERFORMANCE IN HEMIPARETIC WALKING: LOADING ASYMMETRIES, COMPENSATORY MECHANISMS AND RESPONSIVENESS TO LOCOMOTOR TRAINING By Bhavana Raja December 2010 Chair: Steven A. Kautz Major: Rehabilitation S cience Walking requires appropriate integration of sensorimotor input to coordinat e both the legs. Altered walking post stroke is traditionally considered a unilateral motor control problem of the paretic leg. The non paretic leg is generally considered unaffected and left unatten ded during rehabilitation. The purpose of this dissertation was to understand the performance of the non paretic leg during walking. To this end t hree studies were conducted. We investigated loading asymmetries (a major source of afferent input) quantified compensatory mechanisms (a major component of motor output) and investigated the responsiveness of those compensatory mechanisms to locomotor training Participants with chronic hemiparesis and similarly aged healthy control su bjects walked over an instrumented (with force plates) treadmill as kinematic, kinetic and electromyographic data were collected. In the first study, limb loading and unloading during walking was quantified. Results of the analysis revealed limb loading a nd unloading was asymmetric between the paretic and the non paretic leg and significantly different from healthy control subjects walking at matched speeds. Limb loading and unloading was related to knee

PAGE 16

16 angle and leg angle. I ndividuals post stroke make se veral adjustments in the loading and unloading pattern to maintain a steady walking state. We suggest promoting symmetry in limb loading and unloading while training kinematic changes might promote better gait pattern. In the second study, the changes in the motor pattern of the non paretic leg were quantified and evaluated as to whether they represented compensatory mechanisms or impairments. Results demonstrate d that the muscle activity of the non paretic leg differ ed from healthy control leg walking at matched speeds and that most of the changes appear ed to be compensatory in nature. Some of these compensations are simply increased output of the usual dominant power bursts of a healthy leg, while others are novel adaptations in the non paretic leg that a ppear to aid in the production of additional propulsion by the paretic leg. In the third study of this dissertation, we investigate d the response of the non paretic leg motor pattern to a task specific locomotor training i.e. body weight supported treadmill training Walking speed was increased post training and the increase in walking speed was associated with significant changes in the non paretic leg muscle activity. The results of our analysis revealed mixed response to LT post stroke. The muscle activity of non paretic leg was significantly normalized in some individuals walking post LT. Nevertheless, there were some significant i ncreases in the muscle activity in the non paretic leg post LT when compared to healthy control subjects walking at matched speeds. Most notably participants that walked at slower speeds before LT, but acquired moderate walking speed post LT exhibited the greatest increases in muscle activity This suggests that, in addition to the changes that were

PAGE 17

17 occurring in the paretic leg, the non paretic leg appears to have contributed to the increased walking speed through increases in one or more compensatory mech anisms.

PAGE 18

18 CHAPTER 1 LITERATURE REVIEW Introduct ory Statements In the United States approximately 795000 new or recurrent strokes occur each year, with a survivi ng cohort of 6.5 million people 1 Stroke is the prima ry cause of adult disability 2, 3 The residual functioning deficits involve cognitive, communicative, visual and sensorimotor systems 4 Motor impairments, characterized by hemiparesis o r the weakness of one half of the body are reported in 75% of individuals after strok e 4 The sensorimotor deficits include, weakness, impaired selec tive motor control, spasticity, and proprioceptive deficits that interfere with normal adult walking 5 Inability to walk independently after stroke is also associated with poor coordination and inadequate proprioceptive input to both the legs. Proprioc eption from 6 Therefore, they facilitate balance during standing, quasi static positions and walking. Loss of proprioception after stroke might result in asymmetric loading and unloading. While asymmetry in loading and unloading is well documented in static and quasi s tatic tasks 7 10 there is littl e information about asymmetry in loading during walking. One of the primary reasons regarding the importance of investigating asymmetric loading during walking is because it is characterized by ac celeration of the body segments 7 Thus, unlike static and quasi static postures the cent er of mass is mostly outside the base of support, which makes walking more unstable. Furthe rmore, walking is a cyclic task in which loading of one limb is simultaneously associated with the unloading of the other. Nonetheless, there is no specific information about loading of paretic and non paretic leg during steady state walking in individuals post stroke.

PAGE 19

19 Alterations in walking ability post stroke are primarily considered to be related to paretic leg performance. Therefore, most of the rehabilitation and research studies focus on the paretic leg, while assuming the non paretic leg i s unaffecte d; considering it to primarily be compensating for the impairments of the paretic leg. Recent research studies suggest performance is poor in bilateral task as oppose d to unilateral tasks For example, pedaling studies suggested the deteriorated bilateral performance could be attributed to the detrimental influence of the non paretic leg on the paretic leg is performance 11 This strongly indicates bilateral involvement post stroke and therefore the need to assess and treat both the paretic and the non paretic side, to facilitate better performance post stroke. With our increase d understanding of the neural mechanisms of disease an d recovery there has been increased information related to the role of bilateral neural structures and the potential effects of unilateral injury on limbs of both sides is better understood Several recent studies have emphasized the importance of includin g the non paretic leg in the evaluation and rehabilitation post stroke 12, 13 but no study has specifically evaluated the non paretic leg, or reported its response to task specific rehabilitation training. The literature review of this dissertation provides the background information relevant to the specific ai ms of this dissertation. It first broadly discusses alterations in hemiparetic gait in individuals walking after stroke. S econd ly it presents the importance of limb loading (LL) during walking. Third, it overviews the information related to the involvement of non paretic extremity after stroke, relevant to the second study. Finally it describes literature relevant to walking rehabilitation after stroke, the traditional and

PAGE 20

20 modern day concepts with emphasis on the task specific locomotor training (relevant t o study 3). Stroke: Definition, I ncidence Definition. Stroke is defined as a condition characterized by rapidly developing symptoms and signs of a focal brain lesion, with symptoms lasting for more than 24 hours or interrupted by death, with no apparent c ause other than of vascular origin 14 Incidence and s queal Stroke is one of the most debilitating central nervous system conditions resulting in long term disability. Every year approximately 795,000 people are affected by stroke with a surviving cohort of 6.1 million individuals. 1 Over 75% of people affected by stroke have limitation in m obility related to walking 4 Furthermore, only less than 50% of individuals achieve independent community walking status 15 Of those individuals who recover the ability to walk many are still disabled by slow walking speed s and limited endurance. Residual motor weakness, abnormal movement synergies and spasticity result in altered gait patterns and contribute to poor balance, risk for falls and increased energy expenditure 16 Most of the research and clinical work identifies the impairments in walking after stroke as unilateral motor control problems of the paretic leg; however, a detailed investigation suggests that there is bilateral involvement of the limbs after stroke, as discussed in the studies of this dissertation Pathophysiology of S troke The pathophysio logical basis of stroke leading to various impairments can be either occlusive (due to obstruction of a blood vessel by a thrombus or an embolus) or hemorrhagic (due to rupture of a blood vessel) usually on one side of the brain. Insufficiency of blood sup ply is termed as ischemia which deprives brain tissue of

PAGE 21

21 oxygen and glucose and prevents the removal of toxic metabolites. The decrease in blood flow results in the death of neurons and other cellular elements. Hemorrhagic stroke may occur at the brain sur face or intraparenchymal causing a blood clot or haematoma within cerebral hemisphere or brainstem or in the cerebellum. The types and degrees of disability that follow a stroke depend upon multiple factors such as location and size of brain lesion, severi ty of the lesion, degree of spontaneous recovery, and th e duration of stroke onset 17, 18 Characteristics of Hemiparetic Gait Phases of the Gait Cycle Walking function may be quantified in terms of a g ait cycle. A g ait (Stride) cycle is defined as the events occurring from foot strike of one limb to the foot strike of the same limb 19 Each gait cycle is primarily divided into a stance phase (as when one limb is on the ground) and swing phase (as when the same limb is swinging wit h no contact on the ground) 19 Stance and Swing phases of a gait cycle can be further sub divided into different phases. Figure 1 1 presents the detailed description of the sub divisions of a gait cycle. For the purpose of the studies of this di ssertation these phases of gait cycle have been identified as different regions, also identified in figure 1 1. Biomechanical Characteristics of H emiparetic G ait Hemiparesis following stroke is characterized by a multitude of sensorimotor problems, marked by weakness, impaired selective motor control, spasticity and proprioceptive deficits that interfere with invariant features of normal adult gait 5 either home or an assisted living environment. Therefore, ability to walk independently with efficient velocity and endurance is one of the major goals for rehabilitation after

PAGE 22

22 stroke 20 In order to successful ly provide walking rehabilitation, it is important to Biomechanical measures provide an opportunity to investigate the control processes during walking post stroke. This section aims to review the biomechanical characteristics (kinematics, kinetics and EMG) of hemiparetic gait with focus on both the paretic and non paretic leg characteristics. Kinematic C haracteristics Kinematic variables include linear a nd angular positions, their displacements and time derivatives 21 Individuals post stroke display altered kinematics in sagittal plane with peak joint displacements or excursions at the hip, knee and ankle 2 2 24 in both the legs altered relative to healthy subjects walking Common kinematic deviations of the paretic lower limb in the sagittal plane include ankle plantar flexion being increased at initial contact 12, 24 and mid swing and decreased at toe off 25 (figure 1 2a). In addition, paretic leg knee flexion is increased at initial contact 25 and reduced during swing phase 26, 27 (figure 1 2b) while peak paretic hip extension is reduced at toe off and peak paretic hip flexion is increased during swing phase as compared to healthy similarly aged individuals walking at self selected or matched speeds 25, 28 (figure 1 2c). In the horizontal plane, individuals with hemiparesis demo nstrate abnormally large external rotation at the hip joint on the paretic leg throughout the gait cycle 26 Furthermore, some a uthors reported an abnormal abduction and inversion at the paretic hip and ankle in the frontal plane in individuals post stroke 12 In addition to those in the paretic leg, some abnormal joint motions have also been reported on the non paretic side. Hip exten sion at toe off is reduced (figure 1 2c) and knee flexion is maintained at 15 30 degrees during the stance phase and reaches a

PAGE 23

23 peak during the swing phase that is greater than that of the paretic leg ( approx. 20 degrees) and the healthy controls ( 35 43 degrees) 29 (figu re 1 2b). The kinematic variables, primarily the knee and ankle angle have been employed by two studies to group the participants based on the similarity of their characteristics 29, 30 Identification of patterns of gait dysfunction for patients following stroke have be en used to construct a classification system to guide intervention. Relation of the gait patterns to the underlying impairments would allow clinicians to target rehabilitation strategies to 30 De Quervain et al 29 used observational gait analysis and Mulroy et al 30 used non hierarchial cluster analysis technique for this purpose. The groups defined by M ulroy et al 30 classified individuals post stroke into four groups based on their speed and knee and ankle angle characteristics. Individuals in group 1 were fast walkers characterized by decreased knee extension in mid stance and adequate dorsiflexion in swing. Group 2 had intermediate velocity with greater knee flexion in mid stance (as compared to group 1) and adequate dorsiflexion in swing, Group 3 had a very slow velocity and excessive knee flexion in mid stance and inadequate dorsiflexion in swing, while group 4 also had a very slow gait velocity with a motion pattern characterized by knee hyperextension in mid stance and inadequate dorsiflexion in swing 30 Kinetic Characteristics Kinetic variables include forces, joint moments, powers and work. They are a direct consequence of joint moments generated by the muscles about a joint 31 These variables cannot be measured directly (except forces), but can be calculated using the principles of inverse dynamics 32 More than one type of kinematic and kinetic gait pattern have been identified across subjects post stroke, indicating that persons with

PAGE 24

24 stroke use different strategi es to achieve the goal of walking 12 Furthermore, kinetic variables explain the cause of the movement pattern observed at different joints. Ground reaction forces (GRF) of motion, which states that for every action there is an equal and opposite reaction 33 In the study of human motion, this law indicates that all surfaces provide a reaction force and that the individual is acted upon by that force when he/she is in contact with that surface 33 In human motion analysis, GRF is measured as a vector. This vector comprises of three orthogonal components: the anterio r posterior, the medial lateral and the vertical component (figure 1 3a). Extrapolation of these forces in the proximal and distal direction by employing a free body diagram allows for the calculation of subsequent joint reaction forces and is the basis for the inverse dynamics approach of human motion analysis 33 These calculations establish bases for the determination of joint moments, power and thus work, as discussed later. Vertical ground reaction forces Vertical GRF have been used to assess the static tasks 7, 9, 10 In normal human locomotion, a typical vertical GRF curve is an M shaped, double peaked curve separated by a valley However, following a stroke the M shaped pattern is distorted (figure 1 3b). Three different patterns of vertical GRF are typically observed in the paretic leg: doub le peaked similar to the normal M shap ed curve, single peak in mid stance and a plateau with no discernable peak 34, 35 The measures derived from GRF provide a means of qua ntifying bilateral symmetry of LL and limb unloading ( LU ) Individuals with hemiparesis are characterized by asymmetric LL and LU Hemiparetic subjects do not put as much weight on their

PAGE 25

25 paretic leg as on the non paretic leg 36 41 Since walking requires th e full body weight to be borne on either leg du ring single limb support phase, therefore, learning t o load and unload the paretic leg is important for balance and gait training in stroke patients 42 Studies have shown that individuals with stroke have difficulty leaning their body as fa r as possible in all planes. 43, 44 In addition, when shifting weight from a two legged stance to one legged stance or when stepping on stairs of various heights stroke patients show most difficulty in transferring weight to the paretic leg 7, 45, 46 Some studies have reported that the loading of the non paretic leg is difficult too, which could be explained as either due to subtle neuromuscular impairments ipsilateral to the brain lesion or to a reduced ability to control weight shift toward the non paretic side using the leg and hip muscles of the paretic side 10, 37, 45 This suggests that the speed and precision of the weight bearing of both the paretic and the non paretic leg are affected by stroke and thus the stroke patients need more time to shift weight to the paretic leg than to the non paretic leg 8 LL asymmetry and difficulties with actively redistributing weight while standing ha ve been consider ed major contributors to disordered hemiparetic gait 35, 47, 48 The ability to initiate an d control voluntary weight shifts toward either leg is a pre requisite for independent walking. Hence, the prevailing clinical assumption suggests practicing self generated weight shifts in the frontal plane within the base of support would be essential ab ility to train and monitor these patients 39, 46, 49 However, a recent investigation did not support the view point that a reduction in the LL asymmetry during static or quasi static posture would result in improved hemiparetic gait. They concluded that the people, who want to learn to wal k, must be trained to walk 40 Therefore, the treatment strategies

PAGE 26

26 based on the static analysis of posture and balance function may not be appropriate for dynamic activities, like walking, in which significant acceleration of body segments are occurring. Similarly, gait analysis may provide clinicians with an improved understanding of specific probl ems that cause instability and falls. This is the underlying viewpoint of the first study, in which we aim to investigate LL and LU and its relation to various walking parameters post stroke. Neurophysiology of LL and LU. The s ignificance of LL and LU has been studied and reported both in animal and human literature. LL has been regarded as an important afferent input which assists in generating a stepping pattern in absence of supraspinal input 50, 51 These afferent input s are interpreted in the lumbar motorneuron pool in humans 52 Furthermore, Dobkin et al 51 and Dietz et al 53, 54 suggested that the human lumbosacral spinal cord can generate a step like oscillating EMG pattern when adequate and appropriate sensory information associated with weight bearing is provided. In addition, Harkema et al 52 showed that t he EMG activity in the lower leg muscles of spinal cord injury (SCI) person is affected by the lev el of the load carried on the soles during treadmill stepping, implying that the mean EMG amplitude is related to peak load rather than the peak stretch/ stretch loading. This suggests that the LL and LU are important components of walking and therefore, n eed to be assessed and quantified adequately during walking to facilitate better understanding of the hemiparetic gait, which might help us design better rehabilitation techniques. Non neurological changes due to asymmetric loading in individuals with hemi paresis after stroke. Hemiparesis following stroke is not a unilateral motor control problem as often considered and reported, but affects both the paretic and the non

PAGE 27

27 paretic sides The difference in the performance of the paretic and the non paretic limb can be either : 1) functional (or compensatory) 2) due to the neurological damage caused by the stroke or 3) structural or musculo skeletal, due to disuse or the asymmetrical use of the paretic and the non paretic legs. Asymmetry is one of the most impor tant characteristic features of individuals walking post stroke. Additionally, asymmetry in LL is also reported in individuals with amputations, poliomyelitis etc. There is ample evidence in individuals post amputation that suggests osteo arthritis of the joints in the intact limb (e.g. knee 55 and hip 56 ) because of the changes in the direction of the ground reaction forces and greater LL 57 Similarly, subjects with poliomyelitis have asymmetrical disease, with greater forces transmitted across the unaffected limb resulting in greater symptoms on the unaffected leg as compared to the affected leg 58 Similarly, we hypothesize in the first study of this dissertation that there will be increased magnitude of LL on the non paretic leg post stroke, which could thus result in some biomechanical and structural changes. Thereafter, in subsequent studies we will investigate the performance of the non paretic leg and importance of including the non paretic leg as a part of the rehabilitation regime post stroke to ach ieve maximum benefit. Joint moments and power The magnitude of power bursts are smaller on the paretic side as compared to the non paretic side, and smaller in both the limbs as compared to healthy subjects walking at their comfortable speed 23 Recent research suggests that the power gen erated at the hip and ankle joints in the paretic and the non paretic leg in individuals with hemiparesis is significantly related to gait speed 12, 13, 59 Additionally, there are some abnormal patterns of power

PAGE 28

28 genera tion in the paretic and non paretic legs of individuals with stroke (figure 1 4a, b, c). In particular, peak power 12, 13 and positive work 60 of the hip (pull off) and ankle (push off) and negative knee power during pre swing and initial swing on both sides have been shown to be positively related to gait speed in subjects with stroke. Furthermore, Kim et al 12 rep orted an extra power burst in the frontal plane at the hip joint; and a very different pattern of power generation at the knee in the transverse plane and ankle in the frontal plane of the paretic leg as compared to the control subjects. These studies do r eport the abnormalities in the power generation or work done in the non paretic leg but do not consider either the influence of walking speed 60 or do not use a precise quantification method 12 In a seminal work Olney et al 60 reported that the non paretic limb is responsible for performing a greater proportion of work at all speeds, but the work done by the non paretic leg is less than the work done by the healthy contro ls at their self selected speeds. Furthermore, more negative work is done by the non paretic knee in the later part of stance phase of walking. This indicates that the energy absorption occurs in the knee extensors, while the knee is flexing 12 Furthermore, Olne y et al 13 demonstrated in their regression model that the ankle and hip joint positive power of the non pa retic leg were significantly related to walking speed 12 The inverse dynamics analysis reflects only the net result of active and passive muscle forces acting around the joint. However, it is noteworthy that the contribution from the passive structures (non musc le fibers) around the paretic ankle joint in hemiparetic individuals is much greater than the healthy controls (2 49 % as opposed to 5.9 to 6.5%) 61 The increased passive stiffness may compensate for the reduced

PAGE 29

29 moment output due to the weakness of the paretic ankle 61 Therefore kinetic variables alone do not provide detailed insight into the underlying pathophysiological mechanism in locomotor function post stroke, but kinetic analysis in combination with the electrophysiological techniques e.g. electromyography ( EMG ) may be abl e to provide greater understanding of the hemiparetic gait 62 Electromyography Walking involves a series of tasks which are accomplished by muscle activity of adequate magnitude and duration which can be recorded as electrical signals from e lectrodes placed on the skin. This technique of evaluating and recording the electrical activity produced by muscle is called the EMG 63 However, following a stroke the neuromuscular control of walking undergoes radical changes which include muscle weakness on the body side opposite to the lesion resulting in changes in spatial and temporal organization of muscle activity 64 One of the most important impairments after c entral nervous system injury is the inability to generate normal levels of muscle force 65 This lack of force generation influences the success of the individuals trying to accomplish many tasks, like walking. Gait in subjects after stroke is associated with abnormal muscle activation patterns (magnitude and timing). However there are large inter individual differ ences in the EMG of individuals post stroke. The commonly reported deviations include; a. premature activity of the paretic calf muscles during preswing and initial stance 66, 67 b. prolonged stance activity of the hamstrings and quadriceps muscles in both the paretic 24, 68, 69 and the non paretic leg 69, 70 c. the absence of activity in the paretic ankle dorsiflexors during the late swing and early stance phase 66

PAGE 30

30 In a seminal work relating to EMG in individuals post stroke Knutsson and Richards 24 provided an objective criterion for identification of muscle activation patterns during gait They attempted to classify EMG patterns across subjects with stroke into 3 types based on the examination of timing and amplitude of the muscle activation on the paretic side. Type I (spastic) was characterized by a premature activation of the calf muscle during the stance phase. Type II (paretic) showed abnormally low levels of activation in most muscle groups and Type III displayed co activation that spread to several or all muscle groups. Thereafter, Shiavi et al 69 applied this classification to both the paretic and the non paretic side post stroke. They examined bilateral changes in EMG activation taking place with motor recovery post stroke from 1 to 10 weeks, with follow up ranging from 6 24 months. They reported an abnormal synergy exhibited by the contralateral limb (non paretic) was almost always a co contraction. Interlimb Coordination and Need for Studying the Non paretic Leg Humans and other primate an d non primate animals are capable of coordinating various limb movements to accomplish a task. A remarkable spatio temporal coordination is evident in spite of the large differences in the inertial characteristics of the effectors involved. This suggests t he existence of some basic coordination principles that apply across widely different cooperative ensembles 71 Evidence in Animal Literature Interlimb coordination is one of the principal components of effective locomotion. This suggests that the movements of the limbs producing gait must be linked to each other to maintain equilibrium. Evidence from animal literature suggests that the contract ion of muscles on one side of the body leads to an increase in excitability of the contrala teral homologous motor pathways 72 For example, Rossignol et al 72

PAGE 31

31 demonstrated in spinal cats that continuously imposed flexion on one hind limb block ed the rhythm on both sides, suggesting that the rhythmic activity in one limb is i nfluenced by the manipulation of the other leg. This conveys that the spinal cord contains some mechanisms for interlimb coupling. Similarly, the reduction of input on one side produces deficits on the contralateral side. For example, Guiliani et al 73 showed unilateral deafferentation in spinal cats disrupted both ipsilateral and contralateral stepping. In addition to the spinal cord, brainstem and other sub cortica l structures are believed to play an important role in coordination. In experiments in cats with hemisected brainstem it was shown that when one hind limb is prevented from moving, the other hind limb continued its stepping behavior while rhythmic alternat ing bursts persisted in the stationary leg 74, 75 This suggests that the influence of sensorimotor state of one limb over the other and signifies the role of the spinal an d the supra spinal structures in walking. Evidence in Human Literature Evidence in u pper e xtremity l iterature Following unilateral stroke, neurological symptoms are mainly expressed on the side contralateral to the damaged hemisphere. However, careful inve stigation reveals that the side ipsilateral to the lesion also shows various deficits, such as impaired manual dexterity 76 78 muscular weakness 79 81 and reduction of sensory motor function 76, 79, 82 Significant differences have also been reported between healthy control and non paretic hand for fine and gross manual dexterity, motor coordination, global performance for two of four tasks, and thumb kinesthesia 76 Colebatch and Gandevia 79 found that the unilateral damage to motor cortex or its projection produced weakness in the i psilesional limbs. Ipsilateral deficits in hemiparetic

PAGE 32

32 subjects might be due to the interruption of the ipsilateral projection of the corticospinal tract. Although most of the corticospinal fibers decussate in the medulla, a significant proportion remains uncrossed and forms the ventral corticospinal tract 83 In addition, it is possible that a lesion in one hemisphere resulting from a vascular cause interrupts corticobulbar and corticoreticular projections and consequently affects subcor tical structures involved in motor control 84 Therefore, this suggests that the integrity of these descending pathways is necessary to achieve some motor performance on the non paretic side. Evidence in l ower e xtremity The regulation of human walking requires a close coordination of muscle activation between the two legs. During the stance and swing phase of gait, both legs act in a cooperative manner, each limb affects the strength of muscle activation and the time space behaviors of the other 85 Moreover, from a functional perspective interlimb 86 Recently, Dietz et al 87 demonstrated that unilateral stepping movements can lead to a patterned activation of the contralateral but rhythmicall y loaded leg in individuals post SCI They suggested that this influence of locomotor activity on the contralateral static leg is based on the interlimb coordination. However, the patterned activity in the non moving leg was mostly restricted to the leg flexor muscles. 87 The preserved activation of the leg flexo r, but strongly reduced EMG activity in the leg extensor muscles was suggested to be due to the well established differential control of leg flexor activity 85, 88 During normal locomotion the leg extensor activity is modulated continuously by proprioceptive feedback during the stance phase. Therefore, in the

PAGE 33

33 condition mentioned above, with a lack o f roll off of the body over the standing leg the EMG activity is expected to be diminished 86 In addition to the evidence in the spinal cord literature, interesting evidence exists regarding the role of the non paretic leg, post stroke in coordination during pedaling 11, 89 The pedaling literature post stroke suggests that some level of impaired motor control in the paretic leg during bilateral pedaling is due to synchronous motor activity of the non paretic leg and the interlimb afference associated with the non paretic LL and movement 89 It was also suggested that due to disruption of the normal task dependent suppression of interlimb influences, the paretic moto r pattern was more impaired in bilateral pedaling than in unilateral pedaling. The greater impairment of the paretic leg output could be attributed to the coupling of pattern generation of the two legs 89 Furthermore the unilateral non paretic leg pedaling activated a rhythmic locomotor pattern in the paretic leg of subjects with severe hemiparesis, whereas unilateral pedaling in non disabled control subjects did not activate a contralateral, rhythmic locomotor pattern 89 Additionally, the induced pattern in the paretic leg was nearly identical to the pattern observed in bilateral pedaling in the hemiparetic subjects. Ther efore, they concluded that probably the interlimb influences are substantially responsible for the impaired motor pattern of the paretic leg during pedaling 89 Thus, extrapolation of these results suggests that the therapeutic interventions targeting alterations of the sensorimotor state of the non paretic l eg may be highly effective in improving the motor output pattern of the paretic leg and thus facilitate better walking pattern.

