Effects of Dopaminergic Therapy on Locomotor Adaptation and Adaptive Learning in Persons with Parkinson's Disease

MISSING IMAGE

Material Information

Title:
Effects of Dopaminergic Therapy on Locomotor Adaptation and Adaptive Learning in Persons with Parkinson's Disease
Physical Description:
1 online resource (196 p.)
Language:
english
Creator:
Roemmich, Ryan T
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Health and Human Performance, Applied Physiology and Kinesiology
Committee Chair:
Hass, Christopher J
Committee Members:
Coombes, Stephen A
Tillman, Mark D
Fregly, Benjamin J

Subjects

Subjects / Keywords:
adaptation -- disease -- gait -- parkinson -- treadmill
Applied Physiology and Kinesiology -- Dissertations, Academic -- UF
Genre:
Health and Human Performance thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Human gait must be highly adaptable to accommodate constantly changing environments. However, we are only beginning to understand how humans adapt and store gait patterns.  Persons with Parkinson’s disease (PD) exhibit a variety of locomotor deficits, including decreased gait velocity and the presence of gait asymmetry. While some treatments induce modest gait improvements, no interventions exist which restore parkinsonian gait to functionality similar to that of healthy adults.  Thus, understanding how persons with PD adapt and store gait patterns and what neural mechanisms facilitate these processes could have significant impact on gait rehabilitation in PD.  The purposes of this study were to investigate:  1) locomotor adaptation and adaptive learning in persons with PD in their native, unmedicated state and 2) how dopaminergic therapy may affect abilities to adapt and store new gait patterns in persons with PD.  Ten participants with idiopathic PD who were being treated with stable doses of orally-administered dopaminergic therapy participated in this study.  All participants performed two randomized sessions of testing on separate days:  once while optimally-medicated (ON meds) and once after withdrawal from dopaminergic medication for at least 12 hours (OFF meds).  During each session, we investigated locomotor adaptation as the participants walked on a split-belt treadmill (SBT)for ten minutes while the more-affected leg walked at 1.0 m/s and the less-affected leg walked at .50 m/s.  We then assessed locomotor adaptive learning by observing:  1) the magnitude of the aftereffects during de-adaptation (once the belts returned to tied speeds immediately following SBT walking) and 2) the savings during re-adaptation (as the participants performed the same SBT walking task a second time after washout of aftereffects following the initial SBT task). When OFF meds, persons with PD exhibited significant decreases in step length aftereffects such that significantly less step length asymmetry was observed during de-adaptation. However, both locomotor adaptation and savings were relatively unaffected by dopamine.  These findings suggest that dopaminergic pathways influence the storage of aftereffects but do not influence locomotor adaptation or savings. Thus, it appears important that persons with PD should be optimally-medicated if walking on the SBT in a rehabilitation setting.
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Ryan T Roemmich.
Thesis:
Thesis (Ph.D.)--University of Florida, 2013.
Local:
Adviser: Hass, Christopher J.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2014-08-31

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2013
System ID:
UFE0045787:00001


This item is only available as the following downloads:


Full Text

PAGE 1

1 EFFECTS OF DOPAMINERGIC THERAPY ON LOCOMOTOR ADAPTATION AND By RYAN T. ROEMMICH 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 2013

PAGE 2

2 2013 Ryan T. Roemmich

PAGE 3

3 To my family, mentors, and all the participants who contributed to this research

PAGE 4

4 ACKNOWLEDGMENTS I would first like to thank Shay for her patience and selflessness. Her understanding and support of my graduate education and research goals were monumental in allowing me to put forth my best efforts throughout my doctoral studies; words cannot express my gratitude. I would also like to thank my parents and my brother my parents for providing endless support and an entirely hands off approach throughout my academic endeavors and my brother for providing endless comic relief and everything else that go es along with being a great brother. I thank my mentor Dr. been exposed to in my life and for putting faith in me to accomplish anything he set before me. I thank him for teaching me how to become a scientist for always making it easy for me to put forth my best effort, and for being a good friend. I thank Dr. Mark I also thank Dr. B.J. Fregly and Dr. Stephen Coombes. Dr. Fregly has been instrumental in teaching me new analytical techniques and I have greatly appreciated his help and guidance I thank Dr. Coombes for his availability to answer my questions and for the interesting d Dr. Tillman, Dr. Fregly and Dr. Coombes have been great committee members and I have appreciated their upbeat, approachable demeanors. I would also lik e to thank Dr. Michael Okun, Pam Zeilman and the staff at the UF Center for Movement Disorders and Neurorestoration for being instrumental in helping me with my essential tremor project. Finally, I thank all of my friends in the Biomechanics/Applied Neuromechanics Lab and elsewhere in the department for all the help and all the laughs.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURE S ................................ ................................ ................................ ........ 10 LIST OF ABBREVIATIONS ................................ ................................ ........................... 14 ABSTRACT ................................ ................................ ................................ ................... 15 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 17 Background ................................ ................................ ................................ ............. 17 Specific Aims and Central Hypotheses ................................ ................................ ... 20 Specific Aim 1 ................................ ................................ ................................ ... 20 Central Hypothesis 1 ................................ ................................ ........................ 20 Specific Aim 2 ................................ ................................ ................................ ... 20 Central Hypothesis 2 ................................ ................................ ........................ 21 2 REVIEW OF LITERATURE ................................ ................................ .................... 22 ................................ ................................ ............................... 22 ................................ ................................ 22 Neural Control of Human Locomotion ................................ ................................ ..... 27 Locomotor Adaptation ................................ ................................ ............................. 29 ................