PAGE 34

34 Evidence of the R ole of N on P aretic L eg in W alking The role of the non paretic leg has been identified in achieving faster speed in individuals walking post stroke. Kim et al 12 found a high correlation between the gait speed and the non paretic leg saggita l plane kinetics, particularly hip flexors at pull off and ankle plantarflexors during stance and push off. Similar results were supp orted by the regression models of Olney et al 13 which included hip and ankle power variables from both th e paretic and the non paretic side. Parvatneni et al 59 also demonstrated a significant correlation of the non paretic hip e xtensor power in individuals walking faster after stroke. Several other studies have shown that the muscle strength in the non paretic leg is highly correlated with gait speed 90, 91 These studies support the contribution of the non paretic leg performance in the gait of persons with stroke. The high correlation between the gait speed and selected variab les non paretic leg, the performance of the non paretic leg itself may be an important contributing factor to walking ability and justify the inclusion of the non paretic leg in training 12 Therefore, the intervention should not focus on training any s ingle weak link, but should target the parameters on both the sides. Therefore, the evidence from the above mentioned studies strongly supports the idea and need to study the non paretic leg in individuals walking post stroke. Post stroke Rehabilitation Tr aditional Gait Rehabilitation Therapies Until the emergence of the evidence of the spinal modulation of walking pattern in humans, rehabilitation of walking post stroke followed the hierarchical model of motor control 92 This theoretical model hypothesizes that the central nervous system (C NS ) is hard wired and irreparable. Furthermore, an injury to the CNS was deemed permanent

PAGE 35

35 and irreversible. Therefore, the rehabilitation strategies relied on compensatory strategies for non remediable impairme nts and deficits. The aim of rehabilitation was thus to achieve successful completion of the task while the quality of movement was a secondary aim. In late 1950s and 19 60s neurofacilitation techniques were developed. These techniques primarily focused on retraining motor control by facilitation or inhibition of movement patterns. According to these techniques, desirable movement patterns should be facilitated while masse d patterns or synergies should be inhibited 93 According to proponents of these techniques, the stereo typical movements result from a neurological disease and are not functional. Neurofacilitation techniques include Bobath approach, the Brunnstorm approach, and Propioceptive Neuromuscular Facilitation (PNF) The Neurodevelopmental Technique (NDT) establish ed by Bobath 94 assumes that an abnormal postural reflex activity is the m ajor cause of dysfunction, and therefore a significant proportion of therapy time involves inhibiting spasticity and other abnormal responses. In the Brunnstrom technique, synergistic movements are used to strengthen and practice single movements 95 PNF techniques consist of assisted isometric and isotonic leg flexion extension exercises, which are thought to improve strength and control of leg musculature in pr eparation for walking. 96 If a patient is unable to successfully maintain/ achieve a posture, the therapist would provide/instruct to use assistive devices and braces or suggest greater reliance on the non paretic side. These strategies would assist the individual in resuming almost pre injury levels of mobility

PAGE 36

36 using remaining physical strengths and external aids, but is not equivalent to recovery or restor ation of pre injury capabilities 97 Activity Based Gait Rehabilitation Therapies Over recent years, there has been a shift in focus from the nervous system being hard wired to it being malleable 98 This property of the nervous system led to the unders tanding of the ability of the neural systems to undergo plastic changes (neuroplasticity). Neuroplasticity is the mechanism by which the brain encodes experience and learns new behaviors. It is also the mechanism by which the damaged brain relearns lost be havior in response to rehabilitation, referred to as activity dependent plasticity 99 This ability of the nervous system to change and learn presents strong potential for the resumption of functional activities after a nervous system injury if the training conditions are optimized or are task specific, thus p roviding rehabilitation strategies alternative to the traditional approach. Evidence of neuroplasticity following therapy from animal literature Plautz et al. 100 trained adult squirrel monkeys to perform a motor task th at required pellet retrievals from a well (small/ large) to induce repetitive use of a limited set of distal forelimb movements. Detailed analysis of the motor behavior of the monkeys indicated that when monkeys were trained to retrieve food pellets from a small well, requiring skilled use of the digits, progressive increments in motor performance were seen. On the contrary, monkeys trained to retrieve pellets from a larger food well did not acquire any additional skill. Comparisons between pre training and post training maps of cortical movement representations revealed no task related changes in the cortical area devoted to the hand in the large well group. Movement specific changes in the small well group were large and consistent, and corresponded to the actual movement

PAGE 37

37 kinematics of the task. 100 They concluded that repetitive motor activity alone does not produce functional reorganization of cortical maps. However, skilled activity practice is a prerequisite factor in driving representational plasticity in motor cortex 100 Furthermore, several critical insights into the mammalian neural mechanisms controlling stepping have also been made. It was discovered that the adult cats with low thoracic spinal transection could recover hindlimb stepping by walking on a moving treadmill belt when given trunk support, assistance in paw placement and optimal sensory input 50 Furthermore, Forssberg et al 101, 102 and Grillner et al. 103 showed that the kinematics and EMG of spinal kittens were similar to normal cats and the spinalized kittens possessed the ability to adapt to various speeds of the treadmill. This emphasizes the importance of the peripheral afferent signals in adapting the locomotor pattern to external conditions. Along these lines of work it was shown that the age of spinalization and the amount of training in the treadmill had important effects in the locomotor pattern 104, 105 Animals spinalized at 2 weeks of age had a much better locomotor performance than those spinalized at 12 weeks of age. Training in 12 week old cats had an observable effect on weight bearing during locomotion. On the other hand Barbeau and Rossignol 50 demonstrated that appropriate training strategies (duration and intensity) could produce stepping recovery in adult cats as well. Adult cats spinalized at a lower thoracic level were trained in the treadmill environment two/three t imes per week. Over periods of several weeks up to one year of training deficits in balance and volitional movements remained. However, cats remarkably exhibited the ability to maintain hindquarter weight support and to generate rhythmic, coordinated stepp ing with overall similar joint excursions and EMG 106, 107 The neural networks

PAGE 38

38 capable of generating rhythmic motor activity in the absence of sensory feedback are termed Central Pattern Generators (CPG). Another remarkable effect of treadmill walking in the spinalized cats appears to be the specificity of train ing. This implies that in order to learn how to walk, walking needs to be practiced and the neuromotor skill learned during walking does not translate to performance of other tasks e.g. standing. For example Hodgson et al 106 randomized spinalized cats (lower thoracic) into untrained, sta nding trained and treadmill trained group s one month post transection. All the cats were given 2 3 months of training based on the group s they were placed. Following stand training, cats could maintain a standing posture for an extended period of time with out any stimul ation to their tails; however the cats in this group could not step; except for a couple of cats which could take few uncoordinated steps at speeds below 0.2 m/s. On the other hand, the cats in the step trained group could step at faster spee ds (0.62 m/s) but required maximal stimulation at their tails to maintain standing. Follow up experiments included the stand or step training just one week after transaction and continued for six to eight months, then crossed the cats to the other training group. After the cross over the cats learned to perform the original tasks, i.e. if they learned to step before switching to the stand training, they had difficulty stepping, but could stand well. This emphasizes the effect of the specificity of training. Following the strong evidence of the ability of retraining walking post spinal cord transaction in cats, the task specific therapy was employed in individuals post stroke or spinal cord injury. Evidence from human literature Neurorehabilitation is increa singly taking into account novel scientific findings from the basic science. The gait strategies have been recently improved by providing

PAGE 39

39 treadmill training with partial body weight support (BWS) combined with enforced stepping. This method is derived from information gained from animal studies showing that the adult spinal cord can recover a near normal walking pattern after a period of interactive locomotor training (LT) on a treadmill in which weight support for the hind quarters is provided, hence facil itating stepping on a treadmill 50 On the basis of this evidence body weight supported treadmill training (BWSTT) has been employed in individuals post spinal cord injury 108 110 stroke 111, 112 113 Cerebral Palsy 114 115 Several studies have been conducted in individua ls post stroke to test the effectiveness of the BWSTT as a method of rehabilitation of walking 116, 117 Researchers have compared the ability of th e patients following BWSTT with conventional physiotherapy techniques 112, 118 (example: NDT, PNF ) or aggressive bracing techniques 119 Hesse et al 112 conducted a s tudy in which they compared NDT with BWSTT The study was carried out as an ABA single case study with seven patients. The first three week phase (A phase) consisted of 30 min treadmill training each workday. The subsequent three week (B phase) consisted of 45 min physiotherapy sessions daily follow ed by another A phase. The gait parameters were reported to improve only during the A phases. BWSTT in chronic stroke individuals improves balance and gait ability post LT walk at higher speeds associated with improved stride length, stance a nd swing time, initial and terminal double limb support phases 116, 120 122 Additionally, the scores from the Rivermead motor assessment or str oke rehabilitation assessment of motor recovery 111, 123 were also improved post LT. LT also improved cardiovascular status of

PAGE 40

40 these individuals by improving their oxygen uptake 117 However, t here is no consensus regarding the effectiveness of BWSTT as a rehabilitation technique for individuals post stroke. This could primarily be due to the lack of consensus regarding rules to administer and control this therapy. For example, it is important to note that most of the studies mentioned above focused on the performance of the paretic leg alone by employing a trainer to guide the movement of the paretic leg and a trainer on the trunk to facilitate weight shift and maintain upright posture. Moreover, the outcome measures reported were either reported for the paretic leg (stance time, double limb support time) or were focused at both the paretic and the non paretic leg performance, or accomplishment of the task, for example gait speed, functional ambulation scores, or Rivermead mot or scores. The non paretic leg has been largely ignored through these studies and to our knowledge no study has included the non paretic leg in the treatment protocol (except some studies using robotic devices for rehabilitation of walking) however, no stu dy has characterized for the changes in the non paretic leg following LT. Walking is a bilateral task, which requires a coordinated output from both the legs, and the performance of one leg largely influences the output of the other leg. Therefore to impro ve the overall outcome of the rehabilitation of walking post stroke, both paretic and the non paretic legs should be included in the rehabilitation protocols. Furthermore, assessment of the non paretic leg performance post LT would help us to discern if th e locomotor training improves walking function by promoting recovery and thus reduced dependence on the non paretic leg, or compensation by increased dependence on the non paretic leg, or adaptation by development of a new and different motor control

PAGE 41

41 patte rn. This would further improve our understanding regarding the mechanisms underlying BWSTT as rehabilitation techniques. Relevance of the Reviewed Literature to the Studies of the Dissertation Asymmetry is one of the cardinal signs of walking post stroke a nd has been often reported as one of the primary problem in successful accomplishment of certain tasks involving participation of both the limbs, e.g. walking. Asymmetry, i.e. greater contribution of the non paretic leg, has often been reported without bei ng considered to have any pathological importance. Asymmetry has been reported in spatio temporal, kinematic and kinetic measures, but not much information is available regarding the asymmetry of kinetic variables during walking. Since LL and LU form an im portant part of the walking task and are reported to be affected by stroke, through the first study of this dissertation, we aim to investigate the magnitude, pattern and temporal aspects of LL and LU and assess asymmetry (if any) between the paretic and n on paretic leg LL and LU during waking. Since the kinetic parameters are the result of muscle activity, in the next two studies the analysis will concentrate primarily on the EMG and the secondary analysis will include relevant kinetic parameters. Q uantifi cation of the muscle output of the non paretic leg may provide greater insight into the motor control of walking post stroke Furthermore, in the hemiparetic population it is still unclear how the non paretic leg muscle activity may contribute to walking p erformance. There is some evidence of the compensatory mechanisms of the non paretic leg aimed at assisting in achieving a steady walking state. However, there is no single large study to provide detailed information regarding these compensatory mechanisms used by the non paretic leg. Therefore, investigation of the relationship between muscle activity and the biomechanical variables would help determine changes

PAGE 42

42 in the non paretic leg muscles in order to maintain a steady walking state. Thus, the second stu dy of this dissertation aims to quantify the muscle activity of the non paretic leg and determine its relation to biomechanical characteristics of hemiparetic gait. Overall current evidence suggests that BWSTT is an efficient and effective rehabilitation technique to promote walking function in individuals with acute and chronic neurological injuries. However, there is limited or no quantitative information about the changes in the motor control of the non paretic leg and it is unclear how the changes in t he motor control of the non paretic leg relate to the biomechanical parameters of walking performance. Therefore, the third study of this dissertation aimed at quantifying the response of the non paretic leg muscle activity to locomotor training and explai ning the relation to selected biomechanical measures of walking performance post stroke.

PAGE 43

43 Figure 1 1. Represents different phases of the gait cycle. Phase 1 represents the first double limb support, phase 2 represents the first 50% of the single limb support phase, phase 3 is the second 50% of the single limb support phase, phase 4 represents the second do uble limb support phase and phase 5 and 6 represent the first and second 50% of the swing phase (From: Bowden MG, Clark DJ, Kautz SA. Evaluation of Abnormal Synergy Patterns Poststroke: Relationship of the Fugl Meyer Assessment to Hemiparetic Locomotion. Neurorehabil Neural Repair 2010 24(4) 328 337 )

PAGE 44

44 a. b. c Figure 1 2. Represents hip, knee and ankle joint angles subject walking. a) ankle angle, b) knee angle and c) Hip angle in a healthy subject walking at self selected speed. Blue and red represent each leg.

PAGE 45

45 a. b. Figure 1 3. Represents the ground reaction forces. a ) Figure illustrates the 3 D forces applied to the ground from the legs (Fx, Fy, Fz) and the measured Ground reaction forces (Rx, Ry, Rz). x corresponds to me dial lateral, y corresponds to anterior lateral and z corresponds to the vertical component of the ground reaction forces. ( Adapted from http://moon.ouhsc.edu/dthompso/gait/kinetics/ GRFBKGND.HTM assessed on 10.17.2010) b ) Figure represents the vertical, anterior posterior and medial lateral ground reaction forces recorded from a healthy individual walking at self selected walking speed. Red and blue plots represent ground reaction f orces in two legs.

PAGE 46

46 a. b. c. Figure 1 4 Represents power profiles in hip, knee and ankle subject walking at self selected speed a) Hip power b) knee power and c) ankle power in a healthy subject walking at self selected speed. Red and blue plots represent ground reaction forces in two legs.

PAGE 47

47 CHAPTER 2 RATIONALE FOR VARIAB LE SELECTION AND ANA LYSIS Background Hemiparesis post stroke is characterized by asymmetry and incoordination between the paretic and the non paretic leg, resulting in an altered walking pattern. Asymmetry has been reported in terms of spatio temporal 31, 48 kinematic (joint angles) 27, 124 kinetic ( GRF power and work) 12, 59, 60 parameters and muscle activity (EMG) 69, 125, 126 Therefore, these variables form important components for describing the characteristics of movement. GRF can be recorded directly on force plates and the vertical component of the G RF represents the LL and LU) ability. LL and LU has been reported in static and quasi stroke 7 10, 44 However, not much has been investigated with respec t to LL and LU during walking. Walking involves movement of center of mass out of the base of support as opposed to static and quasi static positions where center of mass is always within the base of support. Therefore, it is imperative to assess LL and LU ability in the task of walking. Hemiparesis: Unilateral vs. Bilateral Hemiparesis post stroke, as the name suggests is considered a unilateral motor control problem of the paretic leg, and clinicians and researchers generally consider the non paretic leg as unaffected. Therefore any changes in the non paretic leg output are considered a result of compensation for altered coordination of the paretic leg. For example, there is evidence in the literature that demonstrate significant correlation between the non paretic hip, ankle joint power 59 and work 60 and walking performance

PAGE 48

48 post stroke signifying its role to accomplish the coordinative task of walking 12, 13 On the other hand, some studies have indicated altered non paretic leg muscle activity as co mpared to healthy control subjects, for example, e.g. Shiavi et al 69 and Buurke et al 126 reported changes in the EMG activity of the non paretic leg and suggested they could be due to some residual impairment in the non paretic leg after stroke Evidence of Bilateral Involvement in Upper Extremity Evidence in the upper extremity literature supports bilateral limb involvement post stroke. For example, a study demonstrated that the hemiplegic subjects were impaired in operating a steering wheel wi th their ipsilesional limb 82 Brodal, a neuroscientist observed that his handwriting was less coordinated after a stroke in the internal capsule ipsilateral to his dominant hand 127 Colebatch and Gandevia 79 found that unilateral damage to motor cortex or its descending projections produced weakness in the ipsilesional lim b. These and other results 76, 128 131 suggest that damage to one hemisphere may have a bilateral effect on the control of movement. Causes of Bilateral Limb Inv olvement Following a Unilateral Lesion There is no consensus regarding the bilateral involvement post stroke, but several neurophysiological mechanisms have been reported to identify involvement of both sides after stroke. Deficits in hemiparetic subjects might be due to the interruption of the ipsilateral projection of the corticospinal tract. Although most of the corticospinal fibers decussate in the medulla a significant number remain uncrossed 83 Furthermore, the lesion in one hemisp here might interrupt corticobulbar and corticoreticular projections and consequently affect sub cortical structures involved in motor control 84 The presence of a bilateral component in motor control is supported by observations that a functional activation might hav e a bilateral expression in cortical areas 132 The abnormal

PAGE 49

49 ipsilateral descending output (from the contralesional hemisphere), associated with a unilateral cortical lesion, may contribute to the ipsilateral limbs abnormal function 128, 130 Similar mechanisms might be responsible for the altered performance of the non paretic leg. In addition, there might be greater contribution of the spinal structures in the control of lower extremity functioning as compared to the upper extremity. Goals of Studies of the Dissertation Three studie s were conducted as a part of this dissertation to evaluate and understand the motor control of hemiparetic gait and its response to task specific therapy. Study1. The first study of this dissertation attempts to investigate LL and LU asymmetry during walk ing post stroke and its relation t o functional walking status. We also examine d how LL and LU asymmetry relates to some of the kinematic changes in the leg to maintain steady walking state. Study 2. The analyses of the second study of this dissertation se ek to inspect the changes in the muscle activity in the non paretic leg and whether there is evidence to support that the non paretic leg is affected after stroke. It is important to mention that the performance of the non paretic leg will be evaluated dur ing bilateral walking and not in an experimental condition specifically designed to examine it. Therefore, it would not be possible to explicitly delineate the abilities of the non paretic leg. However, there has been little attention paid to the non paret ic leg performance, so this study fills the gap in our understanding. Study 3. The third study of this dissert ation will answer the question o f whether the motor pattern of the non paretic leg (identified in the second study) can be changed following a task specific LT We will evaluate if the pattern of the non paretic leg

PAGE 50

50 normalizes following the LT and if it appears to do so by reduced compensation or remediated impairment. Variables Used This section discusses the variables selected for each study to accomplish the goals of analyses, benefits and limitations of the variables selected and other potential variables and why they were not selected for t he analyses. Study 1 In this study we examine LL and LU asymmetry during walking and its relation to kinematic measures of the leg (knee angle and leg angle). LL and LU occur during the two double limb support phases of the gait cycle, (where one limb is b eing loaded and the other is being simultaneously unloaded). Therefore all the parameters of this study will be evaluated in these two regions only. The parameters of this study are discussed below. Measure of l imb l oading and u nloading Average normalized vertical GRF This measure will be used to capture the magnitude of LL and LU in the paretic and the non paretic leg during the two double limb support phases while walking on an instrumented treadmill. Benefits opposite reaction of the supporting surface to the body 33 For example, an individual standing on the ground exerts a force on it (i.e. body weight) simultaneously the ground exerts an equal and opposite force on the person, which represents the vertical GRF. An average value of vertical GRF takes into account all the points of the LL and LU phase of the vertical GRF curve. Additionally, normalizing the force value with each

PAGE 51

51 have greater GRF (load). This facilitates int er subject comparisons. Vertical GRF value of 0.5 implies symmetric LL and LU, whereas a value < 0.5 suggests reduced loading (e.g., a lower amount of body weight supported ) and a value > 0.5 means that there is greater LL. Limitations. Average normalized vertical GRF might be affected by walking speed (because vertical GRF is greater during faster walking speeds 133 ). To avoid this we grouped subjects based on their self Household ambulators, 0.4m/s 0.8m/s= Limited community ambulators, >0.8m/s= commun ity ambulators) 15 Other related measures used in literature. LL and LU has been reported in static and quasi static tas ks by using center of pressure displacement 8, 45, 134 peak vertical GRF 135 and the weight borne by each extremity as me asured by force plates 136 and weight scales 46 The center of pressure displacement and velocity provide spatial information of the vertical GRF instead of magnitude of LL and LU. Furthermore, physiologically the peak of LL occurs later than the LL defined for the first study Therefore, the peak value has little significance for the purpose of our study. Timing of l imb l oading and l imb u nloading For the purpose of this study, timi The c ross over point is defined as the specific point in time when the two vertical GRF curves cross each other. It represents the percent gait cycle from the beginning of the double support phase until the majority of weight shifts from the leg in second double support period to the leading leg in first double support phase. Benefits. Timing of LL and LU is an important parameter, because Dequervain et al 29 suggested that the time spent in the pre swing phase is an important predi ctor of

PAGE 52

52 the walking speed of individuals post stroke. Cross over point quantifies the moment of shift of the predominance of the weight, irrespective of the total amount of weight transferred or the peak. It provides accurate and quantifiable information a bout the time (both as percentage of double limb support phase and absolute time in seconds) it takes to shift the predominance of weight. The cross over time in the healthy control subjects (based on the preliminary analysis of study 1) is about 44 4.6% o f the double limb support phase of the gait cycle. This means that at about 44% of the double limb support, the predominance of weight shifts from one leg to the other. Limitations It is an accurate measure, with potentially no major limitations. However, an instrumented split belt treadmill is required to simultaneously record GRF from both legs during steady state walking. Other related measures used in literature Loading time h as been quantified in the literature as time to peak vertical GRF 137, 138 and weight transfer time 135 Time to peak vertical GRF quantifies the time it takes vertical GRF to attain its peak value. However since we desired to capture the moment of the shift of the predominance of weight from one leg to the other (which actually signifies the loading and unloading time), time to peak does not provide that information Furthermore, owing to the altered LL and LU ability, individuals post stroke may have more than one peak GRF or no specific peak ( as seen in individuals with a plateau type LL). A second measure used to report temporal characteristics of LL and LU is weight time transfer 135 It is defined as the duration from the first change in the vertical GRF data (change for either limb of more than 2% of body weight from the mean baseline measurement during quiet standing) to when the force beneath the leading limb reaches

PAGE 53

53 zero, indicatin g that the limb has lost contact with the ground 135 This is a good measure, but it represents duration (period) from the first change in the vertical GRF data, instead of quantifying the moment of weight shift as required in this study. Pattern of l imb l oading and l i mb u nloading The pattern of LL/LU will be calculated from the shape of the vertical GRF curve. The shape of the vertical GRF curve will be quantified using the best fit curve for polynom ial in JMP statistical software (SAS, Cary NC) for the values of the l oading phase (heel strike to just the beginning of single limb support) and unloading phase (end of the single limb support to toe off) for each subject (highlighted with black circle in figure 2 1). A significant first order term will imply a linear slope whereas a significant second order coefficient term will indicate the LL/LU profile has a curvature (positive coefficient is concave and negative coefficient is convex). Benefits It is a simple and elegant method of quantifying the shape of the curves t o understand the pattern of LL and LU. Quantification of the LL and LU pattern will reveal variability in performance of each individual which thus will facilitate classifying individuals post stroke into different categories with respect to the pattern of LL and LU. Limitations. Some subjects might have very irregular vertical GRF curves that are not accurately quantifiable by this technique. However, a closer look at the data revealed that we had only a few subjects with very irregular LL and LU patterns and we did not include them in the analysis. Other potential measures. No one has calculated the shape of the vertical GRF curves, but pattern of loading and unloading has been quantified using center of pressure path 8, 45, 139 However, the center of pressure path provides information about

PAGE 54

54 the path traversed by the point of action of force rather than the actual pattern of LL and LU. Loading rate has also been used to quantify the slope of the curve 140 142 ; however, it gives a single number that quantifies the LL per unit time irrespective of th e shape of the curve. Q uantifying the LL and LU using the polynomial fit provides us more details about the variability and alterations of LL and LU pattern. For examp le, LL and LU in two individuals, one with a concave slope and the other with a convex slope have similar net load ing rate. This could be because they presented with similar change s in the force magnitude per unit time. On the other hand, the polynomial fi t method would group them both into different groups based on the difference in the shape of loading and would inform that the person with concave loading had initial reluctance towards loading, while the person with a convex loading initially, rapidly loa ded their limb, but slowed later on. Kinematic p arameters Medial lateral leg angle. The average medial lateral leg angle is defined as the angle between the vertical and the projection of the line connecting the center of mass of pelvis and the center of m ass of foot in the frontal plane. The leg angle reduces in the first double limb support (LL) phase (e.g., the pelvis is moving toward the foot for the entire phase, with a change of angle of approximately 7 degrees during this phase for control subjects w alking at 0.3m/s) and it increases in the second double limb support (LU) phase (e.g., the pelvis is moving away from the foot for the entire phase, with a change of angle of approximately 7.5 degrees during this phase for slow walking control subjects). T hus, this average will characterize how close the pelvis got to the loading leg for the full the loading period, as opposed to only characterizing the value achieved at

PAGE 55

55 the end of the loading phase. Since we aim to individually investigate the correlation between LL and LU and medial lateral leg angle, we suggest average leg angle value could capture these changes in respective phases of the gait cycle and help us investigate the relation with the LL and LU. Average knee flexion The average knee flexion ex tension angle was calculated during each of the two double limb support phases of the gait cycle. Benefits. We selected to investigate the relation between the medial lateral leg angle and LL and LU post stroke, because evidence in the literature demonstr ates that lateral foot placement asymmetry (the wider paretic foot placement relative to the pelvis than the non paretic foot), strongly correlates with the weight support 143 We aimed to verify th is in the context of our study. Furthermore, knee joint angle goes through a sequence of flexion and extension during the gait cycle. We believe that average is relatively repeatable and can be consistently interpreted, while other variables, like the peak or minimum knee angle or joint angle at a particular percentage of gait cycle have higher variability and are not as repeatable. Limitations Although knee angle and leg angle are significantly related to LL and LU (as discussed in study 1), they might be affected by other factors, like muscle strength etc. However, since we will be using a correlation analysis, a significant correlation would suggest role of knee and leg angle in LL and LU, in addition to the others factors which might be affecting them. Study 2 Muscle activity represents the sole mechanism that allows control over movement of the body segments. The force that is exerted by the muscle needs to be scaled appropriately for each instant of movement sequence, resulting in continuous changes

PAGE 56

56 in the magnitude of output as a function of time. Furthermore, muscles act at joints at every step of the movement, which can be characterized as moments. These joint moments are then used to quantify the work done at the joints. The EMG activity and the wor k done at the joint (ankle, knee and hip) are the primary variables of the analysis. In contrast to the paretic leg, the non paretic leg may not have obvious alterations in its performance. However, the small differences might also influence the performanc e of the non paretic leg and the success in accomplishing the task. Therefore, it is important to carefully select measures to capture the nuances of the non paretic leg performance. EMG t iming v ariable Well timed muscle activity is important to achieve we ll patterned walking. For the purpose of this dissertation EMG timing was recorded over the gait cycle and different regions of the gait cycle (regions of the gait cycle as described in chapter 1 figure 1 1). K means analysis. K means cluster analysis is a statistical technique that groups similar data points in a single cluster. In the context of EMG analysis it differentiates filtered EMG signal 125 The individual da ta points are assigned to k cluster by MATLAB (Mathworks, Inc. Natick, MA ) such that the distance from the centroid is minimized. Therefore, the most similar data points are grouped together. The c luster with the lowest mean value corresponds to inactivity and the other clusters correspond to periods of muscle activity 125 (figure 2 3 ) Benefits of using K means cluster analysis. The k means algorithm is based on minimizing the sum of squared distances from all the points in every cluster to their cor responding cluster center 125 It is a simple algorithm that can be applied to large data

PAGE 57

57 EMG. Since k means individually identifies each data point, it takes care of e ven the short duration of muscle activity, by assigning them to the closest cluster 125 Short bursts of muscle activation are a common occurrence in non paretic leg. Limitations of k means. The adequateness of k means depends to some extent on the cut off frequency of the low pass filter and on the number of clusters set a priori 125 For this dissertation we set the number of clusters to five, because Den Otter et al 125 suggested that five clusters could capture the muscle activity in individuals walk ing after stroke. I n case of a homogeneously distributed data set (a muscle displays no activity or continuous activity) this algorithm fails to provide any useful information 125 However, close inspection of the raw data reveals that there were no such EMG signals in our data set. Finally, there is some possibility of error due to initializing of the cluster centers, but by selecting cluster centers far from each other this type of error can be minimized. F urthermore, we tested the reproducibility of the k means function by performing it on a set of randomly selected walking trials. Visual inspection of the results revealed that they were consistent over several attempts. Percentage duration of activation ov er the gait cycle. Percentage duration of activity of a muscle in a gait cycle is sum of all the "on" points in gait cycle divided by the total number of points in the same gait cycle. The muscle "on" and "off" times are determined using k means clustering for each point in the gait cycle. Percentage duration of muscle activity in a region of the gait cycle. Percentage duration of activity of a muscle in a region is calculated by dividing the

PAGE 58

58 activity of muscle in a region of the gait cycle by the total nu mber of points in that region and multiplying with 100. Demonstration To calculate the percentage activation of TA (Tibialis Anterior) over the gait cycle, assume the gait cycle has X total points, TA had X1 on points in region 1 (which had total points Y1), X2 in region 2 (which had total points Y2), X3 in region 3 (which had total points Y3) and on. Therefore, the percentage duration of activity in the gait cycle would be ((X1+X2+X3+X4+X5+X6)/ X)*100 And the percentage activity o f muscle in the first region of the gait cycle would be (X1/Y1)*100 Benefits of using percentage duration of activity of a muscle over the gait cycle. Quantifying muscle activation over the gait cycle provides global information related to the temporal pa tterning of the non paretic leg. Since k means clustering includes even the short bursts of activation, it captures all the bursts of activation. also takes into cons ideration the variability in duration of the gait cycle. Limitations of percentage duration of activity of a muscle over the gait cycle. In addition to the limitations of the k means analysis discussed above, the percentage activity of a muscle over the ga it cycle is a global measure, which could obscure the important differences in different phases of the gait cycle. In order to obtain detailed characteristics of temporal patterning of muscle activity in the non paretic leg, we also investigated the percen tage duration activation of muscle in different regions of the gait cycle.

PAGE 59

5 9 Benefits of percentage duration of activity in different regions of the gait cycle. It provides information regarding duration of muscle activity in different regions of the gait cy cle. Detailed information of temporal characteristics of muscle activity indicates the region of the gait cycle when the non paretic leg muscle activity is different as compared to healthy control subjects. This could further help to categorize muscle acti vity being due to compensation or impairment, as discussed later. Limitations. The percentage duration of activation of muscle in a region provides detailed information related to the temporal characteristics of muscle activity. It does not have any limit ation except the ones discussed for k means cluster analysis, because k means cluster analysis form s the basis of all the temporal calculations of EMG in this dissertation. Other potential measures. Timing parameters are often derived from the EMG profiles by applying a threshold to detect onset and offset of burst activity. However, there is no consensus in the literature about threshold selection. Moreover, several researchers have suggested that these methods introduce large systematic errors 144 1 46 Another method for tempo ral characteristics used by Mulroy et al 147 was using an tivity that had an intensity of at least 5% of the maximal contraction for minimum 5% of the gait cycle. Kawashima et al 148 on the other hand included muscle activity in their analysis if the muscle was consistently three standard deviations above the recorded resting activity. However, these quantification techniques may not be adequate for pathological gait because it might not be possible for an individual to maintain a true resting state, or a real maximal contraction (due to the pathology). Furthermore,

PAGE 60

60 sometimes the EMG may be well timed to maintain a steady walking state, but it may be less than 5% of the maximum or thr ee standard deviations from the resting. In the cases l ike this the data might be lost because of the threshold of activity required to quantify for inclusion. However, k means cluster analysis has the advantage of being able to include even the short burs ts of activity. EMG m agnitude For the purpose of the studies of this dissertation, the magnitude of EMG activity will be represented by i) the integrated EMG activity over the gait cycle ii) integrated EMG activity of a muscle in different regions of the gait. Integrated EMG activity. The filtered and rectified EMG signal is numerically integrated using a trapezoidal rule, an extended closed formula integration rule with equally spaced abscissas, with respect to time in each region and each gait cycle. In tegration calculates total area under a curve, and numerical integration divides a curve into a bunch of component areas and sums these component areas together to get the total integral (figure 2 2) Benefits. Integrated EMG is a robust measure of muscle activity and has been most often reported in the literature 149 It has proven to be a useful method for evaluation of muscle activity and offers an instrument for documenting muscle rehabilitation. It is robust against movement artifact 150 We hypothesized that there would be significant differences in the magnitude of the muscl e activity in the non paretic leg and healthy control subjects. Therefore, the next step will be to investigate closely different regions of the gait cycle. In order to quantify the magnitude of muscle activity in different regions of the gait cycle we cal culated the integrated EMG activity in each region of the gait cycle.

PAGE 61

61 Limitations. The amplitude of the surface EMG is related to the net motor unit activity, i.e. recruitment and discharge rates of motor units 151 However the amplitude of EMG detected using surface EMG is sensitive to many intrinsic (thickness of subcutaneous fat, distribution of motor unit s, conduction velocities and detection systems used), and extrinsic factors, i.e. the factors which can be influenced by the experimenter (orientation, location and area and shape of the electrode and distance between them 152 ), which renders it highly variable. Furthermo re, the reliability of the magnitude of the EMG in the dynamic contraction is also influenced by length and force change due to stationary shift in relative position of the electrode and change s in tissue conductivity 153 Furthermore, change in the joint angles cause the recording electrodes to shift relative to active muscle fiber and the change in the muscle fiber direction to alter tissue conductivity. Additionally, the length of muscle fiber and the shortening velocity is also inversely r elated. 153 The factors discussed above might increase the variability and reduce t he reproducibility of the EMG magnitude; however, we took careful measures to minimize the error due to extrinsic factors. Other potential measures. To investigate the amplitude of muscle activity over the gait cycle, peak Root Mean Square and mean Root Mean Square have been often reported in the literature. Peak root mean square uses a single value 150 the maximum of the root mean square smoothed signal to represent muscle activation 150 However, it is not robust against movement artifact and error inhe rent within the EMG signal 150 On the other hand, the mean root mean square is a robust measure that limits the effects of movement artifact, however is less sensitive to changes in the EMG signal and may mask the differences in muscle activation and intensity between the experimental

PAGE 62

62 c onditions 150 T herefore, integrated EMG activity is our best choice for the global EMG activation. Similarly, to obtain detailed information regarding the magnitude of activity in different regions of the gait cycle, integrated EMG was calculated. Each of these measures if used adequately would provide important information about the muscle activation and the information obtained from each of them is similar. However, as discussed above, there are some inherent problems associated with each of them. Work Done Work done at a joint is calculated by numerically integrating the joint power curve using a trapezoidal rule, an extended closed formula integration rule with equally spaced abscissas with respect to time in each region and each gait cycle. Integration is finding th e total area under a curve and numerical integration is dividing a curve with several component areas and adding these component areas together to get the total integral. A2, H1, H3 represent the most important and significant phases of positive work durin g the gait cycle and have been shown to relate to the walking performance 12, 13, 59, 60 In addition, other source of power generation is, K2, i.e. positive power by concentric activity of the knee extensors while, A1, K1, K3 and K4 represent phases of negative work. H4 (hip abductor power in frontal plane) is an important source o f positive power generation reported in individuals walking post stroke 12 however it is not found in healthy subjects. Although Kim et al 12 demonstrated that the magnitude of the joint powers in the non saggital planes (esp. at the hip) are significantl y related to the individuals walking performance (as measured by walking speed), stronger correlations were found with the hip and ankle joint in the saggital plane. Therefore, for the purpose of this dissertation we selected to study positive work done at the hip and ankle joint.

PAGE 63

63 Benefits. Work done is an important measure of motor performance 149 It closely represents musculo tendon work 154 and has been used to examine locomo tor deficiencies 155, 156 determine how the muscle functions 157 and identify impairments associated with neurological deficits 59, 60 Limitations. The primary limitation is the inability of work to account for individual muscle contributions, primarily due to co contraction causing the net moment to be less than the absolute sum of the individual muscle flexor and extensor moments and muscle tendon energy storage and release that allows negative work in one phase to be recovered as positive work in a subsequent phase 154 We aim to investigate the correlation between the work done at the joint and the muscles which might have an important contribution to work production. Other potential parameters. Joint power could have been another important variable to investigate the contribution of different muscles. However, power is generally presented as an instantaneous property (as peak or at a particular time in a phase of gait cycle, or gait cycle) while work done is the time integral of the joint power and since we aimed to investigate the magnitude and timing of EMG activity of the muscle over a period of time we chose a measure of motor performance that represents the whole r egion of the gait cycle, like the work done at a joint. Definitions of Important Terms Compensation Compensation is a strategy employed to perform an old task in a new manner following an injury 158 For the purpose of this study, in the context of the non paretic leg, compensation implies the change in the motor control of the non paretic leg, presumably either an increase in output during a normally active period or activity in a

PAGE 64

64 normally inactive period, to perform a function usually performed by a healt hy contralateral leg that is not being performed by the paretic leg due to weakness and/or discoordination For example, the increase in the activity of the ankle plantarflexor (Medial Gastrocnemius (MG), Soleus (SO)) during the pre swing phase of walking may generate greater propulsion to compensate for the impaired ability of the paretic leg. Impairment Impairment, on the other hand, is the deviation from the generally accepted standards in the biomedical status of the body and its functions 159 For the purpose of this dissertation, impairment in the non paretic leg is suggested when there is an excess or deficit in the output of any parameter as compared to healthy control subjects, which would be expected to deteriorate the performance of the non paretic leg itself. We recognize that this is a difficult question and that it is unlikely that we will be able to answer it with certainty. However, we will use a consistent and rigorous a nalysis approach in order to determine how strongly the data support the identification of impairment. Analysis Plan All the variables of EMG magnitude and timing will be calculated for healthy control subjects walking at three pre selected speeds (0.3, 0. 6, 0.9 m/s, to approximately match the self selected speed of individuals post stroke). Once the detailed information regarding all the variables of muscle activity from the healthy control subjects is obtained, it will be then used to evaluate the perform ance of the non paretic leg. The post stroke individuals will be divided into three groups (slow < 0. 4 m/s, moderate 0.4 0 .8m/s, fast > 0.8 m/s) based on their self selected walking speed on the

PAGE 65

65 treadmill. Classification will be based on the treadmill speed because we aim to compare the biomechanical characteristics of these subjects that were recorded while they were walking on the treadmill. Therefore, classifying them based on their self selected over ground speed and comparing their performance on the tr eadmill would result in some ambiguity regarding the results and the conclusions drawn may not be generalizable. The next step of analysis will be to perform independent samples t test to compare 1) integrated EMG activity over the gait cycle for all the muscles (8 variables), 2) percentage duration of muscle activity over the gait cycle (8 variables), between the non paretic and healthy control subjects walking at matched speeds. Thus, timing and magnitude of 8 muscles (16 total variables) in the non pare tic leg will be compared to timing and magnitude of healthy control subjects for each speed group. Furthermore, we will compare the integrated EMG activity between the non paretic leg and healthy control subjects (for specific muscle duration, magnitude or both) in the various regions of the gait cycle. If a significant difference is revealed in the region of predominant activity of a muscle in healthy control subjects, that would suggest compensation. This is because the non paretic leg has to generate gre ater magnitude of activity or fire for a longer duration in order to maintain a steady walking pattern. However, if the prolonged duration of activation or increased magnitude persists beyond the region of predominant activity (as revealed by control subje cts) it might indicate impair ment owing to the inability of the non paretic leg to adequately modulate the muscle activity during walking post stroke.

PAGE 66

66 Hypotheses The hypotheses are based on the preliminary analysis of the control data and the evidence fr om the literature. TA is reported to have two major bursts of activity during the gait cycl e: in region one when it prevents the foot slap and in region five when it contracts to help foot clear the ground. The results of descriptive analysis for healthy control subjects revealed that irrespective of the walking speed, TA had maximum activity in the fifth region of the gait cycle. Therefore, significantly greater EMG activity in the fifth region of the non paretic leg gait cycle would indicate compensation Nevertheless, if the maximum activity of TA is in another region of the gait cycle, we will investigate its role in impairment/ compensation. Furthermore, significantly increased activity of TA in the fourth region of the gait cycle will interfere with the performance of the plantarflexors 160 To test this we wi ll evaluate the correlation between the positive work done at the ankle in the fourth region of the gait cycle and TA activity in the same region. A significant negative correlation would indicate that an increase in the TA activity in the fourth region of the gait cycle will reduce the positive work done at the ankle and thus, indicating impairment. During the second and third region of the gait cycle, TA contracts in conjunction with plantar flexors to maintain the limb in extension to facilitate weight s upport 125, 161 To evaluate the role of TA in postural balance during second and third region of the gait cycle, we will calculate correlation between TA activity and A1 (work done by eccentric contraction of plantar flexor). A sign ificant negative correlation would suggest that an increase in TA contraction reduces the A1, i.e. work done by ankle plantarflexors which might influence the stability of the leg during these regions.

PAGE 67

67 Medial Gastrocnemius ( MG ) and Soleus ( SO ) are the two primary plantarflexor muscles and are responsible for generating maximum positive propulsive work during the pre swing phase to accelerate the leg and trunk. Healthy subjects revealed maximum activity of the MG and SO during the third region of the gait cy cle (late single limb support) at all speeds. Furthermore, literature indicates that the activity of MG and SO is of paramount importance in the fourth region, i.e. pre swing phase, of the gait cycle, for generating the positive ankle plantarflexor power 19, 32, 162 We will investigate the region of predominant activity in the non paretic leg. Signif icantly greater MG and SO activity (time and magnitude) in third or fourth region of non paretic leg will suggest compensation. The increased SO and MG activity will compensate by generating greater positive work at the ankle to propel the leg and increase its kinetic energy. The increased kinetic energy will help the leg to swing through rapidly to reduce the non paretic swing phase, and thus single limb support time of the paretic leg 28 To further confirm the compensation, we hypothesize that the work done at the ankle in the fourt h region of the gait cycle will have a significant positive correlation with the MG and SO activity in the same region of the gait cycle. This will suggest that greater activity of the MG and SO is associated with greater work production at the ankle in no n paretic leg, which has been reported to be important contributor to the walking speed of the subject, thus indicating compensation. Nevertheless, if the maximum activity is found in some other region of the gait cycle, we will investigate its role in imp airment or compensation. Results of simulation analysis demonstrate that MG and SO contract during the single limb support, i.e. second and third region of the gait cycle to extend the leg and support the body weight to maintain balance 161, 162 Increased activity on the non paretic

PAGE 68

68 leg plantarflexors in these regions of the gait cycle would suggest that the muscles are working hard during the extended single limb support phase to support the body weight and maintain balance A significant positive correlation between MG and SO activity and A1 in the second and third region of the gait cycle would indicate compensation by the ankle plantarflexors to maintain stability to support the weight. However, reduced plantarflexor acti vity during this region of the gait cycle (as revealed by the preliminary analysis) would suggest inability of the plantarflexors to maintain the leg in extension to support the weight of the body. This weakness of the plantarflexors is compensated for by increased Vastus Medialis (VM) activity 161, 163 Excessive activity of the non paretic leg plantar flexors in the first region of the gait cycle, in individuals with short non paretic step length, would help i n compensation by generating propulsive forces for the poorly coordinated paretic leg. Step length asymmetry has been quantified as step length ratio and is defined as the ratio of paretic step length to the total stride length 164 Therefore we propose some individuals with high step length ratio (i.e. > 0.52; 0.5 being symmetric) will have a step to gait. To investigate this we will compare the non paretic leg angle in individuals that generate A2 against those who do not generate A2. Significantly increased leg angle in individuals generating A2 would suggest compensation by the non paretic plantar flexors. In region five and six of the gait cycle MG acts in conjunction with primary knee flexors to assist in knee flexion 165 Increased MG activity in these regions would suggest compensation to facilitate ground clearance (counteracting the effects of SO and Rectus Femoris (RF) over activity) or to increase the energy of the non paretic leg to short en the swing phase.

PAGE 69

69 Furthermore, increase in the SO activity in region five and six of the gait cycle would likely suggest impairment, because there is no biomechanical advantage of the SO activity in the fifth and sixth region of the gait cycle, because t hey might just interfere with ground clearance. VM is primarily a knee extensor, which helps to extend the knee and support body weight during stance phase of gait. Descriptive statistics of healthy control subjects revealed maximum activity of VM during t he first region of the gait cycle or during the weight acceptance phase. Therefore, if the results of analysis show increase in activity of VM during the first region it would suggest compensation by the non paretic leg to prevent knee collapse and facilit ate weight acceptance. Simulation analysis demonstrated that VM has an important role in the stance phase to support the body upright 163 VM works in association with the plantarflexors (mainly SO) to maintain the leg in extension. The se studies further stated that the plantarflexors are the first line of action H owever, in case of impaired plantarflexor function, VM contracts to assist in maintaining extension of the leg 162, 163 On similar lines if during the region two and three of the gait cycle the muscle activity is significantly reduced in the plantarflexors (as suggested by the preliminary analysis) and there is a simultaneous increase in the VM muscle activity in the second and third region of the gait cycle this would suggest compensation by the VM for the weakness of the plantarflexors. Furthermore, results of a recent simulation analysis study indicated that the increased activity of VM in mid stanc e facilitated propulsion of the paretic leg.

PAGE 70

70 Therefore, increased VM activity compensates for the reduced propulsion of the paretic leg. In addition, the activity of the VM in the sixth region of the gait cycle contributes to deceleration of the rapidly flexing knee, in preparation for landing. I ncreased activity of VM in this region would suggest greater force r equirement to control the rapidly flexing knee, which could be a result of i ncreased kinetic energy and increased knee flexion of the non paretic leg 28 However, during the swing phase, the knee needs to be flexed to facilitate ground clearance, therefore, excessive activity of the VM during the fifth region would indicate impairment, and reduced ability of the non paretic leg to modulate t he muscle activity. RF is another muscle that belongs to the quadriceps group. There is a controversy in the literature regarding the EMG activity of the RF. Annaswamy et al 166 and others 167 have reported biphasic activity, while Nene et al 168 and Perry et al 19 have reported a single burst of activity during the gait cycle. The preliminary analyses of our control data reveal two prominent phases of activity of the RF muscle: one during the first region of the gait cycle and the second during the region four (pre swing phase) of the gait cycle where the RF acts as a hip flexor in propelling the limb forward into swing. D escriptive greater than the pre swing phase, as reported in literature 19, 168 However, simulation studies do not support this function of the RF, and suggest that the RF generates an extension moment at the hip and the knee d uring the pre swing phase. Furthermore, Hernndez et al 169 in a study of electrical stimulation of the RF during walking reveal that the RF activity before toe off, i.e. region four of the gait cycle diminishes hip and

PAGE 71

71 knee flexion. In addition, several studies have reported that increased activity of the RF dur ing the early swing phase, i.e. region five of the gait cycle, is associated with the stiff knee gait 170, 171 The stiff knee might thus interfere s with ground clearance and result s in alte red walking pattern. Stiff knee gait due to the over activity of the RF muscle in the pre swing and the swing phase is characterized by reduced peak knee flexion 171, 172 Furthermore, the res ults of simulation analysis showed that elimin ation of RF activity during the pre swing and swing phase increased the peak knee flexion angle during swing 172 To confirm the influence of the RF activity on the peak knee flexion angle achieved during the swing phase of the gait cycle we will correlate the peak knee flexion angle during th e fifth region of the gait cycle with the RF activity during the same region. A significantly negative correlation would indicate that the increased RF activity is interfering with the knee flexion. Therefore, we hypothesize that the increased burst of RF activity of the non paretic would indicate impairment. In first through third and sixth regions RF is involved in decelerating the leg and shock absorption and weight support during the single limb support, when the whole limb is maintained in extension. Lateral Hamstrings (LH) and Medial Hamstrings (MH) are some of the major contributors to energy generation over the gait cycle 173 They are active in late stance and beginning of the stance restoring energy to the body near double limb support, a most effective time to utilize the passive locomotion properties of the body 173 Maximum activity of hamstrings muscles was present in region one (0.3m/s) and region six (0.6m/s and 0.9 m/s) of the gait cycle in the healthy subjects. If the maximum activity is present in the first/ sixth region of the gait cycle of the non paretic leg, then we will

PAGE 72

72 compare the integrated EMG activity in these regions. Significantly greate r non paretic leg activity would suggest compensation by the non paretic leg. Nevertheless, if the maximum activity is found in some other region of the gait cycle, we will investigate its role in impairment or compensation. We hypothesize that increased activity of the hamstrings (LH and MH) in the first and sixth region of the gait cycle (depending on the walking speed) would indicate compensation by the non paretic leg to produce greater work to support the non paretic leg to facilitate swing ing of the paretic leg and maintain steady walking state Furthermore, a significant correlation between the hamstrings activity and the H1 positive work in region would further strengthen it Buurke et al 126 reported an extra burst of MH in the stance phase of the non paretic leg and they suggested it assists in extension of the non paretic hip, thereby facilitating the swing of the paretic leg and thus compensating for the diminished swing caused by the lack of push off and reduced force of hip flexors on the paretic side. If an extra burst of the MH is revealed in our data of the non paretic leg, we would evaluate the suggestion made by Buurke et al 126 by evaluating the correlation between the MH activity in region three and four of the non paretic leg and the positive hip extension work Gluteus Medius (GM) contraction during the single limb support phase provides stability to the pelvis in the frontal plane. GM provides body support and forward propulsion in early stance 161, 174 Excessive activity of the GM during the stance phase of the non paretic leg would indicate compensation for increased loading of the same.

PAGE 73

73 Approach to Interpretations First we will investiga te differences in muscle activity over the gait cycle to broadly identify differences between the motor control of the non paretic leg and healthy control subjects. The n ext step will be to identify all the regions of the gait cycle that present altered mu scle activity in the non paretic leg. We will inspect the kinematics (e.g. leg angle) for any difference in the leg configuration that might be related to the difference in the muscle activity. Following this we will evaluate related measures of motor perf ormance, i.e. the work done at the joint, to estimate the biomechanical effect of the same. The information gained from the steps discussed above can thus be used to discuss altered muscle activity in the non paretic leg, relative to the biomechanical char acteristics and our original hypotheses related to muscle activity. The third study of the dissertation, will investigate the response of the non paretic leg to LT In this study, we will evaluate how the impairments and compensation mechanisms change afte r task specific LT. To accomplish this we will first group subjects into different speed based categories (as discussed above). We will compare the performance of their non paretic leg to the healthy control subjects walking at matched speeds before LT si milar to chapter 4 Next we will group the subjects based on their self selected walking speed on the treadmill after the LT. The performance of the non paretic leg will be compared to healthy control subjects walking at the matched speeds. It would be int eresting to evaluate the changes in the non paretic leg, i.e. does it continue to be impaired and compensating, or is there some optimization in the regions of the gait cycle which revealed altered behavior before the training. There might be reduction in impairment, or increased compensation or optimization of the muscle output.

PAGE 74

74 Limitations of the analysis S everal other variables were considered before selecting the variables for the analysis of these studies; however, there are some limitations. EMG vari ables form the core of the analyses for chapter 4 and 5 The information obtained from the EMG recordings may not reflect the true state of the muscles as they may be confounded by external (orientation and location of the electrode, machinery, power lines ) and internal factors (muscle fiber type, depth and the amount of tissue between the muscle and electrode) 151 which cannot be ruled out. Therefore, we performed correlation analysis with the measures of motor performance, i.e. work done to confirm the results of the EMG analysis. Work done is an important quantity in gait analysis and has been often used to analyze energy generation, absorption and/ or transfer within the body segments 175, 176 examine locomotor efficiencies 155, 156 determine how muscle s function 157 and identify impairments associated with neurological deficits 59, 60 A correlation analyses like this provides information regarding the role of the non paretic leg muscle in the function of walking. Furthermore, in a bilateral task like walking, it may be difficult to differentiate between the role of the non paretic leg in compensation and impairment post stroke. However, in this dissertation, we tried to explicitly define compensation and impairment for our studies and then set strict criteria t o demark the non the two. Furthermore, we also performed a detailed analysis of the muscle activity and motor performance measures to ensure correct inference of the non paretic performance.

PAGE 75

75 Figure 2 1. Represents the vertical GRF in a healthy control subject. The blue and red plots represent right and left leg, and the part of the curve in black circles represent the loading and unloading phases.

PAGE 76

76 Figure 2 2 Represents the calculati on of integrated EMG activity First figure represents the raw EMG signal, second panel represents the rectified and filtered EMG signal and the third figure represents integration rule with equally spaced abscissas (red lines) with resp ect to time in the gait cycle

PAGE 77

77 Figure 2 3 Represents the steps of k means analysis. First figure represents the raw EMG magnitude, second figure shows the filtered rectified EMG (linear envelope) and the third figure represents the original raw EMG signal with the detected periods of activity in dicated in gray shaded areas.

PAGE 78

78 C H APTER 3 MAGNITUDE AND RATE O F LIMB LOADING AND U NLOADING DURING HEMIPARETIC GAIT IS ASYMMETRIC AND RELAT ED TO LEG KINEMATICS Background Hemiparesis following stroke results in poorly coordinated and asymmetric gait. In addition to well described kinematic asymmetries, there are also less well described kinetic asymmetries. Kinetic parameters ( e.g. joint moments, powers and GRF ) are particularly important due to their dependence on both kinematics and muscle activity a nd can provide insight into the underlying causes of changes in walking pattern. In particular, vertical GRF has been used to characterize post stroke weight bearing 7, 46 The paretic leg typically exhibits lesser weight bearing than the non paretic leg (i.e., neither LL nor LU are symmetric) 36 41 LL and LU asymmetry are important contributing factors to postural instability, reduced dynamic balance, increased postural sway and thus falls 36 Reduced paretic (as compared to non paretic) LL has been reported in static and quasi static tasks (e.g., standing in different stance positions) 7, 9, 10 stepping on stairs of different heights, and sit to stand from a chair 9 ) and has been related to walking ability post stroke 35, 47, 48 However, studies typically report reduced paretic LL without quantifying the temporal characteristic of LL and LU, which may provide important information to understand leg loading. While it i s often assumed clinically that reduced loading of the paretic leg is indicative of poor walking and recovery, LL and LU asymmetry during walking in individuals with hemiparesis has not been quantified and the relationship with walking performance or strok e severity is unknown. Measuring LL and LU ability during static standing (standing on force plates and transferring weight from one leg to the other while the whole body center of mass stays between the two feet) does not provide much detail about walking ability, because

PAGE 79

79 walking involves acceleration of all individual body segments 7 such that the whole forward and vertical acceleration of the whole body center of mass, walking also involves medial lateral accelerations of the center of mass. Tho ugh the magnitude of movement s and forces in medial lateral direction are small, they are expected to be important to facilitate loading and unloading. This would be consistent with the observation that manual assistance provided at the pelvis by trainers during locomotor training to produce medial lateral movement of body, presumably to facilitate appropriate weight shift, appears to improve walking pattern and increase stability 108 Multiple paretic and non paretic LL and LU patterns can be employed by individuals with hemiparesis to maintain a steady walking state. It is important to quantify these patterns to understand the asymmetry (as well as the differences in asymmet ry between subjects) in steady state walking. The presence of multiple loading patterns would be consistent with the observation that subjects with post stroke hemiparesis exhibit multiple kinematic patterns to successfully maintain a steady walking state. The kinematic characteristics may provide more insight into the various patterns of LL and LU exhibited by different individuals. Therefore, we hypothesize that classifying post stroke individuals based on their LL and LU and evaluating their relationship to kinematic and kinetic measures will provide deeper understanding of weight support and dynamic stability during walking. The main purpose of this study was to determine the relationship between LL and LU asymmetry in individuals with hemiparesis and 1) functional walking status and 2) specific kinematic and kinetic variables during walking. We hypothesize that LL and LU

PAGE 80

80 in the paretic and non paretic leg will differ with functional walking status and that the asymmetry will be related to knee angle and medial lateral leg angle (angle between vertical and the frontal plane projection of the line connecting the center of masses of the pelvis and foot). Since the magnitude of medial lateral forces are directly related to the medial lateral accelerations of the whole body center of mass, we hypothesize it will be intimately related to weight shift, thus we will also establish the relationship between medial lateral forces and LL and LU. Detailed understanding of LL and LU asymmetry will provide deeper insight about the changes in locomotion in post stroke individuals and potentially inspire new strategies to rehabilitate walking. Methods Participants Data were collected from forty five individuals with chronic hemiparesis (age = 60.2 12.3 years, 19 females, 19 left side hemiparesis) and eighteen similarly aged healthy individuals (age = 66.2 10.0 years, 4 men) at the VA UF Human Motor Performance Laboratory, in Gainesville, Florida. To be included in the study, individuals had to have: hemiparesis secondary to a single onset unilateral stroke; ability to ambulate independently with or without an assistive device over 10 m on a level surface; ability to walk regularly at home. Exclusion criteria for the participants from the study were: any orthopedic or neur ologic conditions in addition to stroke that might limit hip and knee extension or ankle plantar flexion to neutral, or inability to provide informed consent. All participants in the study signed a written informed consent and the study was approved by Ins titutional Review Board at University of Florida. Three categories of subjects with different severity were identified based on functional walking status as defined by their self selected overground walking speeds 15 individuals who walked <0.4

PAGE 81

81 m/s (household ambulators), or 0.4 0.8 m/s (limited community ambulators), or > 0.8 m/s (community ambulators). Experimental Prot ocol Retro reflective markers were placed at specified bony prominences on the body. Clusters of reflective markers attached to rigid bodies, were also placed firmly on the pelvis, each thigh, shank and foot of the participants. All participants walked on an instrumented split belt treadmill (Techmachine, Andrezieux Boutheon, France) without using any assistive device or ankle foot orthosis for three walking trials at their self selected treadmill walking speed. Data were collected for 30s after the subject walked for ~ 10s to achieve steady state. 3 D GRF and kinematic parameters were obtained for each leg as the subjects walked on the treadmill. A harness system mounted to the ceiling of the laboratory was used for safety in case participants lost their ba lance. The harness was worn across the shoulders and chest of the participant and no body weight was off loaded by the system. A physical therapist stood near the subject, but did not provide assistance as they walked over the treadmill. Healthy control su bjects walked at their self selected speeds as well as at speeds of 0.3m/s, 0.6m/s, 0.9m/s to provide control data at speeds comparable to that of the hemiparetic individuals. A twelve camera motion analysis system (Vicon Motion Systems, Los Angeles, CA) w as used to collect the data as the subjects walked on the instrumented treadmill. Customized MATLAB (MathWorks Inc, Natick, MA) programs were used for data analysis. Kinematic data were low pass filtered using a fourth order zero lag Butterworth filter wit h a 10 Hz cut off frequency. GRFs were sampled at 2000 Hz and low pass filtered using a fourth order zero lag Butterworth filter with a 20 Hz cut off

PAGE 82

82 expressed as percent o f paretic leg gait cycle i.e. paretic foot contact to paretic foot contact. LL and LU occur during the double limb support phases of the gait cycle; therefore, all the variables in this study were calculated in these two phases. Study Variables Magnitude o f loading/unloading Average LL and LU were quantified by the average vertical GRF in the first and second double support phases, respectively ( f igure 3 1 a 3 2 a ). The legs in first and second double support phase were defined as loading and unloading, respectively (i.e., when one leg is being loaded the other is being unloaded in the same double limb support phase). Timing of loading/unloading The timing of LL was calculated for each leg as the cross over point of the two vertical GRF curves during the double limb support (i.e. the percent gait cycle from the beginning of the double support until the majority of weight shifts from the leg in second double support to the leading leg in first double support ( f igure 3 4 a ) ) Thus, the timing of LL of one leg is equal to that of LU in the other leg and expressed as a percentage of the double support phase (i.e., 50% = symmetric). Pattern of limb loading/unloading The pattern of LL/LU was calculated from the shape of the vertical GRF. We identified three patterns ( f igure 3 1 d ): concave curvature (i.e., initially slow then rapid loading), convex curvature (i.e., initially rapid then slow loading) and linear. The best fit curve for a polynomial was calculated (JMP statistical software, SAS, Cary NC) for the vertical GRF values for the loading phase and unloading phase (end of the single limb

PAGE 83

83 support to toe off) for each subject. Significance was set to p<0.001 as a correction for the large number of calculated curves. A significant first order term meant a li near slope, whereas a second order coefficient term meant the LL/LU profile had a curvature (positive coefficient was concave and negative coefficient was convex). Kinematics The average leg angle in the medial lateral direction was calculated during each double limb support phase as the angle between the vertical and the projection of the line connecting the center of mass of pelvis and the center of mass of foot in the frontal plane ( f igure 3 1 c f igure 3 2 c ). The average knee flexion extension angle was calculated during each of the two double limb support phases of the gait cycle ( figure 3 1 b f igure 3 2b ). Statistical Analysis One way ANOVAs were performed to investigate the differences in the LL and LU between the paretic leg, non paretic leg and the similarly aged healthy individuals walking at matched speeds and their self selected speeds. We then correlated LL (LU) with the kinematic measures during each double limb support phase of the gait cycle. tion between LL and LU and various parameters selected (i.e. knee angle, leg angle, medial lateral GRF). The value of 17 (SPSS, Inc.) was used for statistical analysis Results The control subjects were symmetric (e.g., no statistically significant differences in the LL (LU) and kinematic variables between the two legs). Therefore, the average value of all parameters for both legs was used at all speeds.

PAGE 84

84 Magnitude of Li mb L oading Average magnitude (normalized by body weight) of LL was less in the paretic leg as compared to the non paretic leg (0.5090.102 N/kg vs. 0.6510.146 N/kg, p <0.0001) and to similarly aged healthy control subjects walking at the self selected spee ds (0.5090.102 N/kg vs. 0.5960.0296 N/kg, p = 0.021).There was no significant difference between LL of the paretic leg and control subjects walking at matched speeds (control (0.6m/s/ 0.3m/s) = 0.554 0.117 N/kg, 0.585 0.075 N/kg, p = 0.807 and 0.061). T he reduced paretic LL was related to the functional walking status of individuals with hemiparesis. LL of the paretic leg in the household (0.4590.09N/kg) and limited community (0.5130.069N/kg) ambulatory individuals was less than in the non paretic leg [(0.6800.11N/kg; p =0.001 vs. household ambulators paretic), (0.6460.08N/kg; p =0.001vs. limited community ambulators paretic) and in healthy individuals walking at matched [(0.5850.076N/kg; p =0.001 vs. household ambulators paretic), (0. 5540.117N/kg, p = 0.001 vs. limited community ambulators)] and self selected speeds [(0.5960.373N/kg; p = 0.004 against household ambulators paretic, p = 0.001 against limited community ambulators)]; whereas, the community ambulators did not show any significant difference a s compared to the non paretic leg (0.5260.099N/kg vs. 0.5650.105N/kg) and to the controls (0.5960.373N/kg, p =0.98). ( f igure 3 1 a (individual representative subject, f igure 3 2 a group average)) Magnitude of Limb U nloading The LU was greater (for LU small er value implies greater unloading) in the paretic leg as compared to the non paretic leg (0.3770.085N/kg vs.0.5080.084N/kg, p =0.0001) and similarly aged healthy control subjects walking at the self selected

PAGE 85

85 speeds (0.5110.063N/kg, p = 0.001) and matched speeds (control (0.6m/s/ 0.3m/s)=0.5130.034N/kg/ 0.5040.040N/kg, p =0.001/0.001), as evident from the lower average magnitude of vertical GRF. The increased paretic LU was related to the severity of the gait speed deficit. During the second double limb s upport phase, the LU of the paretic leg in the household (0.3100.06N/kg) and limited community (0.3900.056N/kg) ambulatory individuals was different (note that reduced GRF equals greater LU) from the non paretic leg [(0.5240.09N/kg; p =0.001 against household ambulators paretic), (0.6460.08N/kg; p =0.001against limited community ambulators) and healthy individuals walking at matched [(0.5040.076N/kg; p =0.001 against household ambulators paretic),( 0.5130.04N/kg, p =0.001 against limite d community ambulators) and self selected speeds [(0.5040.063N/kg; p =0.001 against household ambulators paretic, p =0.001 against limited community ambulators)]; whereas, the community ambulators did not show a significant difference as compared to the non paretic leg (0.4380.08N/kg vs. 0.4620.06N/kg) and the controls (0.5040.063N/kg, P= 0.06). ( f igure 3 1 a f igure 3 2 a unloading part of the graph) Timing a nd Pattern o f Limb Loading a nd Unloading In addition to the difference in the magnitude of LL and L U, the timing of LL and LU of the paretic and non paretic legs and healthy controls was significantly different. One way ANOVA revealed that the difference in timing of LL of the paretic leg (LU of the non paretic leg) and LL of the non paretic leg (LU of paretic leg) and healthy controls was significant (p<0.001). The transition of majority of weight from the paretic to the non paretic leg (timing of paretic LU) occurred at 328.2% of the duration of the double limb support phase, which was earlier than t hat from the non paretic leg to the

PAGE 86

86 paretic leg (timing of non paretic LU) (51.69.9%) or control legs (44.54.6%) ( figure 3 4 b ). The above conclusions are drawn from the information gained by the duration of the second double limb support phase of the pa retic (28.0 4.2% of the gait cycle or 0.4270.07s), n on paretic (25.82.9% of the gait cycle or 0.380.09s). This difference in timing of LL and LU was further investigated in each of the functional groups. One way ANOVA analysis revealed that the average LL time of the non paretic leg (transition of the weight from the paretic to non paretic leg) in the household ambulators (22.58.9%) was significantly less (p=0.047 vs. limited community ambulators, and p<0.001 vs. community ambulators) than the limited community ambulators (30.44.8%) and the community ambulators (39.87.7%). Additionally, the LU time of the paretic leg of the household ambulators (57.69.1%) was greater than limited community ambulators (50.18.6%), and the community ambulators (47.16. 9%) (p=0.013 vs. limited community and p=0.006 vs. community ambulators). In addition to the difference in timing, the LL and LU between the paretic leg could be classified into three groups based on their curvatures: for LL; concave curvature (12 subjects 10 Household ambulators, 1 Limited Community am bulator 1 Community ambulator ); convex curvature (20 subjects; 4 Household ambulators, 11 Limited Community ambulators, 5 Community ambulators); and linear (4 subjects; 3 Limited Community ambulators,1 Comm unity ambulator), 8 subjects (6 Limited Community ambulators, 2 Community ambulators) could not be categorized because of very abnormal shape of the GRF curve (Table 2 1).

PAGE 87

87 For LU: concave curvature (31 subjects; 11 Household ambulators, 16 Limited Communit y Ambulators, 4 Community Ambulators), convex curvature (6 subjects; 3 Household Ambulators, 2 Limited Community Ambulators, 1 Community Ambulator ), and Linear (2 subjects; community Ambulators) (Table 2 1). Six subjects could not be categorized because of abnormal shape of LU for their GRF curve (these subjects were different from the subjects mentioned in the pattern of LL analysis above). Household ambulators primarily had concave loading (slow initially then more rapidly 8 out 14 and concave unloading (rapid initially then more slowly 11 out of 14 pattern whereas most of the limited community ambulators had a convex loading (11 out of 21 and a concave unloading pattern (16 out of 21). In addition, the community ambulators did not show any consistent patterns. Relationship b etween Medial Lateral Ground Reaction Forces a nd Limb Loading a nd Unloading Medial lateral forces were not significantly correlated with vertical GRF during both double limb support phases of paretic leg (for first p= 0.91 and second p= 0.342 double limb support) Relationship b etween Knee Angle a nd Limb Loading a nd Unloading There was no statistically significant correlation between average LL and LU and knee angle on the paretic leg, however, there was a trend towards signific ance for LU, suggesting that the individuals with greater unloading (lower average vertical GRF values) may have a greater knee flexion, (r= 0.150, p =0.162 for LL and knee angle; and r= 0.213, p =0.052 for LU and knee angle). The n on paretic leg LL was sig nificantly related to knee angle (r= 0.459, p =0.001) showing that the individuals with greater leg

PAGE 88

88 loading have lesser knee flexion, but the LU showed no significant correlation (r=0.196, P=0.261). Relationship b etween Medial Lateral Leg Angle a nd Limb Lo ading a nd Unloading Subjects with reduced paretic LL placed the paretic leg further lateral relative to the center of mass of the pelvis. LL was negatively associated with the leg angle of the paretic side in the medial lateral direction during the second double support phase of the gait cycle (r= 0.451, p=0.002). The paretic leg medial lateral angle (9.42.6) was always significantly (p<0.05) greater than that of the non paretic (7.32.0) and of healthy controls walking at matched speeds ( f igure 3 1 c f igu re 3 2 c ), demonstrating that the paretic leg was placed more lateral from the center of mass of the body as compared to the non paretic leg and controls Discussion Our results demonstrate that the loading of the paretic leg is significantly less than either the non paretic leg or the control subjects walking at self selected or matched speeds. These results are consistent with the loading abilities of the individuals with hemiparesis during static and quasi static tasks previously reported 7, 8, 10 Furthermore, reduced LL of the paretic leg and the increased compensatory LL of the non paretic leg varies with the severity of stroke (as defined by the self selected walking speed). Our study found that the household walkers had significantly greater loading of their non paretic leg (i.e., greater paretic unloading) during initial double limb support than the limited community a nd community walkers. This compensation by non paretic leg can 31 ,which suggests that greater the impairment of the paretic leg, the greater is the compensation offered by the non paretic leg to maintai n stability of the ongoing task The asymmetry between the

PAGE 89

89 paretic and the non paretic le g in both LL and LU could be attributed to the weakness and decreased control of the paretic leg 7, 177 Physiologically, reduced loading of the paretic leg might be associated with reduced muscle activa tion during walking, which might result in increased impairments during walking, for example, poor support or propulsion; the LU is increased on the paretic side, meaning that the individuals with hemiparesis unload their paretic leg rapidly. In addition to the asymmetry in the magnitude of LL and LU the timing of LL and LU is also important because walking is a cyclic task produced as a result of synchronized phasic stimuli (hip extension and LL and LU). The individuals with hemiparesis, walking at their self selected speed accept body weight much more slowly during the initial double limb support of the paretic leg (than the non paretic leg), whereas during terminal double limb support the weight is rapidly transferred to the non paretic leg. This demonst rates a reluctance to load their paretic leg, resulting in reduced single leg stance phase on the paretic leg and l onger double limb support phase Although, the timing of LL and LU provides details about the LL and LU asymmetry, it does not capture the va riability of biomechanical adjustments made by the individuals with hemiparesis as do the convex, concave or linear shape of the LL and LU curves. The individuals with concave LL pattern were primarily household ambulators, i.e. walking at less than 0.4 m/ s and a majority of the individuals with convex LL pattern were limited community ambulators and the linear pattern comprised both limited community and community ambulating subjects. For 10 subjects who displayed both concave LL and LU, the loading of bot h paretic and non paretic leg was slow, suggesting that they spent a large proportion of the gait cycle in the double limb

PAGE 90

90 support phase and displayed slow loading of both the paretic and the non paretic leg. Similar information about difference in pattern of LL and LU of the paretic and the non paretic leg provides us more details about the changes and variability in walking and LL and LU abilities post stroke. This information might assist in designing better rehabilitation techniques, for example, greate r emphasis could be placed on the weight shifting abilities of the household ambulators to promote faster and more physiologic walking speed. As mentioned before, individuals with hemiparesis make several biomechanical adjustments to minimize the loading o f the paretic leg (i.e. by reducing LL and increasing LU and time to load, joint angle). Knee angle is an important component for supporting the body while accepting weight in order to prevent collapse and has been the most often reported kinematic variabl e in relation to walking ability post stroke 29, 30 Although knee angle was not significantly related to the LL and LU on the paretic leg, the trend towards significance suggests a possible importance of knee angle in understanding LL and LU. We found that the knee was more flexed in the non paretic leg as compared to the paretic leg and healthy controls during both LL and LU. In addition, difference in terminal double limb support could be explained by reduced ability to plantarflex the ankle to create push off power an d the increased knee flexion could be credited to increased power generation requirements of the non paretic leg (at hip, knee and ankle) to compensate for the decreased po wer output from the paretic leg. 28 Additionally, LL and LU in subjects with hemiparesis is influenced by medial lateral leg angle. The results of our study support the observation that the hemiparetic

PAGE 91

91 individuals place their non paretic leg close to their body while placing their paretic leg farther away 143 meaning that the medial lateral leg angle is greater on the paretic leg and significantly less on the non paretic leg. The lateral placement of the paretic leg may be a strategy to increase the base of support while walking or to minimize loading of the paretic leg. Minimizing loading o f paretic leg may be attributed to weakness or impairment, spatial neglect, lack of confidence or fear of falling. Non n eurological c oncerns of i ncreased l oading of the n on p aretic l eg There is evidence in individuals post amputation that suggests osteo a rthritis of the joints in the intact leg (e.g. knee 55 and hip 56 ) because of the changes in the direction of the ground reaction forces and greater leg loading. 57 In addition, subjects with poliomyelitis typically have asymmetrical disease, with greater forces transmitted across the unaffected leg resulting in greater symptoms on the unaffected leg as compared to the affected leg 58 Similarly, greater loading of the non paretic leg might result in musculo skeletal changes which might further interfere w ith the walking ability of the persons with hemiparesis. Therefore, assessment and rehabilitation should not neglect the non paretic leg.

PAGE 92

92 Table 2 1. List of loading and unloading details of hemiparetic individuals ID Speed (m/s) Pattern of LL Pattern of L U Average LL Average LU 1 0.35 concave concave 0.534 0.296 2 0.34 convex concave 0.415 0.294 3 0.21 concave concave 0.463 0.321 4 0.21 concave concave 0.463 0.340 5 0.39 convex concave 0.418 0.280 6 0.19 convex concave 0.400 0.265 7 0.18 concave concave 0.536 0.317 8 0.26 concave convex 0.481 0.444 9 0.35 concave concave 0.380 0.277 10 0.36 concave convex 0.557 0.354 11 0.15 concave concave 0.414 0.248 12 0.35 convex concave 0.270 0.289 13 0.22 concave convex 0.631 0.240 14 0.24 concave concave 0.862 0.388 15 0.4 ** concave 0.696 0.432 16 0.7 convex concave 0.457 0.508 17 0.42 convex concave 0.479 0.436 18 0.45 convex concave 0.466 0.374 19 0.42 concave convex 0.537 0.382 20 0.44 ** concave 0.576 0.378 21 0.46 convex ** 0.422 0.382 22 0.63 convex concave 0.601 0.262 23 0.4 ** concave 0.471 0.379 24 0.73 convex concave 0.523 0.465 25 0.75 convex concave 0.541 0.433 26 0.51 linear concave 0.419 0.308 27 0.57 convex ** 0.458 0.319 28 0.63 convex concave 0.552 0.429 29 0.48 ** concave 0.520 0.356 30 0.65 convex convex 0.491 0.478 31 0.8 ** concave 0.573 0.549 32 0.49 linear ** 0.514 0.350 33 0.42 ** concave 0.424 0.370 34 0.62 convex concave 0.471 0.359 35 0.43 linear concave 0.584 0.295 36 0.99 convex ** 0.050 0.510 37 0.99 ** concave 0.506 0.456 38 1.05 convex concave 0.353 0.516 39 0.84 convex concave 0.545 0.332 40 0.8 ** convex 0.602 0.518 41 0.9 concave linear 0.595 0.430 42 1 convex convex 0.522 0.551 43 0.86 linear linear 0.582 0.363 44 0.87 convex concave 0.619 0.438

PAGE 93

93 a. b. c. d. Figure 3 1. Represents characteristics of one subject in each speed group. a ) the limb loading and unloading b ) knee angle c ) leg angle d ) l oading and unloading pattern in one representative subject of each speed based functional group (column 1 = household ambulator; column 2 = limited community ambulatory; and column 3 = community ambulator. Red is paretic leg, blue is non paretic leg, aqu a blue is control at matched speeds, and green is control at self selected speeds. In 1 d, the black highlight represents the concave pattern, purple shows the convex pattern and the yellow is linear pattern. Household Ambulators Limited Community Ambulators Community Ambulators

PAGE 94

94 a. b. c. Figure 3 2. Represents avera ge characteristics of all subjects in each speed group. a ) Limb loading and unloading b ) knee angle c ) medial lateral leg angle of all individuals in each functional group (based on speed). First column represents household ambulators; second column limite d community ambulators while the third column community ambulators. Red represents paretic leg, blue: non paretic leg, green is control at self selected speed, aqua blue is control at matched speeds.

PAGE 95

95 Figure 3 3 Represents individuals with different loading and unloading patterns against their self selected walking speed. Self selected walking speed is used in this study to categorize individuals into three groups (<0.4m/s= household Ambulators, 0.4 0.8 m/s= Limited Community Ambulators, >0.8 m/s= Co mmunity Ambulators). Each point on the scatter plot represents a hemiparetic individual. X axis represents the different loading and unloading patterns. NOTE: the number of points looks less than the total number of subjects because subjects with similar s peeds are all represented with a single point.

PAGE 96

96 Figure 3 4 Represents the cross over point (circled in black) during paretic loading and unloading during walking.

PAGE 97

97 CHAPTER 4 COMPENSATION AND IMPAIRMENT PATTERNS OF THE NON PARETIC LEG IN HEMIPARETIC GAIT. Background Hemiparesis post stroke is usually considered a unilateral motor control problem of the paretic leg, while the non paretic leg is assumed to have no deficits (sometimes even being used as a reference for comparison of the performance of th e paretic side). Nevertheless, non paretic leg performance differs from that of a healthy leg with many differences presumed of value for compensation 12, 13, 59, 60 and some that might indicate impairment 178 For example, Kim et al 12 reported a high correlation between ankle plantarflexion angle in the saggital plane and ankle push off and hip pull off power bursts in the saggita l plane of the non paretic leg and gait speed; suggesting that the compensatory performance of the non paretic leg itself may be an important contributing factor to the gait performance post stroke. While non paretic leg compensatory mechanisms may a ssist a person to attain a steady walking state even with a discoordinated paretic leg, it may also be the case that these compensations might also inhibit expression of a more appropriate biomechanical pattern of the paretic leg This interference could be eith er a mechanical phenomenon (e.g. doing more than half of the task requirements does not allow expression of a normal pattern without requiring increased output ) or it could be a neural phenomenon (e.g. the altered sensorimotor state of the non paretic leg might interfere with normal bilateral pattern formation ) For instance, pedaling studies in individuals with hemiparesis post stroke ha ve demonstrated a strong negative influence of the non paretic leg participation in pedaling on the paretic leg performan ce, thereby further deteriorating the performance of the paretic leg pedaling performance 11, 89 Thus, it is likely that the altered

PAGE 98

98 sensorimotor state of a compensating non paretic leg may interfere with expression of a normal motor pattern by the paretic leg. If this finding can be extrapolated to walking the performance of the non paretic leg should also be a clinical concern even w hen the focus is to improve performance of the paretic leg Although there is ample evidence to suggest that non paretic leg output is a substantial contributor to the altered walking pattern post stroke, to our knowledge no study has explicitly investiga ted and reported the performance of the non paretic leg during walking after stroke. Through this study, we aim to investigate the performance of the non paretic leg and quantify how it differs from a healthy control leg. We will test each motor pattern ch ange we observe and present the evidence for and against the change representing compensation for the weakness and altered coordination of the paretic leg and/ or a direct impairment. Compensation is a strategy employed to perform an old task in a new manne r following an injury 158 For the purpose of this study, in the context of the non paretic leg, compensation implies the change in the motor control of the non paretic leg, presumably either an increase in output during a normally active period or ac tivity in a normally inactive period, to perform a function usually performed by a healthy contralateral leg that is not being performed by the paretic leg due to weakness and/or discoordination. Impairment, on the other hand, is the deviation from the gen erally accepted standards in the biomedical status of the body and its functions 159 For the purpose of this dissertation, impairment in the non paretic leg is suggested when there is an excess or deficit in the output of any parameter as compared to healthy control subjects, which would be expected to deteriorate the performance of the non paretic leg itself.

PAGE 99

99 In this study we will investigate changes in EMG activity in the non paretic leg relative to a healthy control leg walking at the same speed by evaluating kinetic and kinematic measures during the regions of change. While observation of walking by itself will not allow us to definitely state that non paretic leg impairment exists, we will be able to document changes in biomechanical variables that suggest compensation and demonstrate when the biomechanical consequences of an activation change may deteriorate performance. We will use kinesiological measures, such as EMG, kinetics an d kinematics to evaluate the motor performance of the non paretic leg. We will then associate these changes with the mechanical work done by a joint moment, which is an important measure of motor performance 149 Interpretation of these measures will also be made within the context of the potentially different kinema tics, which may allow muscle activity to have a different consequence than it would have with usual walking kinematics (e.g., if the non paretic leg is placed behind the paretic leg in a step to gait, then plantarflexor activity could produce propulsion ve ry early in the non paretic leg gait cycle as it will spend more of the cycle behind the center of mass of the body). Methods Participants For the purpose of this study sixty individuals with chronic hemiparesis and twenty similarly aged healthy individua ls were tested. The inclusion criteria for the study were: subjects should have hemiparesis secondary to a single onset unilateral stroke; should be able to ambulate independently with or without an assistive device over 10 m on a level surface; have the a bility to walk on a regular basis at least at home; do not have any significant lower extremity joint pain and any major sensory deficits; do not have any significant lower limb contractures and any significant cardiovascular or respiratory

PAGE 100

100 symptoms contra indicative to walking. Criteria for the exclusion of participants from the study were: any orthopedic or neurologic conditions in addition to stroke, or any significant musculoskeletal problems other than stroke that would limit hip and knee extension or a nkle plantar flexion to neutral, or inability to provide informed consent. All participants in the study signed a written informed consent and the study was approved by Institutional Review Board of University of Florida. Subjects were grouped into three g roups based on their self selected walking speed and compared to healthy control subjects walking at matched speeds. Three groups were defined as slow speed: walking at 0 <0.4m/s (compared to controls at 0.3m/s), moderate speed group: walking at 0.4 0.8m/ s (compared to control subjects walking at 0.6m/s) and the fast speed group: walking at >0.8m/s (compared to control subjects walking at 0.9m/s). Procedures Subject preparation Reflective marker spheres and rigid body clusters were positioned on specified bony prominences and limb segments respectively, to acquire 3D motion data. Marker positions were based on the Vicon PlugInGait marker set (modified Helen Hayes set). In addition, disposable bipolar Ag AgCl surface electrodes (Vermed, Inc, bellows Falls, VT) were used to record muscle activity from eight lower extremity muscles bilaterally (TA: Tiabialis Anterior, SO: Soleus, MG: Medial Gastrocnemius, VM: Vastus Medialis, RF: Rectus Femoris, BF: Biceps Femoris LH: Lateral Hamstrings, MH: Medial Hamstrings and GM: Gluteus Medius). Adequate skin preparations were done (electrode placement site was shaved and wiped with alcohol) to facilitate maximum skin conduction. Reference electrode was placed on electrically neutral patella.

PAGE 101

101 Data Collection The subjects walked on the split belt instrumented treadmill (Techmachine, Andrezieux Boutheon, France) for 30 seconds (3 trials) at their self selected walking speed. The treadmill was started at a slower speed and the speed was gradually self selected comfortable treadmill walking speed was attained. While walking on the treadmill subjects wore a safety harness (without any BWS ) that was mounted to the ceiling, and would provide support in case of loss of balance. A physical therapist was also present for close supervision and providing assistance if required. Notes were made if any assistance was used during data collection and those data were not used in the final analysis. All subjects walked without ankle foot orthosis or walking aid. Rest breaks were provided as needed by the subjects between the walking bouts. Additionally, healthy control subjects walked at a self selected speed and four additional walking trials, speeds introduced randomly from slow to normal: 0.3, 0.6, 0.9 an d 1.2 m/s. These speeds were selected a priori to represent a likely range within which persons with stroke would elect to walk. Literature reports normal gait speed is 1.2 m/s and three incremental speeds slower than normal were chosen 15 Data Recording and Processing Twelve camera VICON motion capture system (Vicon Motion Systems, Los Angeles, CA) and instrumented treadmil l and telemetric EMG system (Konigsberg Instruments, Pasadena, CA) were used to capture the kinematic, kinetic and EMG data as the subjects walked on the split belt instrumented treadmill. GRFs were acquired from each foot separately at a sampling rate of 2000 Hz, and low pass filtered using a fourth order, zero lag Butterworth filter with a 20 Hz cut off

PAGE 102

102 frequency. A 13 segment musculoskeletal model was created using Visual 3D (V3D) (C Motion, Inc., Germantown, MD). V3D models were used to conduct inverse dynamics analyses for calculation of intersegment joint kinetics. In addition, the EMG signals were collected as analog signals at 2000 Hz, which were then converted to digital signals and stored on the laboratory computer. The analog signals were amplifie d and pre filtered before being digitized and stored. The EMGs were then filtered with a 40 Hz high pass filter, demeaned, rectified and then low pass filtered (25 Hz). Filtering was done with 4th order Butterworth filters in Matlab (MathWorks Inc, Natick, MA). Data were processed with K means cluster analysis (figure 4 3 ) to determine whether the muscle was active or not for each point in the gait cycle 125 K means cluster analysis is a statistical technique that groups similar data points in a single cluster. In between the data points of the rectified and filtered EMG signal 125 The individual data points are assigned to k clu ster s by MATLAB such that the distance from the centroid is minimized. Therefore, the most similar data points are grouped together. The c luster with the lowest mean value corresponds to inactivity and the other clusters correspond to periods of muscle activity 125 Calculations of S tudy V ariables The same variables were recorded and calculated for healthy individuals and individuals with hemiparesis during walking. All the variables were calculated by averaging across all gait cycles of each trial. Since there was no statistically significant diffe rence between the performance of the left and right side of the healthy control subjects, the mean value of the two sides was used for all the analys e s.

PAGE 103

103 EMG Variables EMG magnitude. Integrated EMG was used to quantify the magnitude of EMG activity over the gait cycle and for different regions of the gait cycle. It was calculated by numerically integrating absolute EMG signals using an extended open closed formula with equally spaced abscissas with respect to time (figure 4 2 ) EMG timing. Percentage durati on of activity of a muscle in each gait cycle/region is sum of all the "on" points divided by the total number of points in the gait cycle/region. Kinetic and kinematic variables Total positive work done by the ankle plantarflexor s hip extensor s and hip flexors was calculated in different regions of the gait cycle as defined in figure 4 1 Kinematics were calculated similarly for the same regions of the gait cycle Statistical A nalyses Independent samples t test were performed to compare the magnitude and timing of muscle activity in the non paretic leg: 1) over the entire gait cycle, and 2) in different regions of the gait cycle to healthy control subjects walking at matched speeds. Th e slow speed group (0 <0.4m/s) was compared to control subjects at 0.3m/s, the moderate speed group (0.4m/s 0.8m/s) were compared to control subjects at 0.6m/s and the fast speed group (0.8m/s 1.0m/s) were compared to control subjects walking at 0.9m/s. percentage average activation for the muscle in a specific region and the measures of motor performance Significance for all tests was set at alpha < 0.05. All statistics were run using SPSS version 1 7 .0 (SPSS, Inc.).

PAGE 104

104 Results The results of the analyses reveal that the muscle activation in the non paretic leg is significantly different than muscle activation of healthy control subjects during walking. In addition, we also performed a detailed analysis of the muscle activity and motor performance measures to ensure correct inference of the non paretic leg output. Magnitude of EMG Activity Integrated EMG activity over the gait cycle. The dominant feature of the magnitude of integrated EMG activity in the non paretic leg gait cycle was an increase in the majority of muscles. Figure 4 4 presents the magnitude of integrated EMG activity in the non paretic leg gait cycle as compared to the healthy control subjects walking at the matched spe eds. Five of the eight recorded muscles in the slow and moderate speed group s (MG, VM, RF LH, MH) had significantly increased integrated EMG activity in the non paretic leg while the fast speed group had significantly incre ased activity in MG, VM, RF, LH MH, GM. Integrated EMG activity in different regions of the gait cycle in slow speed group. The individuals in the slow speed group revealed significantly increased integrated EMG activity in quadriceps (RF and VM) muscles throughout stance phase in the non paretic leg (p< 0.05), w hile all the muscles had significantly increased activity in the first region of the gait cycle ( figure 4 5 ). Integrated EMG activity in different regions of the gait cycle in moderate speed group. The upper leg muscles were mo stly increased while the distal muscles were mostly unaffected. The quadriceps and hamstring muscles had significantly increased magnitude throughout the stance phase of gait cycle (except RF in region one) TA and SO did not reveal significant difference in any region of the gait cycle

PAGE 105

105 (except SO in region six ) while the magnitude of MG was increased in first, second and third regio n of the gait cycle (figure 4 6 ). Integrated EMG activity in different regions of the gait cycle in fast speed group. The upp er leg muscles had significantly increased magnitude in the stance phase of gait cycle. Furthermore, MG also had significantly increased activity in the stance phase (SO only region four) of the gait cycle All muscles (except TA) had increased activity in the fourth region, while there was almost no difference in the magnitude of muscle activity in fifth and sixth region (swing phase) of the gait cycle (except VM in region 5 and GM in region 6) (figure 4 7 ). Timing of EMG Activity Percenta ge duration of mu scle activity for the entire gait cycle Only the subjects in the slow speed group showed significant changes (all muscles increased, TA activity reduced) in the percent duration of muscle activity for the entire gait cycle. Unlike the magnitude measures i n which there were many changes, the individuals in moderate and fast speed groups had no significant difference in the percentage duration of activity as compared to healthy control subjects (except SO, p=0.009) (figure 4 8 ). Percentage duration of muscle activity in different regions of the gait cycle Individuals in slow group revealed most differences in the muscle activity, specifically most muscles revealed significant difference in the sixth and first region of the gait cycle. LH revealed significantly increased activity in all the regions of the gait cycle ( except region five) (figure 4 9 ) On the other hand the duration of activity in individuals walking at moderate speed revealed significant difference in the duration of activity in SO and LH during the stance phase (figure 4 10 ). Participants walking at faster speeds

PAGE 106

106 did not differ from the healthy control subjects walk ing at the matched speeds except for SO and LH in region four and MH in the second region of the gait cycle (figure 4 11 ). Correlation A nalysis H1 and LH & MH in region one of the gait cycle. In the slow and fast speed group s the magnitude of LH activity w as significantly positively (r= 0.536, p=0.002 and r = 0.966, p= 0.008) related to H1. However, individuals in the moderate speed group did not show any significant relation between any of the hamstrings muscles and positive work at hip in region one Howe ver, the duration of activity of MH in the first region had a significant negative correlation with H1 (r = 0.433, p = 0.01). SO and MG muscle activity and A2 in region four of the gait cycle. Although the duration of SO (also fast) and MG (also slow) in the fourth region of the gait cycle was significantly greater in the moderate speed groups, there was no significant correlation between the muscle activity and the positive work done at the an kle in that region of the gait cycle. Furthermore, the increased magnitude of SO in the fourth region of the moderate and fast speed group also did not reveal any significant correlation. TA muscle activity and A2 in region four of the gait cycle. There wa s no significant correlation between the magnitude and duration of the TA activity in region four in all the three speed groups. RF muscle activity and peak knee flexion angle in region four and five of the gait cycle. The percentage duration of the RF act ivity in the fourth region of the gait cycle was significantly negatively correlated with the peak knee flexion in individuals in the moderate speed group (r = 0.464, p = 0.03). However, no significant correlation was found in individuals in the slow and fast speed groups.

PAGE 107

107 MH and LH muscle activity and positive hip extensor work in region three and four A significant correlation was found between the magnitude of LH and the positive hip extensor work (p=0.536, r= 0.002) in individuals walking at the slow speed. However, no significant correlation was revealed in any other group. MG and SO muscle activity and leg angle in region one No significant correlation was present between the increased plantarflexor activity and the leg angle in the first region of the gait cycle. Discussion This study demonstrates that the muscle activity of the non paretic leg differs from healthy control leg of subjects walking at matched speeds and many of these changes appear to be compensatory in nature. This is because the activity is either increased in the muscles and regions known to be the primary power bursts (e.g., H1, A2 and H3) or in other specific regions of the gait cycle that revealed statistically significant correlation with the measures of motor performance. Th is implies that individuals walking post stroke employ several compensatory mechanisms in order to attain a steady walking state. Some of these com pensations are simply increased output of the usual dominant power bursts of a healthy leg, such as increased activity in the plantarflexors during late stance and the hip extensors during early stance. Others are novel adaptations in the non paretic leg s uch as hamstrings activity in late stance to generate propulsion, early stance plantarflexor activity that generates propulsion because the non paretic leg is not placed as far in front of the body, and mid to late stance activity in the knee extensors tha t facilitates paretic leg swing. With respect to limitations of identifying impairment, it is important to emphasize that the performance of the non paretic leg was evaluated during bilateral walking and

PAGE 108

108 not in an experimental condition specifically design ed to test the possibilities of non paretic leg motor control. Nevertheless, we explicitly defined compensation and impairment for our study. Furthermore, because all of the muscles as well as the kinetic and kinematic measures influence walking and are no t independent, we did not correct for multiple comparisons in our study. Since the results of the statistical analyses were mostly in agreement with our initial hypotheses that the muscle activity will be increased in the non paretic leg based on some evid ence in the literature, as opposed to many more random appearing changes in different directions, we feel that this choice was justified. Nonetheless we believe that the reader can appropriate ly assess importance of the changes observed in the coordinatio n of the non paretic leg. Muscle Activity over the Gait Cycle Integrated EMG and percentage duration of activity over the gait cycle demonstrate that in general the muscles in the non paretic leg are activated at a significantly greater magnitude and for a significantly greater duration than those of the healthy control subjects walking at matched speeds. This therefore suggests the non paretic leg should not be considered unaffected or performance of the paretic leg. The pa ttern of difference in the global measures of EMG between the non paretic leg of individuals in the three speed groups and the control subjects at matched speeds suggests that during walking the individuals with severe stroke had greater changes in the mot or pattern of the non paretic leg than the individuals with less severe stroke. F or instance the total integrated muscle activity was significantly greater in five five and six muscle s in the non par etic leg of individuals in the slow, medium and fast sp eed groups respectively Furthermore only individuals post stroke categorized in the slow speed

PAGE 109

109 group revealed significantly greater percentage duration of activity in all eight recorded muscles (except TA, TA had reduced percentage duration of activity) N o difference was present in moderate and fast speed walkers. T he overall increased percentage duration of activity of different muscles over the gait cycle in the non paretic leg suggests that these muscles are firing for greater percentage duration, wh ich could either indicate compensation for the impaired coordination of the paretic leg after stroke, or impairment of the non muscles efficiently, during walking. To discern the role of the non paretic leg in compensation vs. impairment we investigated the muscle activity in the different regions of the gait cycle. Muscle A ctivity during D ifferent R egions of the G ait C ycle Reduced propulsion and thus reduced walking speed is one of the primary concerns after stroke. The non paretic leg typically generate s greater propulsive force 179 to compensate for the weak and discoordinated paretic leg. Several mechanisms are thus involved. One compensatory me chanism identified was increased hamstrings activity during early stance, presumably to increase propulsion by positive work done at the hip, i.e. H1. To this end, the results of our analysis reveal that the magnitude and duration (except in the fast speed group) of LH and MH are significantly increased in the first region of the gait cycle of the non paretic leg (presumably contributors to the H1 power burst) The increased firing of the MH and LH muscles facilitate an increase in the magnitude of the usua lly dominant positive H1 power burst to increase propulsion. Simulation analysis reported an important role of the hamstrings in forward propulsion of the trunk 163 (instead of support and stability 163, 180 ) duri ng early stance. Thus increased

PAGE 110

110 hamstring activity might compensate for reduced propulsion by promoting forward acceleration of the trunk at a time when the paretic leg is in pre swing and is not contributing as much propulsion as a healthy leg normally wo uld This is further supported by a significant positive correlation between the positive work done by the hip extensors H1 and the hamstring (MH and LH) activity during this phase of the gait cycle. Increased activity in the plantarflexors earlier in the stance phase was likely a novel compensation to produce increased propulsion that might possibly be facilitated by the altered asymmetric kinematics. The increased plantarflexors activity in this region of the gait cycle may facilitate propulsion if the no n paretic leg is not placed ahead of the body cente r of mass (as happens in a step to gait 160 ). Studies have suggested that shorter non paretic leg step results in reduced paretic propulsion (because the paretic plantarflexors are put in a biomechanical disadvantage position 160 ). Therefore, the increased activity of the SO and MG in this region of the gait cycle may compens ate for the reduced propulsion generation. Although as a group there was no significant correlation between the non paretic leg angle and the plantarflexor activity in first region, there was a significantly greater positive work done by the plantarflexors in this region, particularly in individuals walking at slow speeds. Thus, those at slow speeds were in a kinematic position where they were able to plantarflex the ankle in this region, which is normally characterized by dorsiflexion. Another potentially novel compensation by the non paretic leg is increased VM and RF activity in the second and third region of the non paretic gait cycle in the slow and moderate speed group individuals walking post stroke According to a rece nt simulation study the increased VM activity in the stance phase of the non paretic leg

PAGE 111

111 might facilitate the propulsion o f the swinging paretic leg 181 They also reported that the contribution of RF and VM was incre ased in the non paretic leg of group of slow walkers (limited community ambulatory) as compared to group of fast walke rs (community ambulatory) 181 consistent with the increased VM activity in slow and moderate speed group and not in the fast speed group of our analyses. The increased VM firing in the stance phase increases propulsion in the paretic leg to maintain a walking state. Furthermore, the increased VM and RF activity in the stance phase of the non paretic leg also generates greater force to maintain the stability, while the total body weigh t is being supported. A ccording to the simulation analysis, SO is the primary muscle to maintain the limb in extension to support bod y weight during the second and third regio n 163 However, our analysis show s that the activity of SO was reduced in the slow and moderate speed group therefore VM and RF contract to compensate 161 A third likely novel compensation in the non paretic leg muscle activ ity is a much more pronounced hamstring burst in the late stance phase of the gait cycle that appears to be associated with increased propulsion A s imilar pattern of activity has bee n reported by Shiavi et al. 69 and Buurke et al 126 in individuals post stroke and in some studies in i ndividuals post amputation 182 Based on the information obtained from the amputation literature, Buurke et a l. 126 suggested that th e contraction of hip extensors (hamstrings ) helps in generation of extension of the non paretic limb, which might thereby facilitate the swing of the paretic leg, thus co mpensating for the poor plantarflexors and hip flexors. Generation of an extra extension force on the non paretic leg, places the paretic leg in a mechanically advantageous position, which facilitates its swing without actually transferring energy to the s wing leg. This, atypical power burst

PAGE 112

112 (found on the non paretic side of individuals post stroke) increases the propulsion of the otherwise weak and discoordinated paretic leg. This phenomenon is further supported by a significant positive correlation betwee n the positive hip extensor work and the MH and LH activity during the third region of the gait cycle in the slow speed group However, no significant correlation was found in the moderate and fast speed group despite significantly increased magnitude of t he LH and MH muscle. This could be due to high variability in the slow speed group whereby only some individuals extensively employ this mechanism in order to maintain a steady walking state. An expected compensation that was observed was increased plantar flexor output during the usual A2 power burst. The magnitude and duration of the plantarflexors were increased in the fourth region of slow and moderate speed group individuals in order to increase the magnitude of A2. A2 is the most important source of p ositive work generation in the gait cycle and thus helps to generate greater propulsive force to attain faster walking speed. This greater contraction of the plantarflexors on the non paretic side generates greater force t o produce increased propulsion to maintain steady walking s tate In addition to increased plantarflexor activity there is an additional burst of activity in the TA during the fourth region of the gait cycle. In theory, the increased magnitude of activity of TA in this region of the gait cy cle may interfere with generation of propulsion during this phase 160 An increased burst in the RF muscle during the pre swing phase (both duration and timing) was exhibited in the non paretic leg of individuals in the slow and moderate speed groups. Excessive activity of RF has been cited as a contributor to reduced knee flexion 183, 184 Furthermore, RF tends to decrease the peak knee fle xion velocity 170, 171

PAGE 113

113 and over activity during region four and five is often implicated in stiff knee gait 170 Additionally, Hernandez et al. 185 in their electrical stimulation study reported that the stimulation of the RF before the toe off largely reduced the peak knee flexion during swing as opposed to the stimulation during the swing phase. Interference with knee flexion is supported by a significant negative correlation between the peak knee flexion angle and the RF activity in the moderate speed group. However, lack of any significant correlation in the slow speed groups could be attributed to the increased contribution by other muscles ( MG iliopsoas, and hamstrings) to knee flexion. Implications of A nalysis The conventional or novel c ompensat ory mechanisms by the non paretic leg are aimed at either providing support, stability or fo r providing greater propulsion. This ma y be accomplished by increase in normally ex isting sources of positive work. For instance, increased hamstring activity in the first region of the gait cycle provides greater forward propulsion of the trunk 163 Additionally, increased activity of the plantarflexor muscles (MG and SO) also facilitate generation of greater propulsion of both the trunk (SO) 162 and leg (MG) 162 in the fourth region of the gait cycle. Additionally, novel power generation pat terns of the non paretic leg: increased activity of the VM and RF in the second and third region of the gait cycle provides increased support to fully loaded non paretic leg 161 I n addition it has also been attributed to increase the propulsion of the swinging paretic leg 181 Finally, the extra burst of the MH in the late stance or third an d fourth region of the gait cycle generates extensor moment and provides biomechanical advantage to the paretic leg, thereby compensating for the weak plan tarflexors and hip flexors 126, 182

PAGE 114

114 On the other hand the non paretic leg exhibits some increases in muscle activity which may not have a positive influence on walking and are might suggest impairments. For example, increased activity of TA in the pre swing phase ( r egion four) of the gait cycle, that likely interferes with generation of propulsion by the non paretic leg 160 and the increased burst of RF during the fourth and fifth region of the gait cycle, possibly negatively influences knee angle and knee flexion velocity during the swing phase 170, 171, 185 Clinical R elevance The clinical relevance of this study lies in providing evidence for the altered performance of the non paretic leg post stroke (which is usually considered unaffected). Therefore, the results of this study emphasize that stroke is not a unilateral motor control problem as often suggest ed, and there is a bilateral involvement. This study presents compensatory mechanisms of the non paretic leg and understanding of these mechanisms is important to provide effective rehabilitation. For example, increased MH activity in late stance increases extension of the non paretic leg, which in turn provides mechanical advantage to the paretic leg to swing the leg through without greater recruitment of muscle fibers. However, if the excessive extension of the non paretic leg is prevented during therapy, it would facilitate greater use of the paretic leg plantarflexors a n d hip flexors to swing the leg in order to attain a functional walking state. In summary, to improve walking after stroke, it is important not only to investigate and understand the mecha nics of paretic leg output, but understanding of the non paretic leg performance provide s additional insights.

PAGE 115

115 DLS SLS SLS DLS SWG SWG Region 1 Region 2 Region 3 Region 4 Region 5 Region 6 Figure 4 1. Represents the different regions of the gait cycle. At the top are the marked the different physiological phases of the gait cycle (DLS= double limb support, SLS= single leg support, SWG= swing phase) and at the bottom are the corresponding regions used in this study.

PAGE 116

116 Figure 4 2. Represents the calculation of integrated EMG activity. F irst figure represents the raw EMG signal, second panel represents the rectified and filtered EMG signal and the third figure represents integration rule with equally spaced abscissas (red lines) with resp ect to time in the gait cycle

PAGE 117

117 Figure 4 3. Represents the steps of k means analysis. First figure represents the raw EMG magnitude, second figure shows the filtered rectified EMG (linear envelope) and the third figure represents the original raw EMG signal with the detected periods of a ctivity indicated in gray shaded areas

PAGE 118

118 Figure 4 4. Integrated EMG activity for the entire gait cycle in the different groups is plotted to represent the magnitude of muscle activity. Group 1 is slow speed group, group 2 is moderate speed group and group 3 is fast speed group. Y axis represents Integrated EMG activity over the gait cycle in Vs and x axis represents different muscle (TA= Tibialis Anterior, MG= Medial Gastrocnemius, SO= Soleus, VM= Vastus Medialis, RF= Rectus Femoris, LH= Lateral Hamstrings, MH= Medial hamstrings, GM= Gluteus Medius). Red bars represent the activity of the non paretic leg and Blue bars represent the control subjects walking at matched speeds.

PAGE 119

119 Figure 4 5. Represents intergrated EMG activity in different regions of the gait cycle in individuals in slow speed group. Each panel of plots represents different regions of gait cycle, ordered first to sixth. Y axis represents integrated EMG (Vs), x axis is differ ent muscles (TA= Tibialis Anterior, MG= Medial Gastrocnemius, SO= Soleus, VM= Vastus Medialis, RF= Rectus Femoris, LH= Lateral Hamstrings, MH= Medial hamstrings, GM= Gluteus Medius). Red bars represent the activity of the non paretic leg and Blue represent the control subjects walking at matched speeds.

PAGE 120

120 Figure 4 6. Represents intergrated EMG activity in different regions of the gait cycle in individuals in moderate speed group. Each panel of plots represents different regions of gait cycle, ordered first to sixth. Y axis represents integrated EMG (Vs), x axis is different muscles (TA= Tibialis Anterior, MG= Medial Gastrocnemius, SO= Soleus, VM= Vastus Medialis, RF= Rectus Femoris, LH= Lateral Hamstrings, MH= Medial hamstrings, GM= Gluteus Medius). Red bars represent the activity of the non paretic leg and Blue represent the control subjects walking at matched speeds.

PAGE 121

121 Figure 4 7 Represents intergrated EMG activity in d ifferent regions of the gait cycle in individuals in fast speed group. Each panel of plots represents different regions of gait cycle, ordered first to sixth. Y axis represents integrated EMG (Vs) x axis is different muscles (TA= Tibialis Anterior, MG= Medial Gastrocnemius, SO= Soleus, VM= Vastus Medialis, RF= Rectus Femoris, LH= Lateral Hamstrings, MH= Medial hamstrings, GM= Gluteus Medius). Red bars represent the activity of the non paretic leg and Blue represent the control subjects walking at matched speeds.

PAGE 122

122 Figure 4 8 Represents percentage duration of EMG activity in different speed groups (slow, moderate and fast speed). Y axis represents p ercentage duration activated over the gait cycle in Vs and x axis represents different muscle (TA= Tibialis Anterior, MG= Medial Gastrocnemius, SO= Soleus, VM= Vastus Medialis, RF= Rectus Femoris, LH= Lateral Hamstrings, MH= Medial hamstrings, GM= Gluteus Medius). Red bars represent the activity of the non paretic leg and Blue bars represent the contro l subjects walking at matched speeds.

PAGE 123

123 Figure 4 9 Represents percentage duration of EMG activity in different regions of the gait cycle in indi viduals in slow speed group. Each panel of Bar plots is different region of the gait cycle, from 1 6. Y axis represents p ercentage duration activated in region of gait cycle and x axis r epresents muscle (TA=Tibialis Anterior, MG= Medial Gastrocnemius, SO= Soleus, VM=Vastus Medialis, RF= Rectus Femoris, LH=Lateral Hamstrings, MH=Medial hamstrings, GM= Gluteus Medius). Red bars represent the non paretic leg and Blue bars represent the control subjects walking at matched speed

PAGE 124

124 Figure 4 1 0 Represents percentage duration of EMG activity in different regions of the gait cycle in individuals in moderate speed group. Each panel of Bar plots is different region of the gait cycle, from 1 6. Y axis represents p ercentage duration activated in region of gait cy cle and x axis r epresents muscle (TA=Tibialis Anterior, MG= Medial Gastrocnemius, SO= Soleus, VM=Vastus Medialis, RF=Rectus Femoris, LH=Lateral Hamstrings, MH=Medial hamstrings, GM= Gluteus Medius). Red bars represent the non paretic leg and Blue bars represe nt the control subjects walking at matched speeds.

PAGE 125

125 Figure 4 11 Represents percentage duration of EMG activity in different regions of the gait cycle in individuals in fast speed group. Each panel of Bar plots is different region of the gait cycle, from 1 6. Y axis represents Percentage duration activated in region of gait cycle and x axis r epresents muscle (TA=Tibialis Anterior, MG= Medial Gastrocnemius, SO= Soleus, VM=Vastus Medialis, RF=Rectus Femoris, LH=Later al Hamstrings, MH=Medial hamstrings, GM= Gluteus Medius). Red bars represent the non paretic leg and Blue bars represent the control subjects walking at matched speeds.

PAGE 126

126 CHAPTER 5 RESPONSE OF NON PARETIC LEG TO LOCOM OTOR TRAINING POST STROKE Background Stroke is one of the leading causes of death and long term disability in the United States of America 4 75% of individuals affected by stroke have some level of long term disability 4 and less than 50% individuals can gain independent walking status 15 Ability to walk independently is one of the most common stated goal of individuals post stroke 20 and therefore, is the primary concern of therapists during rehabilitation. Most rehabilitation techniques focus primarily on the performance of t he paretic leg and the non paretic leg is ignored without much concern about the mechanisms used by the non paretic leg to compensate for the weakness and discoordination of the paretic leg. In chapter 4 of this dissertation we investigated the performanc e of the non paretic leg during walking post stroke and revealed several expected and novel compensatory mechanisms employed by the non paretic leg to accomplish walking. One conclusion from that study is that it is likely that the increased compensation b y the non paretic leg would inhibit development of biomechanically correct patterns in the paretic leg. Therefore, the altered performances of the non paretic leg should receive clinical attention during rehabilitation that seeks to improve paretic leg mot or output We believe that it is important to emphasize that allowing altered performance of the non paretic leg while promoting proper paretic leg movement is likely not optimal for developing coordinated walking patterns. For example, increased activity by the non paretic hip extensors in the stance phase ( i.e., compensation by non paretic leg) might result in reduced employment of the ankle plantarflexors and hip flexors of the paretic leg during the swing phase. If so training that reduc es the non pare tic hip extension in

PAGE 127

127 the late stance phase might result in better expression of the paretic leg muscle s Furthermore, evidence in the pedaling literature 11, 89 suggests that the impaired motor pattern of the paretic leg during pedaling arises in part from generation of non paretic leg motor activity and the interlimb afference associated with non paretic leg loading and moveme nt. Therefore, it is important to focus on both the paretic and the non paretic leg in the rehabilitation regimen to modulate the activity of both and achieve optimal output. Muscle activity or EMG is one of the most commonly used measures to evaluate mot or coordination. There have been very few studies which have reported changes in muscle coordination during walking after rehabilitation following stroke. One of the early studies was reported by Shiavi et al 69 whereby they reported abnormal muscle activation patterns in both the paretic and the non paretic leg. Lately, studies have reported no change in the temporal characteristics of muscles in both the paretic and non paretic leg following rehabilitation 126 Another study reported an increase in the magnitude of semimembranosus muscle after training, in both low response and high response group and additionally, the high response group demonstrated an increase in the average intensity of the soleus muscle 147 While these results may not be in conflict, note that there were difference s in training methodology the use of task specific locomotor training (LT) i.e. body weight supported treadmill training (BWSTT)) in the latter study and the use of N euro Developmental therapy (NDT) in the former study. However, to our knowledge, no study has reported changes in non paretic leg muscle activity after a task specific training, like the BWSTT.

PAGE 128

128 BWSTT is a task specific rehabilitation technique based on th e principles of activity dependent plasticity and focuses on the capacity to capitalize on inherent central nervous system plasticity by creating a permissive environment in which the focus of therapy shifts from compensatory mechanisms to recovery of func tion through production of normalized movement patterns 108, 123, 186 1 89 BWSTT allows spatio temporally and kinematically effective lower limb stepping practice with both legs, with adequate body weight support to facilitate proper trunk and limb alignment and weight shift during stepping. The primary goal of the present study is to investigate if the motor pattern of the non paretic leg (identified in the second study) can be change d following a task specific locomotor training. We will evaluate if the pattern of the non paretic leg normalizes following the locomotor training such that the compensatory mechanisms are reduced Methods Subject Demographics Individuals with chronic s troke were recruited for this study at the Department of Veterans Affairs Medical Center, Gainesville, FL. Thirty individuals with chronic hemiparesis (19 males; 60.213.2 (SD) years of age; affected side left 18) participated in the study. Data for healt hy control subjects was not collected for this study; instead, we used the data previously reported from our laboratory ( chapter 4 ). The healthy subjects did not undergo any training sessions, but only went through full biomechanical testing. Inclusion cri teria for persons post stroke included; hemiparesis secondary to a single unilateral stroke and older than 18 years of age with a stroke within last five years (> 1 year). The subject should ha d to residual paresis in the lower extremity, Fugl Meyer

PAGE 129

129 score < 34, but ability to sit unsupported for 30 s and be able to walk at least 10 feet with maximum one person assistance. In addition the subject needed to have a self selected speed of <0.8 m/s and be able to follow three step commands and provide informed c onsent. Whereas, subjects who lived in nursing home s prior to the stroke or were unable to walk 150 feet prior to stroke had/ have any neurological or cardio vascular or orthopedic injury other than stroke were excluded from the study. Subject Classificati on All subjects were classified into two groups before training based on their self selected walking speed and compared to healthy control subjects walking at matched speeds. The groups before training; slow = 0 < 0.4 m/s (compared to control subjects walk ing at 0.3 m/s n= 17 subjects ) and moderate = 0.4 0.8 m/s (compared to control subjects walking at 0.6m/s n=13 subjects ). After LT all the subjects were grouped into three speed based groups and again compared to healthy control subjects. The groups af ter training were defined as: slow = 0< 0.4m/s (compared to control subjects walking at 0.3 m/s n=6 subjects ), moderate = 0.4 0.8 m/s (compared to control subjects walking at 0.6 m/s n= 17 subjects and fast >0.8m/s (compared to control subjects walkin g at 0.9 m/s n=7 subjects ). Furthermore, individuals in the moderate group were grouped into two groups a) subjects who increased their speed post LT (slow to moderate ; n=13 ) and moved up one group, b) participants who increased their speed post LT, but w ere still grouped in the moderate speed category (n=4) Intervention Subjects post stroke underwent 36 sessions of LT on a BWSTT system Three sessions each week for twelve weeks were provided. Missed sessions were compensated for at the end. Each session included 20 minutes of stepping on treadmill

PAGE 130

130 (TM) and 20 30 minutes of training over ground. Rest breaks were provided as required. The t otal duration of training session was 60 75 minutes. Vitals were measured before each session after every five minutes of stepping during the session and ensure safety Subjec ts wore a harness which was hooked to the overhead motorized system which provided body weight support (BWS) to facilitate stepping (Robomedica Inc, Mission Viejo, CA). The level of BWS was adjusted to promote upright standing posture while ensuring maximu m loading of both the legs. In addition trainers and physical therapist s provided requisite assistance at the trunk and both paretic and non paretic leg to mimic kinematics of the normal walking pattern. In the beginning of the training the non paretic le g was provided greater guidance by the leg trainer which was reduced gradually over time The guidance was aimed at lateral and anterior posterior placement of the foot. Individuals post stroke tend to rapidly place their non paretic leg close to the midl ine to facilitate greater loading of the non paretic leg, while reducing the loading of the par etic leg (as indicated by chapter 3 ). Therefore, the non paretic leg trainer provides guidance to place the foot laterally to reduce the loading of the non paret ic leg, while trying to increase the loading of the paretic leg. Furthermore, the anterior posterior guidance was provided not only to achieve adequate step length, but also to increase the non paretic leg swing time, which thus facilitates increased paret ic single leg support phase. This promotes symmetry while also increasing paretic loading. In addition the anterior posterior guidance by the non paretic leg trainer also ensured adequate hip flexion during the s wing phase. This

PAGE 131

131 would reduce the hip extens or activity in the non paretic stance which provides compensation to weak paretic hip flexors a nd plantarflexors during swing. This might result in forced use of the paretic leg hip flexors and plantarflexors. Furthermore, the non paretic leg trainers also cued the knee to extension and flexion from stance to swing respectively. Some participants tended to maintain their non paretic knee in flexion, which may be detrimental to stability during weight bearing (see chapter 3) In the initial few sessions, ass istance to the non paretic leg was typically provided throughout the training sessions. However, once the participants developed adequate control and pattern (usually after a few sessions) the amount of assistance to the non paretic leg was subsequently r educed. In addition to the efforts by the leg trainer s the trunk trainer also worked in collaboration to acquire normal walking pattern. Trunk trainers controlled the weigh t shift, by regulating the amount (decrease it) and time of loading (see chapter 3) to the non paretic leg. They also prevented the non paretic hip from hyperextending during stance. In doing so they ensured appropriate movement at the pelvis. As the subjects progressed through the training, focus was on an independent and normal walking pattern. BWS was reduced gradually, treadmill speeds were increased and the assistance provided by the trainers on both legs and the trunk was reduced. Biomechanical D ata C ollection All subjects underwent a full scale biomechanical testing before the onse t of training, at the end of training period. Training consisted of thirty six sessions of manually assisted body weight supported treadmill training. Biomechanical data collection was conducted within one week before and after thirty six sessions of locomotor training. Participants walked on a split belt instrumented

PAGE 132

132 treadmill (Techmachine, Andrezieux Boutheon, France) at their self selected comfortable walking speed. No ankle foot orthosis or assistive device was used during data collection. However, all participants wore a harness, mounted to the ceiling of the lab, which eliminated the risk of falling and could hold the subjects in case they lost thei r balance. A physical therapist was also present at all times for vigilance and providing assistance if required. Notes were made, if assistance was used and those data were not used in the final analysis. All the participants walked at their self selected speed on the split belt instrumented treadmill for 30 seconds (3 trials). The treadmill was started at a slower speed and then speed was gradually adjusted until the subjects reached their self selected comfortable walking speed. Data R ecording and P rocessing Twelve camera VICON motion capture system (Vicon Motion Systems, Los Angeles, CA) and instrumented treadmill and telemetric EMG system (Konigsberg Instruments, Pasadena, CA) were used to capture the kinematic, kinetic and EMG data as the subjects walked on the split belt instrumented treadmill. Ground reaction forces were acquired from each foot separately at a sampling rate of 2000 Hz, and low pass filtered using a fourth order, zero lag Butterworth filter with a 20 Hz cut off frequency. A 13 seg ment musculoskeletal model was created using Visual 3D (V3D) (c motion, Germantown, MD) processing that fits the model to marker trajectories. V3D models were used to conduct inverse dynamics analyses for calculation of intersegmental joint kinetics. The E MG signals were collected as analog signals, and were then converted to digital signals and stored on the lab computer. The analog signals are amplified and pre filtered before being digitized and stored. The EMGs are then filtered with a 40 Hz high pass f ilter, demeaned, rectified and then low

PAGE 133

133 pass filtered (3 Hz). Filtering was done with 4 th order Butterworth filters in Matlab (Mathworks, Inc., Natick, MA) Calculations of S tudy V ariables The s ame variables were recorded and calculat ed for healthy individ uals and subjects with hemiparesis during walking. All the variables were calculated by averaging across all gait cycles of each trial. Since there was no statistically significant difference between the performance of the left and right side of the health y control subjects, mean value of the two sides was used for all the analysis. EMG v ariables EMG magnitude. Integrated EMG was used to quantify the magnitude of EMG activity over the gait cycle and for different regions of the gait cycle (figure5 1) It wa s calculated by numerically integrating absolute EMG signals using an extended open closed formula with equally spaced abscissas with respect to time (figure 5 2) EMG timing. Percentage duration of activity of a muscle in each gait cycle/region is sum of all the "on" points divided by the total number of points in the gait cycle/region. means analysis technique 125 (figure 5 3) Kinetic and kinematic variables Total positive work done by the ankle pl antarflexor in region four hip flexors in region four and hip extensors in region one, three and four was calculated. Kinematic parameters of leg and knee were calculated during different regions of the gait cycle.

PAGE 134

134 Statistical A nalyses Independent samples t test and Mann Whitney U test were performed to compare the magnitude and timing of muscle activity in the non paretic leg pre and post LT to healthy control subjects walking at speeds matched to pre and post LT walking speeds (respe ctively, subject classification described above). Comparisons were made : 1) over the gait cycle, 2) in different regions of the gait cycle (subject grouping described in methods section above) Thus, for each muscle we will determine if the pre LT activity differed from speed matched control and the post LT muscle activity differs from its speed matched control, instead of comparing pre LT to post LT directly. We have chosen this comparison because we believe that it will best account for changes in speed. correlations were conducted to evaluate the relationship between percentage average activation for the muscle in a specific region and the measures of motor performance (work, angles) in that region. Significance for all tests was set at alpha < 0.05. All statistics were run using SPSS version 1 7 .0 (SPSS, Inc.). Results The results of the analyses reveal a ll subjects experienced increase in walking speed post LT (except participants walking at slow speed before and after LT ) ; in add ition to changes in the non paretic leg performance. H owever the amount of change was different f or different groups. Specifically, individuals that increased their speed from moderate to fast revealed the maximum change (0.40m/s), followed by subjects wal king at slower pre speeds who proposed to moder ate post speeds (0.34m/s). The people walking at moderate speeds pre and post LT had an increase of 0.17 m/s.

PAGE 135

135 Participants in the slow group before and after training exhibited a non significant increase of on ly 0.02 m/s. Magnitude of EMG A ctivity in the E ntire Gait Cycle The results of our analysis suggest that the individuals walking at moderate speed post LT had the m ost compensation by the non paretic leg, while the individuals walking at slow speed had the least compensation. Specifically, individuals walking at moderate speed post LT had increased magnitude of activity in most muscles (all except SO for subjects from slow moderate or moderate moderate) while those walking at slow speeds post LT rev ealed increased activity in only VM (Appendix A, Figure A 1) The individuals increasing speed from the moderate pre to fast post LT did not show any change in the magnitude of muscle activity as compared to speed matched control subjects (Appendix A, Figu re A 2). Magnitude EMG A ctivity in D iff erent R egions of the G ait C ycle Slow speed group Individuals walking at slow speed pre LT presented different changes in the muscle magnitude (compared to the appropriate speed matched control group) depending on whe ther they still walked at slow speed after LT or they increased their speed to moderate level (figure 5 4 Appendix A: figure A 3 A 5 ) Continued c ompensation post LT was provided by increased MG in region one and VM in region two ( each are previously defined as novel compensations) and LH in region one (conventional compensation H1 burst ) in subjects walking at slow speed after LT (6 subjects) On the other hand, individuals who subsequently walked at moderate speeds ( 13 subjects figure 5 5 Appendix A: figure A 4 ) after LT revealed significantly more activity in the thigh muscles during the entire stance phase relative to the speed matched

PAGE 136

136 controls than they did relative to the speed matched controls before LT (i.e., in 20 of 20 p ossible muscle region combinations after LT and 10 of 20 before LT). They also increased TA throughout the stance phase after LT when it had been reduced in most of the stance phase before LT. Note that while there was also much increas ed activity in swing phase, the absolute magnitude was much lower in most muscles and so it may not be clinically significant. Moderate speed group. Individuals walking at moderate speed at both times showed fewer differences from the same speed matched control data after LT with respect to the thigh muscles (there were 8 of 20 muscle region combinations increased after LT as opposed to 16 of 20 before LT) Appendix A: figure A 3 to A 7 As to the specific compensations we outlined in chapter 4: A2 in region 4 is present at bot h pre and post LT H1 in regions 6, 1 or 2 is present at both pre and post LT ; the novel MG in region 1 is only present at pre LT; the novel VM activity in regions 2 and 3 is present at both times; and the novel hamstrings activity in regions 3 and 4 is on ly present at pre LT (Appendix A, figure A 7) Furthermore, i ndividuals attaining faster speed post LT (7 subjects) (figure 5 7 Appendix A, figure A 8 ) had significantly greater magnitude of activity in the upper l e g muscles as compared to speed matched control subjects. The compensatory mechanisms involved were: hamstrings in region 1 to increase H1present both before and after LT. Among the novel compensations MG was present in region one, VM in region two and three (figure 5 6) (mid stance) and hamstr ings in region three and four both before and after LT

PAGE 137

137 Percentage D uration of M uscle A ctivity in the E ntire G ait C ycle (Appendix B, table B 1) Slow group. T he individuals walking at slow speed pre LT exhibited normalization of percentage duration of activity post training in most muscles. I ndividuals continuing to walk at slower speed had increased compensation from upper and lower leg muscles ( SO, VM, RF), w hile the group of participants that increased walking speed from slow to moderate speed had a n increase in compensation only from the plantarflexor muscles (MG, SO) Moderate speed group. Participants w alking at moderate speed before LT exhibited increased percentage duration of activity primarily in the power producing muscles. In individuals walking at moderate speed after LT increased compensation was provided primarily by the two plantarflexor muscle s On the other hand the individuals walking at faster speed post LT revealed significantly increased percentage duration of activity in MG, SO, LH and MH. Percentage D uration of M uscle A ctivity in D ifferent R egions of the G ait C ycle Slow speed group. The percentage duration of muscle activity during swing phase post LT was similar to control subjects walking at matched speeds, while differences were present in muscle activity during the stance phase.(figure 5 4) For the specific compensations by the non pa retic leg in individuals walking at slow speed after LT: increased A2 by increased MG and SO was present both before and after LT (only MG) Among the novel compensatory mechanisms of the non paretic leg: increased activity of VM in region t hree (pre LT) o r two (post LT) (mid stance) and increased plantarflexor activity in region one (only MG post LT) was present both pre and post LT,

PAGE 138

138 while increased hamstrings activity in the late stance phase was present only before training (Appendix B, Table B 4) On t he other hand, t he individuals walking at moderate speed after LT (slow pre) had significantly increased muscle activity in several muscles during stance phase (figure 5 5 Appendix B, Table B 5 ) For the specific compensations: A2 was present both pre (on ly SO) and post LT, while increased H1 due to hamstrings activity was present in region one only post LT. Among the novel compensatory mechanisms increased plantarflexor activity was present pre and post LT (only MG post) and so was the increased VM activity in region two and three. However, increased hamstrings activity in region four was present only pre training. Moderate speed group. There was significant normalization of temporal patterning of other muscles in various reg i ons of the gait cycle in individuals continuing to walk at moderate speed after LT (figure 5 6 Appendix B Table B 6 ). The trend of specific compensatory mechanisms in individuals walking at moderate speed post LT was: increased A2 due to increased MG wa s present pre and post LT and increased hamstrings in region 1 was also present both before and after LT (only LH after LT). However, among the novel compensatory mechanisms, all the three were present before LT, while none of them were revealed after LT. On the other hand individuals walking at faster speed after LT had normalized temporal patterning of muscles in region three to six of the gait cycle (figure 5 7 Appendix B Table B 7 ) The specific compensations as outlined in chapter 4: increased hamstr ings activity in region one was [resent both before and after training, while increased activity of the plantarflexors in region four was present only pre LT. with

PAGE 139

139 respect to the novel compensatory mechanisms of the non paretic leg, SO activity was signifi cantly greater than speed matched control subjects both pre and post LT and also increased activity of VM muscle in region two (region three only pre LT) was present before and after training. Additionally, the increased activity of LH was present in regio n three only before LT. Correlational Analysis H1 and LH & MH in region 1 of the gait cycle. There was no significant correlation between the magnitude of hamstring activity and positive work done by the hip extensor in the first region of the gait cycle o f the slow speed group both before and after LT. On the other hand, individuals in the moderate speed group had significant positive correlation between the MH and H1 following LT, (r= 0.515, p= 0.04) (not LH). SO and MG muscle activity and A2 in region 4 of the gait cycle. Although the magnitude of SO and MG in the fourth region of the gait cycle was significantly greater in the slow speed groups before LT, there was no significant correlation between the muscle activity and the positive work done at the ankle in that region of the gait cycle both before and after LT. TA muscle activity and A2 in region 4 of the gait cycle. There was no significant correlation between the magnitude and duration of the TA activity in slow speed group before and after LT. RF muscle activity and peak knee flexion angle in region 4 and 5 of the gait cycle. The percentage duration of the RF activity in the fourth region of the gait cycle was significantly negatively correlated with the peak knee flexion in individuals in the fas t speed group (r= 0.792, p= 0.002). No significant correlations were present after LT.

PAGE 140

140 MH muscle activity and positive hip extensor work. In the slow speed group a significant positive correlation (r= 0.528, p= 0.026) was present before LT, however after LT there was a significant negative correlation between the two (r= 0.654, p=0.015). MG muscle activity and leg angle in first region. There was no significant correlation between the magnitude and duration of the MG activity in slow and moderate speed gr oup before and after LT. Discussion The results of our analysis revealed mixed response to LT post stroke. The mus cle activity of non paretic leg was significantly normalized in some individuals walking post LT. Nevertheless, there were some significant di fferences in the muscle activity in the non paretic leg post LT when compared to healthy control subjects walking at matched speeds. Specifically, participants that walked at slower speeds before LT, but acquired moderate walking speed post LT exhibited th e most diff erences. This suggests that, in addition to the changes that were occurring in the paretic leg, the non paretic leg also appears to have contributed to the i ncreased walking speed through increases in one or more compensatory mechanisms These c ompensatory mechanisms may be just increase in the normally occurring positive phases of work generation in the gait cycle (i.e. LH and MH for H1 in region one, SO and MG for A2 in region four). They could also be novel compensatory mechanisms developed b y the non paretic leg in an attempt to compensate for the poorly coordinated paretic leg (i.e. increased MG and SO in region one, or VM in region two and three or increased activity in hamstrings in region three and four of the gait cycle.

PAGE 141

141 With respect to limitations, it is important to note, that because all of the muscles as well as the kinetic and kinematic measures influence walking and are not independent, we did not correct for multiple comparisons in our study. Furthermore, the amplitude of EMG detec ted using surface electrodes is sensitive to many intrinsic (thickness of subcutaneous fat, distribution of motor units, conduction velocities and detection systems used), and extrinsic factors, i.e. the factors which can be influenced by the experimenter (orientation, location and area and shape of the electrode and distance between them 152 ), which renders it highly variable. T he reliability of the magnitude of the EMG in the dynamic contraction is also influenced by length and force change due to stationary shift in re lative position of the electrode and changes in tissue conductivities 153 Furtherm ore, the changes in the joint angles can cause the recording electrodes to shift relative to active muscle fiber and the change in the muscle fiber direction to alter tissue conductivity. 153 Thus all comparisons based on the magnitude measures must be cautiously evaluated. However, the amplitude of the surface EMG is related to the net motor unit activity, i.e. recruitment and discharge rates of motor units 151 and forms an important component of the EMG analysis. Care was taken while recording and analyzin g EMG activity to minimize errors and caution is warranted with the interpretation from strictly amplitude measures Musc le Activity over the Entire Gait Cycle Increased magnitude of muscle activity over the entire gait cycle was related to the severity of deficits and walking speed attained. Analysis of i ntegrated EMG activity over the gait cycle in individuals with severe speed deficits was associated with significantly greater magnitude in more number of muscles (slow moderate: seven muscles as opposed to five muscles pre LT ) as compared to ind ividuals with moderate deficits

PAGE 142

142 (moderate fast: same four muscles both pre and post LT ). Although integrated EMG activity provides important information regarding the firing or recruitment of muscle fibers durin g the gait cycle, it is important to take under consideration the limitations associated with the processing of EMG amplitude (discussed above). Furthermore, t he tempor al patterning in nearly all individuals ( walking at all the different speeds ) after LT revealed fewer differences from healthy control subjects walking at the matched speeds than were evident before LT These results are contrary to the pre existing evidence 126, 190 which suggests that the temporal patterning of the muscles is unaltered following rehabilitation. This difference could be attributed to the task specific nature of the gait training provided in our study (i.e. BWSTT), as oppose to pedaling aerobic training and Neuro developmental therapy in the ot her studies mentioned However, increased duration of activity was present in some muscles even after LT, suggesting some of the compensatory mechanisms used by the non paretic leg continued to be used To obtain more details about the compensation mechan ism of the non paretic leg after a task specific BWSTT and investigate the changes in its performance relative to healthy control subjects walking at similar speeds, we investigated the changes in muscle activity in different regions of the gait cycle. Mus cle A ctivity in D iff erent Regions of the Gait Cycle Ability to walk faster accompanied by normal kinematic and kinetic patterns is the important aims of the contemporary task specific therapies T herefore it is important to reduce the compensation by the non paretic leg, while targeting to increase the contribution of the paretic leg, as was done in this study. However, some of the compensatory mechanisms present before LT continued after LT, most notably in

PAGE 143

143 people who exhibited increase in speed from slow to moderate Compensation was provided by either increasing the normally existing phases of positive work generation over the gait cycle or increased contribution of the novel compensatory mechani sms developed in the non paretic leg post stroke. The f irst compensatory mechanism wa s the increase d positive work done by hip extensors (hamstrings) in stance phase of the gait cycle in both slow and moderate speed groups before and after LT ( except in individuals who walked at slow speed before and after training) I t may be a technique used to facilitate propulsion by promoting forward acceleration of the trunk in order to attain faster spe eds This is also supported by a significant positive correlati on between the positive work done by the hip extensors H1 and magnitude of hamstring (MH) activity during this phase of the gait cycle in moderate speed group. The second, compensatory mechanism was increases from the ankle plantarflexors in the fourth reg ion of the gait cycle in individuals walking at slower speeds pre and post LT. This suggests that the individuals walking at slower speeds may attain faster speed by increased compensation from the ankle plantarflexors. This can be inferred from the incr eased magnitude (only MG) and duration (except MG in slow group) of ankle plantarflexors (MG and SO) present in the fourth region of the gait cycle before LT in individuals walking at slow and moderate speed as compared to healthy control subjects walking at matched speeds respectively. Some novel mechanisms of compensation continued to exist even after LT. First the additional activity of the hamstrings in the late stance phase or region three and four of the gait cycle in participants that progressed f rom sl ow pre LT moderate post LT )

PAGE 144

144 was present both before and after LT This extra burst of hamstrings (also reported by other researchers 126 69 ) increases extension of the non paretic hip, which thus facilitates the swing of the paretic leg, thereby compensating for the poor activity of the paretic plantarflexors and hip flexor muscles 126 Alternatively, the increased activity of th e hamstring muscles in the late stance phase may hold head, arm and trunk and prevent postural lean 21 in subjects ten ding to walk faster This phenomenon is further supported by a significant positive correlation between the positive hip extensor work and the MH activity during the third region of the gait cycle both before and after the training. Secondly, the increase d activity of VM in the mid stance phase of the non paretic leg has been indicated to facilitate propulsion of the swinging paretic leg 181 and was present in almost all groups before and after LT (specifically individuals progressing to next speed groups) The increased VM firing in the stance phase is a novel non paretic leg adaptation aimed to facilitate increased propulsion in the paretic leg to achieve faster speed Third, all the individuals before and after (except moderate speed participants walking at moderate speed even after LT) LT presented an imp ortant compensatory mechanism by the ankle plantarflexors in the first region of the gait cycle. It is a novel compensation mechanism to produce increased propulsion that might possibly be facilitated by the altered asymmetric kinematics. The increased plantarfl exors activity in this region of the gait cycle may facilitate propulsion if the non paretic leg is not placed ahead of the body center of mass (as happens in a step to gait 160 ). Implications of Analysis. Individuals grouped in the slow speed group revealed significantly increased magnitude and duration of most recorded muscles while they

PAGE 145

145 walked at moderate speed after LT. This s uggests that the increase in walking speed could be due to increased contribution by the compensatory mechanisms of the non paretic leg. On the other hand, individuals walking at moderate speed before LT, but fast speed post LT revealed little difference i n magnitude of muscle activity ( although note limitations of magnitude discussed above), while the temporal pattern of muscle activity was significantly improved. And in the individuals who presented with a speed change without a change in speed group pos t LT (i.e. slow slow and moderate moderate) exhibited significantly reduced contributions of the non paretic leg during walking as compared to healthy control subjects. This suggests that, although changes in speed groups may not have occurred, compensati on by the non paretic leg may have reduced and the patterns of walking may have improve d in these participants This mi ght be associated with increased and/or improved motor activity of the paretic leg. The results of this study suggest that the LT using B WSTT is an effective technique of g ait rehabilitation post stroke Despite the limitations of comparison of EMG amplitude between subjects and between time points in the same subjects in consideration, we believe the increased magnitude of most muscles in individuals walking from slow to moderate speed suggest s additional involvement of compensatory mechanisms We suggest that empha sizing reduction of the compensatory mechanisms of the non paretic leg would force increased use of the paretic leg (in additio n to the conventional emphasis of increasing the paretic leg use). This dual pronged approach to modulate the performance of individuals walking after stroke might produce more effective response to therapy.

PAGE 146

146 DLS SLS SLS DLS SWG SWG Region 1 Region 2 Region 3 Region 4 Region 5 Region 6 Figure 5 1. Represents the different regions of the gait cycle. At the top are the marked the di fferent physiological phases of the gait cycle (DLS= double limb support, SLS= single leg support, SWG= swing phase) and at the bottom are the corresponding regions used in this study. Figure 5 2. Represents the calculation of integrated EMG activity. F irst figure represents the raw EMG signal, second panel represents the rectified and filtered EMG signal and the third figure represents integration rule with equally spaced abscissas (red lines), with respect to time in the gait cycle.

PAGE 147

147 Figure 5 3. Repre sents the steps of k means analysis. First figure represents the raw EMG magnitude, second figure shows the filtered rectified EMG (linear envelope) and the third figure represents the original raw EMG signal with the detected periods of activity indicated in gray shaded areas

PAGE 148

148 Regions 1 2 3 4 5 6 TA MG SO VM RF LH MH GM Regions 1 2 3 4 5 6 TA MG SO VM RF LH MH GM Figure 5 4. Represents difference in muscle activity in individuals walking at slow speed (0 < 0.4m/s) before and after locomotor training as compared to healthy control subjects walking at matched speeds. Top panel shows the differences in the integrated EMG activity pre (left) and post training (right) when compared to the speed matched controls Bottom panel shows the percentage duration of activity pre (left) and post training (right) Red squares represent increased activity in the non paretic leg and green represent reduced activity in the non paretic leg as compared to controls. More details in appendix A and B Regions 1 2 3 4 5 6 TA MG SO VM RF LH MH GM Regions 1 2 3 4 5 6 TA MG SO VM RF LH MH GM

PAGE 149

149 Regions 1 2 3 4 5 6 TA MG SO VM RF LH MH GM Figure 5 5. Represents difference in muscle activity in individuals walking at slow speed (0 <0.4m/s) before and moderate speed (0.4 0.8m/s) after locomotor training as compared to healthy control subjects walking at matched speeds. Top panel shows the differences in the integrated EMG activity pre (left) and post training (right) when compared to the speed matched controls Botto m panel shows the percentage duration of activity pre (left) and post training (right). Red squares represent increased activity in the non paretic leg and green represent reduced activity in the non paretic leg as compared to controls. More details in app endix A and B. Regions 1 2 3 4 5 6 TA MG SO VM RF LH MH GM Regions 1 2 3 4 5 6 TA MG SO VM RF LH MH GM Regions 1 2 3 4 5 6 TA MG SO VM RF LH MH GM

PAGE 150

150 Regions 1 2 3 4 5 6 TA MG SO VM RF LH MH GM Regions 1 2 3 4 5 6 TA MG SO VM RF LH MH GM Figure 5 6 Represents difference in muscle activity in individuals walking at moderate speed (0.4 0.8m/s) before and after locomotor training as compared to healthy control subjects walking at matched speeds. Top panel shows the differences in the integrated EMG activity pre (left) and post training (right) when compared to the speed matched controls Bottom panel shows the percentage duration of activity pre (left) and post training (right). Red squares represent increased activity in the non paretic leg and green represent reduced activity in the non paretic leg as compared to controls. More details in appendix A and B. Regions 1 2 3 4 5 6 TA MG SO VM RF LH MH GM Regions 1 2 3 4 5 6 TA MG SO VM RF LH MH GM

PAGE 151

151 Regions 1 2 3 4 5 6 TA MG SO VM RF LH MH GM Regions 1 2 3 4 5 6 TA MG SO VM RF LH MH GM Figure 5 7 Represents difference in muscle activity in individuals walking at moderate speed (0.4 0.8m/s) before and fast speed (> 0.8m/s) after LT as compared to healthy control subjects walking at matched speeds. Top panel shows the differences in the integrated EMG activity pre (left) and post training (right) when compared to the speed matched controls Bottom panel shows the percentage duration of activity pre (left) and post training (right). Red squares represent increased activity in the non paretic leg and green represent reduced activity in the non paretic leg as compared to controls. More details in appendix A and B. Regions 1 2 3 4 5 6 TA MG SO VM RF LH MH GM Regions 1 2 3 4 5 6 TA MG SO VM RF LH MH GM

PAGE 152

152 CHAPTER 6 CONCLUSION H emiparetic gait is characterized by asymmetry in motor output of the paretic and the non paretic leg Usually, the non paretic leg is perceived to compensat e for the weakness and discoordination of the paretic leg. Therefore, the non paretic leg has traditionally been considered unaffected, even to the point that it has sometimes been used as a standard to gauge the performance of the paretic leg. Some studies have documented significant contribution of the non paretic leg output to walking performance o f individuals post stroke 12, 13, 59 w hile other studies have highlighted ( in pedaling literature 11, 89 ) that the non paretic leg exerts a strong negative in f luence on the paretic leg performance. We suggest that evaluating the non par etic leg mechanisms should be important in order to identify its unique compensatory patterns. Furthermore, allowing normal performance of the non paretic leg, while trying to promote proper paretic leg movement is not likely optimal for restoring the pre stroke biomechanical pattern of walking. While asymmetry in limb loading and unloading has been well documented in sta tic and quasi static tasks 9 7, 8 there is limited information available related to limb loading and unloading during steady state walking. Therefore, in the first study of this dissertation, we quantified magnitude of limb loading and unloading during the two double limb suppo rt periods of the gait cycle. We also evaluated its timing, pattern and relation to medial later al leg angle and knee angle. Our results showed that the limb loading and unloading was asymmetric between the paretic and the non paretic leg, with the non paretic leg sustaining greater load and for longer duration (as revealed by cross over time ( the ti me to switch the predominance of weight from the leg in second

PAGE 153

153 double support to the leg in first double support ). Our analysis revealed that the loading and unloading asymmetry related significantly with the medial lateral leg angle and knee angle indicating the biomechanical compensations employed to maintain the steady state walking despite asymmetry in loading an d unloading. Our findings also showed the variability in the pattern of loading and unloading which provides greater insight into biomec hanical adjustments made by individuals post stroke. A symmetry in limb loading and unloading results in structural 55, 56 and biomechanical changes 57, 58 in the non paretic leg thus making it of importance to address clinically I n the second study of this dissertation we exclusively focused on the biomechanical compensation and impairment mechanisms of the no n paretic leg. The n on paretic leg is often reported to be compensating for the poorly coordinated paretic leg, without any detail about the mechanisms of these compensatory patterns. Understanding of the compensatory mechanisms provides greater insight in to motor control of hemiparetic gait (spe cifically the non paretic leg) and t herefore might guide in developing more effective rehabilitation techniques. The results of our study show significant difference in muscle activity between the non paretic leg an d the healthy control subjects walking at the matched speeds, associated with significant correlations with simultaneous measures of motor performance. Specifically, some of these compensations are simply increased output of the important sources of work d one in the normal gait as indicated by significant positive correlations between the muscle activity and work done at a joint. For example, A2 was associated with SO and MG in the fourth region of the gait cycle or and H1 was associated with MH and LH were related in the first region of the gait cycle. O ther compensations however were novel adaptations of

PAGE 154

154 the non paretic leg to compensate for the paretic leg and achieve steady walking state. For example, hamstrings activity in late stance to generate prop ulsion, early stance plantarflexor activity that generates propulsion because the non paretic leg is not placed as far in front of the body, and mid to late stance activity in the knee extensors that facilitates paretic leg swing. Information obtained fr om this study emphasizes that stroke is not a unilateral motor control problem as often suggested, and there is a bilateral involvement. This indicates that gait related rehabilitation post stroke should focus not just on the paretic leg but also on the no n paretic leg. Thus, in the third study of this dissertation we evaluated the response of the non paretic leg to a task specific locomotor training, i.e. body weight supported treadmill training. Body weight supported training provides a permissive enviro nment in that it affords individuals a walking experience that more closely approximates the actual sensorimotor pattern of walking when compared with walking overground 97 Furthermore, manual assistance is provided by trainers on both the non paretic and the paretic leg and the trunk to maintain kinematically appropriate pattern of the non paretic leg. The results of our study reveal that body weight supported treadmill tr aining facilitates increased walking speed with associated changes in the non paretic leg muscle activity. For example individuals with severe gait deficits acquire greater speeds by increased compensation by the non paretic leg (both conventional and nove l adaptation mechanisms), while individuals who did not increase to moderate (from slow) or fast (from moderate) walking group reveal reduced compensation by the non paretic

PAGE 155

155 leg. This improved non paretic leg muscle activity is also associated with better performance of the paretic leg muscles Therefore we can state that improved walking performance in the individual post stroke appears to have contribution from both the paretic and the non paretic leg. Thus, in order to maximize the benefits of the body w eight supported treadmill training, it is essential to train both the paretic and the non paretic legs. We suggest that focusing on both the non paretic leg and the paretic leg provides a two pronged approach to the rehabilitation that is better than the c onventional unilateral approach focused on the paretic leg It can be supported firstly by greater increases in walking speed in our study participants as oppose d to other studies 111, 119, 191 evaluating the benefits of the body weight supported treadmill training ( we had a trainer for both the paretic and the non paretic leg, while most studies in literature only had a trainer only on the paretic leg) Secondly, including the non paretic leg in rehabilitation results in a normalized sensorimotor state for the non paretic leg, which possibly pro vides facilitatory input to the paretic leg (based on extrapolation of results from pedaling studies 11, 89 ). Therefore, we suggest that it is important to emphasize that allowing altered performance of the non paretic leg, while trying to promote proper paretic leg movement is likely not optimal for developing coordinated walking patterns. Future Directions The aim of this dissertation was to evaluate the non paretic leg performance in hemiparetic walking in relation to loading asymmetries, compensatory mechanisms and responsiveness to locomotor training. Through the three studies performed as a part of this dissertation we were successful in providing detailed and interesting insights about the non paretic leg performance t hus, filling in the previously existing gap in our

PAGE 156

156 knowledge. This however leads us to some more valuable questions related to the motor control of hemip aretic gait, which still need to be answered. The chapters 3 and 4 of this dissertation elucidate the altered performance of the non paretic leg. Extrapolating the results from the pedaling studies in stroke 11, 89 and walking studies post SCI 148 we believe that the performance of the non paretic leg is likely to influence the output of the paretic leg. Therefore, future studies should be aimed at designing experiments to study the performance of each leg independently while biom echanically controlling the performance of the other leg, in a task of walking (or similar to walking). These studies should also focus to investigate the influence of performance of one leg on the performance of the other (i.e. paretic on non paretic and non paretic on paretic). In the fifth chapter of this dissertation we established the benefits of including the non paretic leg in the rehabilitation protocol (as compared to other studies which did not include the non paretic leg). Future studies could e valuate the effect of locomotor training in two groups of individuals, where one group is provided guidance on both the legs, while the other groups receives guidance on only the paretic leg. Also it would be interesting to investigate the effect of locomo tor training with augmentative therapy (e.g. functional electrical stimulation) on the paretic leg and only manual guidance on the non paretic leg compared to only manual guidance on both the legs. In summary, it is important to train both the legs and max imize the output of the paretic leg, while trying to minimize the compensation by the non paretic leg. Summary The findings of this dissertation further illuminate the asymmetrical nature of the hemiparetic gait. Asymmetry implies altered output of both th e paretic and the non

PAGE 157

157 paretic leg. Altered output of the non paretic leg presents as either impairment or compensation. The compensation of the non paretic leg might be increased during the conventional phases of the positive work in the gait cycle, or mig ht be novel adaptive patterns. These compensatory patterns can to some degree be normalized following a task specific rehabilitation technique.

PAGE 158

1 58 APPENDIX A MAGNITUDE OF INTEGRA TED EMG ACTIVITY OVE R THE ENTIRE GAIT CY CLE Figure A 1. Integrated EMG activity in individuals walking at slow speed pre LT. Top plot represents individuals in slow speed group pre LT, middle bar plots represent individuals walking at slow speed post LT and bottom represents individuals progressing to moderate speed post LT. Y axis represents integrated EMG in Vs and X axis represents various muscles recorded. TA=tibialis anterior, MG= medial gastrocnemius, SO= soleus, VM= vastus medialis, LH= lateral hamstrings, MH= medial hamstrings and GM= gluteu s medius.

PAGE 159

159 Figure A 2 Integrated EMG activity in individuals walking at moderate speed pre LT. Top plot represents integrated EMG in individuals in moderate speed group pre LT, middle represent s individuals walking at moderate speed post LT and bottom represents individuals progressing to fast speed post LT. Y axis represents integrated EMG in Vs and X axis represents various muscles recorded. TA=tibialis anterior, MG= medial gastrocnemius, SO= soleus, VM= vastus medialis, LH= lateral hamstring s, MH= medial hamstrings and GM= gluteus medius.

PAGE 160

160 Figure A 3. Represents integrated EMG activity in different regions of the gait cycle in individuals walking at slow speed pre LT. Y axis represents integrated EMG in Vs and X axis represents vari ous muscles recorded. TA=tibialis anterior, MG= medial gastrocnemius, SO= soleus, VM= vastus medialis, LH= lateral hamstrings, MH= medial hamstrings and GM= gluteus medius.

PAGE 161

161 Figure A 4. Represents integrated EMG activity in different regions of the gait cycle in individuals walking at slow speed pre and post LT. Y axis represents integrated EMG in Vs and X axis represents various muscles recorded. TA=tibialis anterior, MG= m edial gastrocnemius, SO= soleus, VM= vastus medialis, LH= lateral hamstrings, MH= medial hamstrings and GM= gluteus medius.

PAGE 162

162 Figure A 5. Represents integrated EMG activity in different regions of the gait cycle in individuals walking at slow speed pre and moderate speed post LT. Y axis represents integrated EMG in Vs and X axis represents various muscles recorded. TA=tibialis anterior, MG= medial gastrocnemius, SO= soleus, VM= vastus medialis, LH= lateral hamstrings, MH= medial hamstrings and GM= gluteus medius.

PAGE 163

163 Figure A 6. Represents integrated EMG activity in different regions of the gait cycle in individual s walking at moderate speed pre LT. Y axis represents integrated EMG in Vs and X axis represents various muscles recorded. TA=tibialis anterior, MG= medial gastrocnemius, SO= soleus, VM= vastus medialis, LH= lateral hamstrings, MH= medial hamstrings and GM= gluteus medius.

PAGE 164

164 Figure A 7 Represents integrated EMG activity in different regions of the gait cycle in individuals walking at moderate speed pre and post LT. Y axis represents integrated EMG in Vs and X axis represents various muscles r ecorded. TA=tibialis anterior, MG= medial gastrocnemius, SO= soleus, VM= vastus medialis, LH= lateral hamstrings, MH= medial hamstrings and GM= gluteus medius.

PAGE 165

165 Figure A 8 Represents integrated EMG activity in different regions of the gait cycle in individuals walking at moderate speed pre LT and at fast speed post LT. Y axis represents integrated EMG in Vs and X axis represents various muscles recorded. TA=tibialis ante rior, MG= medial gastrocnemius, SO= soleus, VM= vastus medialis, LH= lateral hamstrings, MH= medial hamstrings and GM= gluteus medius.

PAGE 166

166 APPENDIX B TABLES FOR PERCENTAG E DURATION OF ACTIVI TY BEFORE AND AFTER LOCOMOTOR TRAINING Table B 1. Represents the perce ntage duration of activity pre and post LT over the entire gait cycle in individuals walking at different speeds. All values are represented as percentages. Pre LT slow Post LT slow Control slow slow moderat e post LT Control moderat e Pre LT moderat e Pos t LT moderat e moderate fast post LT controls fast TA 39.88 12.3 35.34 3.3 53.27 12.6 50.26 12 55.70 9.76 50.47 8.6 52.05 4.9 51.21 8.4 53.07 10.1 MG 52.31 9.75 51.56 6.8 43.63 6.8 50.63 9.25 44.70 7.8 57.11 8.7 62.15 6.9 55.42 10.7 46.21 5.9 SO 64.23 11.6 65.26 2.65 53.85 5.6 56.89 7.1 53.20 5.8 64.27 7.3 64.965 3.4 58.33 4.08 51.85 3.7 VM 55.89 12.4 64.48 6.9 45.69 8.8 49.96 11.7 43.96 8.3 52.19 7.5 53.847 17.2 47.24 7.8 41.13 9.9 RF 58.44 11.6 64.38 3.7 48.60 11.7 55.45 11.8 47.71 10.4 54.86 15.6 54.195 13.2 57.51 8.2 51.69 9.2 LH 56.77 12.8 56.58 12.01 45.60 12.4 52.32 12.62 46.28 10.9 58.36 10.2 49.028 13.6 58.83 3.9 45.00 10.6 MH 62.07 14.2 53.55 9.9 49.78 11.9 58.21 9.4 49.51 11.7 62.98 9.7 47.36 10.4 60.82 6.6 48.18 11.9 GM 61.04 13.7 54.13 8.1 54.87 6.6 59.74 9 54.64 5.4 60.59 11.4 65.00 3 57.14 8.9 54.07 6.6 Table B 2. Represents percentage duration of activity of each muscle in different regions of the gait cycle in individuals walking at slow speed (0 <0.4m/s) pre LT. All values are in percentages. Column 1 represents all the muscles. Column1 Region 1 Region 2 Region 3 Region 4 Region 5 Region 6 TA 41.5116.7 38.9120.5 25.2420.3 35.9913.5 84.3217.2 53.2221.8 MG 64.7211.9 68.0115.4 55.6019.8 43.7217.3 11.0521.6 24.6421.8 SO 76.6213.1 77.0315.5 71.5517.3 59.2713.3 13.7824 27.9524.4 VM 79.7815.1 70.5718.5 61.4824 27.4721.5 32.4130.7 61.9329.4 RF 72.3117.2 70.2918.9 63.1018.8 41.3616.5 48.7234.1 58.8628.2 LH 75.0116.66 68.1923.5 44.1030.2 39.7318.2 40.7032.0 69.8524.7 MH 82.4516.5 58.7419.6 34.2624.2 48.7921.4 69.6127.2 83.2416.8 GM 84.0716.2 85.2615.2 72.7421 31.9318.4 32.9432.8 37.5129.2

PAGE 167

167 Table B 3. Represents percentage duration of activity of each muscle in different regions of the gait cycle in individuals walking at moderate speed (0.4 0.8m/s) pre LT. All values are in percentages. Column 1 represents all the muscles. Column1 Region 1 Region 2 Region 3 Region 4 Region 5 Region 6 TA 56.4417.2 40.4219.6 27.8117.1 40.2515.3 89.6412 67.5920.4 MG 69.1818 79.5317.6 70.04725.3 46.1519.7 21.2616.9 35.7929.7 SO 80.109.1 89.1812.8 87.0110.6 58.8112.5 15.919.4 19.9314.9 VM 89.978.85 69.9520 47.8924.9 17.7613.1 11.6013.2 56.1917.4 RF 77.9323.4 62.8131.5 48.5630.7 45.7913.5 35.6521.6 38.7719.2 LH 83.379.9 62.9126.9 56.9230.5 33.9719.6 27.7322.3 76.6119.7 MH 93.489.8 61.1623.5 40.47326.2 40.0016.1 44.1226.6 90.887.9 GM 93.958.7 85.4823.2 66.5924.8 24.7411.6 40.9223.8 32.2616.3 Table B 4. Represents percentage duration of activity of each muscle in different regions of the gait cycle in individuals walking at slow speed (0 <0.4m/s) post LT. All values are in percentages. Column 1 represents all the muscles. Column1 Region 1 Region 2 Region 3 Region 4 Region 5 egion 6 TA 49.9110.5 39.2615.6 12.326.1 23.8310.4 67.9233.1 60.01 20.2 MG 63.745.6 68.4316.6 45.1322.6 44.499.4 6.6759.3 21.5016 SO 75.039.1 72.244.7 63.6614.1 68.968.2 6.6810.4 17.208.7 VM 89.466.9 84.199.7 85.07511.4 33.4918.4 18.6833.7 60.6017.1 RF 81.988.6 79.5011.3 68.0222.5 50.047.3 28.7422.5 59.9212.6 LH 75.4519.2 71.4523.5 51.7131.1 37.4613.9 40.6830.1 66.8810.8 MH 82.4012.2 52.6023.2 15.3812.4 39.3525.8 66.7431.4 77.4315 GM 89.698 86.1515.2 59.9125.9 14.186.6 18.3114.7 31.6218.5 Table B 5. Represents percentage duration of activity of each muscle in different regions of the gait cycle in individuals walking at slow speed (0 <0.4m/s) pre LT and moderate speed (0.4 0.8 m/s) post LT. All values are in percentages. Column 1 represents all the m uscles. Column1 Region 1 Region 2 Region 3 Region 4 Region 5 Region 6 TA 62.8618.9 51.4223.3 29.9821.9 42.6516.7 76.3325.2 57.7427.3 MG 66.6114.9 62.129.9 62.1617.3 48.2414.45 9.2611.3 20.8123.6 SO 69.141.8 70.2414.6 73.5519 61.2512.6 13.282.4 17.4417.6 VM 84.3910.8 67.5518.3 52.6727.6 11.3310 9.7810.27 59.6524.7 RF 69.4816.3 60.3517.9 54.8526.3 45.5713.8 46.4228 47.4928.1 LH 82.2012.5 51.0631.6 43.2527.4 32.4718.8 21.0016.2 64.826.8 MH 88.047.8 45.8118.7 29.9314.6 47.9318 51.6926.9 78.6219.9 GM 89.9710.5 81.7316.8 60.1718.4 28.7921.7 39.3633 36.0220.5

PAGE 168

168 Table B 6. Represents percentage duration of activity of each muscle in different regions of the gait cycle in individuals walking at moderate speed (0.4 0.8 m/s) pre and post LT. All values are in percentages. Column 1 represents all the muscles. Column1 Region 1 Region 2 Region 3 Region 4 Region 5 Region 6 TA 83.7412.4 58.2927.2 15.3512.9 20.805.7 68.1039 70.0010.8 MG 68.9837.7 80.2316.4 87.8311.7 55.4219.3 24.4922.7 41.189.7 SO 73.4718.8 84.818.7 96.664.7 59.4425.4 20.9220.5 28.3430.8 VM 95.411.9 72.7429 29.3326.4 24.2723.9 26.048.1 59.0226.8 RF 82.9118.2 67.5729.3 39.2118.3 42.7810.6 42.6827.7 34.8117.7 LH 89.4513 65.0242.8 26.4419.5 13.4121.6 3.973.3 80.7810.8 MH 88.4217 36.8918.7 10.718.4 21.523 31.7423.2 93.747.1 GM 97.493.4 93.7110.5 73.1918.2 26.3012 42.4325.5 39.4526.7 Table B 7. Represents percentage duration of activity of each muscle in different regions of the gait cycle in individuals walking at fast speed (>0.8m/s) post LT. All values are in percentages. Column 1 represents all the muscles. Column1 Region 1 Region 2 Region 3 Region 4 Region 5 Region 6 TA 57.6413.8 57.6823.5 44.7822.5 38.7416.9 76.2315.3 62.7715 MG 49.1215 72.2713.9 73.2318.2 38.408.2 17.9421.5 7.578.9 SO 64.6312.4 91.126.9 89.717.75 51.629.4 5.126.1 7.5710.5 VM 82.0315.4 61.4620.4 41.2217 20.3811.3 10.8513.4 42.4919.4 RF 79.8615.7 67.2125.6 47.9224.9 27.0672619 20.5113.7 37.2112.8 LH 68.8820.7 43.6122.4 31.2621.5 28.4315.6 39.0522.1 55.2620.4 MH 72.8819.1 45.6321.7 28.7618.4 32.7620.8 44.3222.4 65.1816.6 GM 80.6611.5 96.944.54 88.728.8 25.6315 22.1516.6 20.4514.7 Table B 8. Represents percentage duration of activity of each muscle in different regions of the gait cycle in healthy control individuals walking at slow speed (0.3m/s). All values are in percentages. Column 1 represents all the muscles. Column1 Region 1 Region 2 Region 3 Region 4 Region 5 Region 6 TA 57.6413.8 57.6823.5 44.7822.5 38.7416.9 76.2315.3 62.7715 MG 49.1215 72.2713.9 73.2318.2 38.408.2 17.9421.5 7.578.9 SO 64.6312.4 91.126.9 89.717.75 51.629.4 5.126.1 7.5710.5 VM 82.0315.4 61.4620.4 41.2217 20.3811.3 10.8513.4 42.4919.4 RF 79.8615.7 67.2125.6 47.9224.9 27.06 19 20.5113.7 37.2112.8 LH 68.8820.7 43.6122.4 31.2621.5 28.4315.6 39.0522.1 55.2620.4 MH 72.8819.1 45.6321.7 28.7618.4 32.7620.8 44.3222.4 65.1816.6 GM 80.6611.5 96.944.54 88.728.8 25.6315 22.1516.6 20.4514.7

PAGE 169

169 Table B 9. Represents percentage duration of activity of each muscle in different regions of the gait cycle in healthy control individuals walking at matched moderate speed (0.6m/s). All values are in percentages. Column 1 represents all the muscles. Column1 Region1 Region 2 Region 3 Region 4 Region 5 Region 6 TA 65.4111.5 49.1122 40.1524.4 41.5211.8 76.9715.9 62.03715.6 MG 44.017.5 75.3918.8 88.4013.5 31.3512.3 22.54 12.6812 SO 64.212.3 91.017.8 94.027.6 47.3610.7 9.9611.6 11.339.48 VM 90.27.1 53.4019.5 23.4216.8 14.9410.4 13.5610.6 54.3519.3 RF 85.0714 57.9526.7 27.6323.4 28.7713.4 27.1111.3 37.9615.7 LH 70.7715.8 42.0919.1 26.4120.9 22.5915.0 42.2024.5 79.8617.3 MH 73.0618.4 46.4919.2 25.1521.6 26.609 44.7623.6 85.8511.3 GM 89.628.3 93.579.7 73.4718.1 19.8812.7 26.01516.6 28.9119.3 Table B 10. Represents percentage duration of activity of each muscle in different regions of the gait cycle in healthy control individuals walking at matched fast speed (0.9m/s). All values are in percentages. Column 1 represents all the muscles Column1 Region 1 Region 2 Region 3 Region 4 Region 5 Region 6 TA 66.9813.5 41.8321.5 29.0322.1 40.3414.9 76.6511.6 63.6714.9 MG 46.3616.6 81.4016 93.417.4 29.0810.5 15.382.2 15.3815.5 SO 62.608.7 91.898.3 95.406.9 43.8912.3 7.097.4 7.915.9 VM 92.664.8 45.0824.3 18.0813.9 13.9010.8 15.4411.5 54.4518.5 RF 89.758.2 57.8120.3 27.4622.8 38.3517.5 41.5520.9 47.1616.5 LH 68.8518 43.119.8 21.9818.6 16.4313 36.2023.7 85.1417.2 MH 73.2317.4 44.0821.4 24.8023.1 23.4122 33.3721.9 89.377.9 GM 92.256.1 91.6710.2 58.4224.6 16.8311.2 30.5418.3 34.7520.7

PAGE 170

170 LIST OF REFERENCES 1. Lloyd Jones D, Adams R, Carnethon M, De Simone G, Ferguson TB, Flegal K, Ford E, Furie K, Go A, Greenlund K, Haase N, Hailpern S, Ho M, Howard V, Kissela B, Kittner S, Lackland D, Lisabeth L, Marelli A, McDermott M, Meigs J, Mozaffarian D, Nichol G, O'Donn ell C, Roger V, Rosamond W, Sacco R, Sorlie P, Stafford R, Steinberger J, Thom T, Wasserthiel Smoller S, Wong N, Wylie Rosett J, Hong Y. Heart disease and stroke statistics -2009 update: A report from the american heart association statistics committee and stroke statistics subcommittee. Circulation 2009;119:480 486 2. Wade DT, Hewer RL. Functional abilities after stroke: Measurement, natural history and prognosis. J Neurol Neurosurg Psychiatry 1987;50:177 182 3. Wade DT, Wood VA, Heller A, Maggs J, Langton Hewer R. Walking after stroke. Measurement and recovery over the first 3 months. Scand J Rehabil Med 1987;19:25 30 4. Gresham G DP, Stason WB. Post stroke rehabilitation guidelines technical report. Agency fo r health care policy and research. Rockville, md. 1995 5. Stein J, Harvey RL, Macko RF, Winstein CJ, Zorowitz RD. Stroke recovery and rehabilitation. 2009 6. Dietz V, Duysens J. Significance of load receptor input during locomotion: A review. Gait Postur e 2000;11:102 110 7. Pai YC, Rogers MW, Hedman LD, Hanke TA. Alterations in weight transfer capabilities in adults with hemiparesis. Phys Ther 1994;74:647 657; discussion 657 649 8. de Haart M, Geurts AC, Dault MC, Nienhuis B, Duysens J. Restoration of weight shifting capacity in patients with postacute stroke: A rehabilitation cohort study. Arch Phys Med Rehabil 2005;86:755 762 9. Eng JJ, Chu KS. Reliability and comparison of weight bearing ability during standing tasks for individuals with chroni c stroke. Arch Phys Med Rehabil 2002;83:1138 1144 10. Goldie PA, Matyas TA, Evans OM, Galea M, Bach TM. Maximum voluntary weight bearing by the affected and unaffected legs in standing following stroke. Clin Biomech (Bristol, Avon) 1996;11:333 342 11 Kautz SA, Patten C, Neptune RR. Does unilateral pedaling activate a rhythmic locomotor pattern in the nonpedaling leg in post stroke hemiparesis? J Neurophysiol 2006;95:3154 3163

PAGE 171

171 12. Kim CM, Eng JJ. Magnitude and pattern of 3d kinematic and kinetic gait profiles in persons with stroke: Relationship to walking speed. Gait Posture 2004;20:140 146 13. Olney SJ, Griffin MP, McBride ID. Temporal, kinematic, and kinetic variables related to gait speed in subjects with hemiplegia: A regression approach. Phys Ther 1994;74:872 885 14. Stroke -1989. Recommendations on stroke prevention, diagnosis, and therapy. Report of the who task force on stroke and other cerebrovascular disorders. Strok e 1989;20:1407 1431 15. Perry J, Garrett M, Gronley JK, Mulroy SJ. Classification of walking handicap in the stroke population. Stroke 1995;26:982 989 16. Cunha IT, Lim PA, Henson H, Monga T, Qureshy H, Protas EJ. Performance based gait tests for acu te stroke patients. Am J Phys Med Rehabil 2002;81:848 856 17. Alexander MP. Stroke rehabilitation outcome. A potential use of predictive variables to establish levels of care. Stroke 1994;25:128 134 18. Thirumala P, Hier DB, Patel P. Motor recovery after stroke: Lessons from functional brain imaging. Neurol Res 2002;24:453 458 19. Perry J. Gait analysis: Normal and pathological function Thorofare, NJ: Slack, Inc.; 1992. 20. Bohannon RW, Andrews AW, S mith MB. Rehabilitation goals of patients with hemiplegia. International Journal of Rehabilitation Research 1988;11:181 183 21. Olney sJ, Richards C. Hemiparetic gait following stroke. Part i: Characteristics. Gait & Posture 1996;4:136 148 22. Chen C L, Chen HC, Tang SF, Wu CY, Cheng PT, Hong WH. Gait performance with compensatory adaptations in stroke patients with different degrees of motor recovery. Am J Phys Med Rehabil 2003;82:925 935 23. Kerrigan DC, Karvosky ME, Riley PO. Spastic paretic stif f legged gait: Joint kinetics. Am J Phys Med Rehabil 2001;80:244 249 24. Knutsson E, Richards C. Different types of disturbed motor control in gait of hemiparetic patients. Brain 1979;102:405 430

PAGE 172

172 25. Burdett RG, Borello France D, Blatchly C, Potter C. Gait comparison of subjects with hemiplegia walking unbraced, with ankle foot orthosis, and with air stirrup brace. Phys Ther 1988;68:1197 1203 26. Kuan TS, Tsou JY, Su FC. Hemiplegic gait of stroke pa tients: The effect of using a cane. Arch Phys Med Rehabil 1999;80:777 784 27. Lehmann JF, Condon SM, Price R, deLateur BJ. Gait abnormalities in hemiplegia: Their correction by ankle foot orthoses. Arch Phys Med Rehabil 1987;68:763 771 28. Chen G, Pa tten C, Kothari DH, Zajac FE. Gait differences between individuals with post stroke hemiparesis and non disabled controls at matched speeds. Gait Posture 2005;22:51 56 29. De Quervain IA, Simon SR, Leurgans S, Pease WS, McAllister D. Gait pattern in the early recovery period after stroke. J Bone Joint Surg Am 1996;78:1506 1514 30. 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 2003;18:114 125 31. Olney SJ, Richards C. Hemiparetic gait following stroke. Part i: Characteristics. Gait Posture 1996. 32. Winter DA. The biomechanical and motor control of human gait: Normal, elderly and pathological. Waterloo Biomechanics; 1991. 33. Hamill J KK. Biomechanical basis of human movement. Media, PA: Limppincott Wiliams & Wilkins; 1995. 34. Wong AM, Pei YC, Hong WH, Chung CY, Lau YC, Chen CP. Foot contact pattern analysis in hemiplegic stroke patients: An implication for neurologic status determination. Arch Phys Med Rehabil 2004;85:1625 1630 35. Carlsoo S, Dahlof AG, Holm J. Kinetic analysis of the gait in patients with hemiparesis and in patients with intermittent claudication. Scand J Rehabil Med 1974;6:166 179 36. Sackley CM. Falls, sway, and symmetry of weight bearing after stroke. Int Disabil Stud 1991;13:1 4 37. Laufer Y, Dickstein R, Resnik S, Marcovitz E. Weight bearing shifts of hemiparetic and healthy adults upon stepping on stairs of various heights. Clin Rehabil 2000;14:125 129

PAGE 173

173 38. Rode G, Tiliket C, Boisson D. Predominance of postural imbalance in left hemiparetic patients. Scand J Rehabil Med 1997;29:11 16 39. Shumway Cook A, Anson D, Haller S. Postural sway biofeedback: Its e ffect on reestablishing stance stability in hemiplegic patients. Arch Phys Med Rehabil 1988;69:395 400 40. Winstein CJ, Gardner ER, McNeal DR, Barto PS, Nicholson DE. Standing balance training: Effect on balance and locomotion in hemiparetic adults. Arc h Phys Med Rehabil 1989;70:755 762 41. Mizrahi J, Solzi P, Ring H, Nisell R. Postural stability in stroke patients: Vectorial expression of asymmetry, sway activity and relative sequence of reactive forces. Med Biol Eng Comput 1989;27:181 190 42. Bru nnstrom S. Walking preparation for adult patients with hemiplegia. Phys Ther 1965;45:17 29 43. Dettmann MA, Linder MT, Sepic SB. Relationships among walking performance, postural stability, and functional assessments of the hemiplegic patient. Am J Phys Med 1987;66:77 90 44. Goldie PA, Matyas TA, Spencer KI, McGinley RB. Postural control in standing following stroke: Test retest reliability of some quantitative clinical tests. Phys Ther 1990;70:234 243 45. Turnbull GI, Charteris J, Wall JC. Deficiencies in standing weight shifts by ambulant hemiplegic subjects. Arch Phys Med Rehabil 1996;77:356 362 46. Bohannon RW, Larkin PA. Lower extremity weight bearing under various standing conditions in independentl y ambulatory patients with hemiparesis. Phys Ther 1985;65:1323 1325 47. Bogarth E, Richards, C. Gait analysis and re learning of gait control in hemiplegic patients. Physiotherapy Canada 1981;33:223 230 48. Wall JC, Turnbull GI. Gait asymmetries in residual hemiplegia. Arch Phys Med Rehabil 1986;67:550 553 49. Wannstedt GT, Herman RM. Use of augmented sensory feedback to achieve symmetrical standing. Phys Ther 1978;58:553 559 50. Barbeau H, Rossignol S. Recovery of locomotion after chronic spinalization in the adult cat. Brain Res 1987;412:84 95

PAGE 174

174 51. Grillner S, Dubuc R. Control of locomotion in vertebrates: Spinal and supraspinal mechanisms. Adv Neurol 1988;47:425 453 52. Harkema SJ, Hurley SL, Pat el UK, Requejo PS, Dobkin BH, Edgerton VR. Human lumbosacral spinal cord interprets loading during stepping. J Neurophysiol 1997;77:797 811 53. Dobkin BH, Harkema S, Requejo P, Edgerton VR. Modulation of locomotor like emg activity in subjects with comp lete and incomplete spinal cord injury. J Neurol Rehabil 1995;9:183 190 54. Dietz V, Colombo G, Jensen L, Baumgartner L. Locomotor capacity of spinal cord in paraplegic patients. Ann Neurol 1995;37:574 582 55. Burke MJ, Roman V, Wright V. Bone and jo int changes in lower limb amputees. Ann Rheum Dis 1978;37:252 254 56. Kulkarni J, Adams J, Thomas E, Silman A. Association between amputation, arthritis and osteopenia in british male war veterans with major lower limb amputations. Clin Rehabil 1998;12 :348 353 57. Robbins S, Waked E, Krouglicof N. Vertical impact increase in middle age may explain idiopathic weight bearing joint osteoarthritis. Arch Phys Med Rehabil 2001;82:1673 1677 58. Harrington IJ. Knee joint forces in normal and pathological g ait. In: Niwa S, Perren, S.M., Hatton, T. eds, ed. Biomechanics in orthopaedics Tokyo: Springer Verlag; 1992:121 146. 59. Parvataneni K, Olney SJ, Brouwer B. Changes in muscle group work associated with changes in gait speed of persons with stroke. Clin Biomech (Bristol, Avon) 2007;22:813 820 60. Olney SJ, Griffin MP, Monga TN, McBride ID. Work and power in gait of stroke patients. Arch Phys Med Rehabil 1991;72:309 314 61. Lamontagne A, Malouin F, Richards CL. Contribution of passive stiffness to ankle plantarflexor moment during gait after stroke. Arch Phys Med Rehabil 2000;81:351 358 62. Lamontagne A, Stephenson JL, Fung J. Physiological evaluation of gait disturbances post stroke. Clin Neurophysiol 2007;118:717 729 63. Robertson Dea. Research methods in biomechanics. Champaign, IL: Human Kinetics Publ.; 2004.

PAGE 175

175 64. Den Otter AR, Geurts AC, Mulder T, Duysens J. Gait recovery is not associated with changes in the tempor al patterning of muscle activity during treadmill walking in patients with post stroke hemiparesis. Clin Neurophysiol 2006;117:4 15 65. Bourbonnais D, Vanden Noven S. Weakness in patients with hemiparesis. Am J Occup Ther 1989;43:313 319 66. Burridge JH, Wood DE, Taylor PN, McLellan DL. Indices to describe different muscle activation patterns, identified during treadmill walking, in people with spastic drop foot. Med Eng Phys 2001;23:427 434 67. Lamontagne A, Malouin F, Richards CL. Locomotor speci fic measure of spasticity of plantarflexor muscles after stroke. Arch Phys Med Rehabil 2001;82:1696 1704 68. Hirschberg GG, Nathanson M. Electromyographic recording of muscular activity in normal and spastic gaits. Arch Phys Med Rehabil 1952;33:217 225 69. Shiavi R, Bugle HJ, Limbird T. Electromyographic gait assessment, part 2: Preliminary assessment of hemiparetic synergy patterns. J Rehabil Res Dev 1987;24:24 30 70. Wortis SB, Marks M, Hirschberg GG, Nathanson M. Gait analysis in hemiplegia. Tra ns Am Neurol Assoc 1951;56:181 183 71. Swinnen SH, H., Massion J., Casaer, P. Interlimb coordination neural, dynamical and cognitive constraints. San Diego, CA: Academic Press Inc.; 1994. 72. Rossignol SPS, M.C., Perreault, T. Drew K, Perason KG, Belanger M. Intralimb and interlimb coordination in the cat during real and fictive rhythmic motor programs. Semin Neurosci 1993;5 73. Giuliani CA, Smith JL. Stepping behaviors in chronic spinal cats with one hindlimb deafferented. Journal of Neuroscience 1987;7:2537 2546 74. Grillner S, Rossignol S. On the initiation of the swing phase of locomotion in chronic spinal cats. Brain Res 1978;146:269 277 75. Duysens J, Pearson KG. Inhibition of flexor burst generation by loading ankle extensor muscles in walking cats. Brain Res 1980;187:321 332 76. Desrosiers J, Bourbonnais D, Bravo G, Roy PM, Guay M. Performance of the 'unaffected' u pper extremity of elderly stroke patients. Stroke 1996;27:1564 1570

PAGE 176

176 77. Spaulding SJ, McPherson JJ, Strachota E, Kuphal M, Ramponi M. Jebsen hand function test: Performance of the uninvolved hand in hemiplegia and of right handed, right and left hemiple gic persons. Arch Phys Med Rehabil 1988;69:419 422 78. Sunderland A, Bowers MP, Sluman SM, Wilcock DJ, Ardron ME. Impaired dexterity of the ipsilateral hand after stroke and the relationship to cognitive deficit. Stroke 1999;30:949 955 79. Colebatch JG, Gandevia SC. The distribution of muscular weakness in upper motor neuron lesions affecting the arm. Brain 1989;112 ( Pt 3):749 763 80. Smutok MA, Grafman J, Salazar AM, Sweeney JK, Jonas BS, DiRocco PJ. Effects of unilateral brain damage on contrala teral and ipsilateral upper extremity function in hemiplegia. Phys Ther 1989;69:195 203 81. Watkins MP, Harris BA, Kozlowski BA. Isokinetic testing in patients with hemiparesis. A pilot study. Phys Ther 1984;64:184 189 82. Jones RD, Donaldson IM, Par kin PJ. Impairment and recovery of ipsilateral sensory motor function following unilateral cerebral infarction. Brain 1989;112 ( Pt 1):113 132 83. Nathan PW, Smith MC, Deacon P. The corticospinal tracts in man. Course and location of fibres at different segmental levels. Brain 1990;113 ( Pt 2):303 324 84. HGJM K. Anatomy of descending pathways. In: Brrokheart JG MV, eds, ed. Handbook of physiology Bethesda, Md: American Physiology Society; 1981:597 660. 85. Dietz V. Human neuronal control of automatic functional movements: Interaction between central programs and afferent input. Physiol Rev 1992;72:33 69 86. Swinnen SP, Duysens J. Neuo behavioral determinants of interlimb coordination: A multidisciplinary approach. 2004 87. Dietz V. Do human bipeds use quadrupedal coordination? Trends Neurosci 2002;25:462 467 88. Dietz V. Proprioception and locomotor disorders. Nat Rev Neurosci 2002;3:781 790 89. Kautz SA, Patten C. Interlimb influences on paretic leg function in poststroke hemiparesis. J Neurophysiol 2005;93:2460 2473

PAGE 177

177 90. Kim CM, Eng JJ. The relationship of lower extremity muscle torque to locomotor performance in people with stroke. Phys Ther 2003;83:49 57 91. Suzuki K, Nakamura R, Yamada Y, Handa T. Determinants of maximum walking speed in hemiparetic stroke patients. Tohoku J Exp Med 1990;162:337 344 92. Shumway Cook A, & Woollacot, M. Motor control: Theory and practical applications. P hiladelphia: Lippincott Williams and Wilkins; 2001. 93. Gans BM. Medicine and rehabilitation. Principles and practice. 2005 94. Bobath K. A neurophysiological basis for the treatment of cerebral palsy. Philadelphi a: Lippincott 1980. 95. Sawner KA, & Lavi nge, J.M. Brunnstrom's movement therapy in hemiplegia: A neurophysiological approach Philadelphia: Lippincott 1992. 96. Knott M VD. Proprioceptive neuromuscular facilitation patterns and techniques New York: Harper and Row publishers; 1968. 97. Behrman AL, Harkema SJ. Physical rehabilitation as an agent for recovery after spinal cord injury. Phys Med Rehabil Clin N Am 2007;18:183 202. 98. Hallett M. Plasticity of the human motor cortex and recovery from stroke. Brain Res Brain Res Rev 2001;36: 169 174 99. Kleim JA, Jones TA. Principles of experience dependent neural plasticity: Implications for rehabilitation after brain damage. J Speech Lang Hear Res 2008;51:S225 239 100. Plautz EJ, Milliken GW, Nudo RJ. Effects of repetitive motor trainin g on movement representations in adult squirrel monkeys: Role of use versus learning. Neurobiol Learn Mem 2000;74:27 55 101. Forssberg H, Grillner S, Halbertsma J, Rossignol S. The locomotion of the low spinal cat. Ii. Interlimb coordination. Acta Physiol Scand 1980;108:283 295 102. Forssberg H, Grillner S, Halbertsma J. The locomotion of the low spinal cat. I. Coordination within a hindlimb. Acta Physiol Scand 1980;108:269 281 103. Grillner S, McClellan A, Perret C. Entrainment of the spinal pattern generators for swimming by mechano sensitive elements in the lamprey spinal cord in vitro. Brain Res 1981;217:380 386 104. Bregman BS, Goldberger ME. Infant lesion effect: Ii. Sparing and recovery of function after spinal cord damage in newborn and adult cats. Brain Res 1983;285:119 135

PAGE 178

178 105. Robinson GA, Goldberger ME. The development and recovery of motor function in spinal cats. Ii. Pharmacological enhancement of recovery. Exp Brain Res 1986;62:387 400 106. Hodgson JA, Roy RR, de Leon R, Dobkin B, Edgerton VR. Can the mammalian lumbar spinal cord learn a motor task? Med Sci Sports Exerc 1994;26:1491 1497 107. Lovely RG, Gregor RJ, Roy RR, Edgerton VR. Weight bearing hindlimb stepping in treadmill exercised adult spinal cats. Brain Res 1990;514:206 218 108. Behrman AL, Harkema SJ. Locomotor training after human spinal cord injury: A series of case studies. Phys Ther 2000;80:688 700 109. Dietz V, Wirz M, Jensen L. Locomotion in patients with spinal cord injuries. Phys Ther 1997;77:508 516 110. Basso DM. Neuroanatomical substrates of functional recovery after experimental spinal cord injury: Implications of basic science research for human spinal cord injury. Phys Ther 2000;80:808 817 111. Visintin M, Barbeau H, Korner B itensky N, Mayo NE. A new approach to retrain gait in stroke patients through body weight support and treadmill stimulation. Stroke 1998;29:1122 1128 112. Hesse S, Bertelt C, Jahnke MT, Schaffrin A, Baake P, Malezic M, Mauritz KH. Treadmill training wit h partial body weight support compared with physiotherapy in nonambulatory hemiparetic patients. Stroke 1995;26:976 981 113. Miyai I, Fujimoto Y, Ueda Y, Yamamoto H, Nozaki S, Saito T, Kang J. Treadmill training with body weight support: Its effect on p arkinson's disease. Arch Phys Med Rehabil 2000;81:849 85 2 114. Schindl MR, Forstner C, Kern H, Hesse S. Treadmill training with partial body weight support in nonambulatory patients with cerebral palsy. Arch Phys Med Rehabil 2000;81:301 306 115. Ulrich DA, Ulrich BD, Angulo Kinzler RM, Yun J. Treadmill training of infants with down syndrome: Evidence based developmental outcomes. Pediatrics 2001;108:E84 116. Trueblood PR. Partial body weight treadmill training in persons with chronic stroke. Ne uroRehabilitation 2001;16:141 153

PAGE 179

179 117. Danielsson A, Sunnerhagen KS. Oxygen consumption during treadmill walking with and without body weight support in patients with hemiparesis after stroke and in healthy subjects. Arch Phys Med Rehabil 2000;81:953 9 57 118. Hesse S, Konrad M, Uhlenbrock D. Treadmill walking with partial body weight support versus floor walking in hemiparetic subjects. Arch Phys Med Rehabil 1999;80:421 427 119. Kosak MC, Reding MJ. Comparison of partial body weight supported tread mill gait training versus aggressive bracing assisted walking post stroke. Neurorehabil Neural Repair 2000;14:13 19 120. Pohl M, Mehrholz J, Ritschel C, Ruckriem S. Speed dependent treadmill training in ambulatory hemiparetic stroke patients: A randomized controlled trial. Stroke 2002;33:553 558 121. Hesse S, Bertelt C, Schaffrin A, Malezic M, Mauritz KH. Restoration of g ait in nonambulatory hemiparetic patients by treadmill training with partial body weight support. Arch Phys Med Rehabil 1994;75:1087 1093 122. Ada L, Dean CM, Hall JM, 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 2003;84:1486 1491 123. Barbeau H, Visintin M. Optimal outcomes obtained with body weight support combined with treadmill training in stroke su bjects. Arch Phys Med Rehabil 2003;84:1458 1465 124. Knutsson E. Gait control in hemiparesis. Scand J Rehabil Med 1981;13:101 108 125. Den Otter AR, Geurts AC, Mulder T, Duysens J. Abnormalities in the temporal patterning of lower extremity muscle ac tivity in hemiparetic gait. Gait Posture 2007;25:342 352 126. Buurke JH, Nene AV, Kwakkel G, Erren Wolters V, Ijzerman MJ, Hermens HJ. Recovery of gait after stroke: What changes? Neurorehabil Neural Repair 2008;22:676 683 127. Brodal A. Self observa tions and neuro anatomical considerations after a stroke. Brain 1973;96:675 694 128. Carey JR, Baxter TL, Di Fabio RP. Tracking control in the nonparetic hand of subjects with stroke. Arch Phys Med Rehabil 1998;79:435 441

PAGE 180

180 129. Haaland KY, Harrington DL. Hemispheric control of the initial and corrective components of aiming movements. Neuropsychologia 1989;27:961 969 130. Winstein CJ, Pohl PS. Effects of unilateral brain damage on the control of goal directed hand movements. E xp Brain Res 1995;105:163 174 131. Fisk JD, Goodale MA. The effects of unilateral brain damage on visually guided reaching: Hemispheric differences in the nature of the deficit. Exp Brain Res 1988;72:425 435 132. Cramer SC. Stroke recovery. Lessons f rom functional mr imaging and other methods of human brain mapping. Phys Med Rehabil Clin N Am 1999;10:875 886, ix 133. Nilsson J, Thorstensson A. Ground reaction forces at different speeds of human walking and running. Acta Physiol Scand 1989;136:217 227 134. Cheng PT, Wu SH, Liaw MY, Wong AM, Tang FT. Symmetrical body weight distribution training in stroke patients and its effect on fall prevention. Arch Phys Med Rehabil 2001;82:1650 1654 135. Mercer VS, Freburger JK, Chang SH, Purse r JL. Measurement of paretic lower extremity loading and weight transfer after stroke. Phys Ther 2009;89:653 664 136. Lomaglio MJ, Eng JJ. Muscle strength and weight bearing symmetry relate to sit to stand performance in individuals with stroke. Gait Po sture 2005;22:126 131 137. Li L, Hamill J. Characteristics of the vertical ground reaction force component prior to gait transition. Res Q Exerc Sport 2002;73:229 237 138. Monger C, Carr JH, Fowler V. Evaluation of a home based exercise and training programme to improve sit to stand in patients with chronic stroke. Clin Rehabil 2002;16:361 367 139. Laufer Y, Sivan D, Schwarzmann R, Sprecher E. Standing balance and functional recovery of patients with right and left hemiparesis in the early stages o f rehabilitation. Neurorehabil Neural Repair 2003;17:207 213 140. Hesse SA, Jahnke MT, Bertelt CM, Schreiner C, Lucke D, Mauritz KH. Gait outcome in ambulatory hemiparetic patients after a 4 week comprehensive rehabilitation program and prognostic facto rs. Stroke 1994;25:1999 2004 141. MT Jahnke SH, C Schreiner, K H Mauritz. Dependences of ground reaction force parameters on habitual walking speed in hemiparetic subjects. Gait & Posture 1995;3:3 12

PAGE 181

181 142. Pohl M, Rockstroh G, Ruckriem S, Mrass G, Mehrholz J. Immediate effects of speed dependent treadmill training on gait parameters in early parkinson's disease. Arch Phys Med Rehabil 2003;84:1760 1766 143. Balasubramanian CK, Neptune RR, Kautz SA. Foot pl acement in a body referenc frame during walking and its relationship to hemiparetic walking performance. Clin Biomech (Bristol, Avon) 2010;25 144. Staude G, Wolf W. Objective motor response onset detection in surface myoelectric signals. Med Eng Phys 1999;21:449 467 145. Winter DA. Pathologic gait diagnosis with computer averaged electromyographic profiles. Arch Phys Med Rehabil 1984;65:393 398 146. Bonato P, D'Alessio T, Knaflitz M. A statistical method for the measurement of muscle activation in tervals from surface myoelectric signal during gait. IEEE Trans Biomed Eng 1998;45:287 299 147. Mulroy SJ, Klassen T, Gronley JK, Eberly VJ, Brown DA, Sullivan KJ. Gait parameters associated with responsiveness to treadmill training with body weight sup port after stroke: An exploratory study. Phys Ther .90:209 223 148. Kawashima N, Nozaki D, Abe MO, Akai M, Nakazawa K. Alternate leg movement amplifies locomotor like muscle activity in spinal cord injured persons. J Neurophysiol 2005;93:777 785 149. K autz SA, Brown DA. Relationships between timing of muscle excitation and impaired motor performance during cyclical lower extremity movement in post stroke hemiplegia. Brain 1998;121 ( Pt 3):515 526 150. Renshaw D, Bice, M.R., Cassidy, C., Eldridge, J.A ., A comparison of three computer based methods used to determine emg signal amplitude. Int J Exerc Sci 2010;3:43 48 151. Farina D, Merletti R, Enoka RM. The extraction of neural strategies from the surface emg. J Appl Physiol 2004;96:1486 1495 152 Burdette BH, Gale EN. Reliability of surface electromyography of the masseteric and anterior temporal areas. Arch Oral Biol 1990;35:747 751 153. Farina D. Interpretation of the surface electromyogram in dynamic contractions. Exerc Sport Sci Rev 2006; 34:121 127 154. Sasaki K, Neptune RR, Kautz SA. The relationships between muscle, external, internal and joint mechanical work during normal walking. J Exp Biol 2009;212:738 744

PAGE 182

182 155. Winter DA. A new definition of mechanical work done in human movemen t. J Appl Physiol 1979;46:79 83 156. Aissaoui R, Allard P, Junqua A, Frossard L, Duhaime M. Internal work estimation in three dimensional gait analysis. Med Biol Eng Comput 1996;34:467 471 157. Jacobs R, Macpherson JM. Two functional muscle groupings during postural equilibrium tasks in standing cats. J Neurophysiol 1996;76:2402 2411 158. Levin MF, Kleim JA, Wolf SL. What do motor "Recovery" And "Compensation" Mean in patients following stroke ? Neurorehabil Neural Repair 2009;23:313 319 159. W.H.O. International classification of functioning, disability and health. Geneva: World Health Organization; 2001. 160. Turns LJ, Neptune RR, Kautz SA. Relationships between muscle activity and anterop osterior ground reaction forces in hemiparetic walking. Arch Phys Med Rehabil 2007;88:1127 1135 161. Higginson JS, Zajac FE, Neptune RR, Kautz SA, Delp SL. Muscle contributions to support during gait in an individual with post stroke hemiparesis. J Biom ech 2006;39:1769 1777 162. Neptune RR, Kautz SA, Zajac FE. Contributions of the individual ankle plantar flexors to support, forward progression and swing initiation during walking. J Biomech 2001;34:1387 1398 163. Neptune RR, Zajac FE, Kautz SA. Mus cle force redistributes segmental power for body progression during walking. Gait Posture 2004;19:194 205 164. Balasubramanian CK, Bowden MG, Neptune RR, Kautz SA. Relationship between step length asymmetry and walking performance in subjects with chron ic hemiparesis. Arch Phys Med Rehabil 2007;88:43 49 165. Tachibana A, McVea DA, Donelan JM, Pearson KG. Recruitment of gastrocnemius muscles during the swing phase of stepping following partial denervation of knee flexor muscles in the cat. Exp Brain Re s 2006;169:449 460 166. Annaswamy TM, Giddings CJ, Della Croce U, Kerrigan DC. Rectus femoris: Its role in normal gait. Arch Phys Med Rehabil 1999;80:930 934 167. Andersson EA, Nilsson J, Thorstensson A. Intramuscular emg from the hip flexor muscles during human locomotion. Acta Physiol Scand 1997;161:361 370

PAGE 183

183 168. Nene A, Byrne C, Hermens H. Is rectus femoris really a part of quadriceps? Assessment of rectus femoris function during gait in able bodied adults. Gait Posture 2004;20:1 13 169. Hernandez A, Lenz AL, Thelen DG. Electrical stimulation of the rectus femoris during pre swing diminishes hip and knee flexion during the swing phase of normal gait. IEEE Trans Neural Syst Rehabil Eng .18:523 530 170. Goldberg SR, Anderson FC, Pandy MG, Delp SL. Muscles that influence knee flexion velocity in double support: Implications for stiff knee gait. J Biomech 2004;37:1189 1196 171. Goldberg SR, Ounpuu S, Arnold AS, Gage JR, Delp SL. Kinematic and kinetic fac tors that correlate with improved knee flexion following treatment for stiff knee gait. J Biomech 2006;39:689 698 172. Reinbolt JA, Fox MD, Arnold AS, Ounpuu S, Delp SL. Importance of preswing rectus femoris activity in stiff knee gait. J Biomech 2008; 41:2362 2369 173. Zajac FE, Neptune RR, Kautz SA. Biomechanics and muscle coordination of human walking: Part ii: Lessons from dynamical simulations and clinical implications. Gait Posture 2003;17:1 17 174. Arnold AS, Anderson FC, Pandy MG, Delp SL. M uscular contributions to hip and knee extension during the single limb stance phase of normal gait: A framework for investigating the causes of crouch gait. J Biomech 2005;38:2181 2189 175. Caldwell GE, Forrester LW. Estimates of mechanical work and ene rgy transfers: Demonstration of a rigid body power model of the recovery leg in gait. Med Sci Sports Exerc 1992;24:1396 1412 176. Willems PA, Cavagna GA, Heglund NC. External, internal and total work in human locomotion. J Exp Biol 1995;198:379 393 1 77. van Asseldonk EH, Buurke JH, Bloem BR, Renzenbrink GJ, Nene AV, van der Helm FC, van der Kooij H. Disentangling the contribution of the paretic and non paretic ankle to balance control in stroke patients. Exp Neurol 2006;201:441 451 178. Bagnato S, Boccagni C, Boniforti F, Trinchera A, Guercio G, Letizia G, Galardi G. Motor dysfunction of the "Non affected" Lower limb: A kinematic comparative study between hemiparetic stroke and total knee prosthesized patients. Neurol Sci 2009;30:107 113

PAGE 184

184 179. Bow den MG, Balasubramanian CK, Neptune RR, Kautz SA. Anterior posterior ground reaction forces as a measure of paretic leg contribution in hemiparetic walking. Stroke 2006;37:872 876 180. Anderson FC, Pandy MG. Individual muscle contributions to support in normal walking. Gait Posture 2003;17:159 169 181. Hall LH. Understanding changes in post stroke walking ability through simulation and experimentala nalyses. 2010;PhD 182. Czerniecki JM. Rehabilitation in limb deficiency. 1. Gait and motion analysis. Arch Phys Med Rehabil 1996;77:S3 8 183. Gage JR, Perry J, Hicks RR, Koop S, Werntz JR. Rectus femoris transfer to improve knee function of children with cerebral palsy. Dev M ed Child Neurol 1987;29:159 166 184. Perry J. Distal rectus femoris transfer. Dev Med Child Neurol 1987;29:153 158 185. Hernandez A, Lenz A, Thelen D. Electrical stimulation of the rectus femoris during pre swing diminishes hip and knee flexion during the swing phase of normal gait. IEEE Trans Neural Syst Rehabil Eng 186. Barbeau H. Locomotor training in neurorehabilitat ion: Emerging rehabilitation concepts. Neurorehabil Neural Repair 2003;17:3 11 187. Barbeau H, Ladouceur M, Norman KE, Pepin A, Leroux A. Walking after spinal cord injury: Evaluation, treatment, and functional recovery. Arch Phys Med Rehabil 1999;80:22 5 235 188. Field Fote EC, Lindley SD, Sherman AL. Locomotor training approaches for individuals with spinal cord injury: A preliminary report of walking related outcomes. J Neurol Phys Ther 2005;29:127 137 189. Behrman AL, Lawless Dixon AR, Davis SB, Bowden MG, Nair P, Phadke C, Hannold EM, Plummer P, Harkema SJ. Locomotor training progression and outcomes after incomplete spinal cord injury. Phys Ther 2005;85:1356 1371 190. Kautz SA, Duncan PW, Perera S, Neptune RR, Studenski SA. Coordination of he miparetic locomotion after stroke rehabilitation. Neurorehabil Neural Repair 2005;19:250 258 191. Sullivan KJ, Brown DA, Klassen T, Mulroy S, Ge T, Azen SP, Winstein CJ. Effects of task specific locomotor and strength training in adults who were ambulat ory after stroke: Results of the steps randomized clinical trial. Phys Ther 2007;87:1580 1602

PAGE 185

185 BIOGRAPHICAL SKETCH Bhavana Raja graduated from Holy Angels School in Rajpura, Punjab, India in 19 96 Sh e received her Bachelor in Physical Therapy from Guru Nanak Dev University, Amritsar, Punjab, India Her keenness to help people with neurological impairments lead a better quality of life encouraged her to pursue the Rehabilitation Science Doctoral program at University of Florida. She was funded by the Alumni fellowship for four years of her graduate education and worked under the expertise and guidance of Dr. Steve Kautz.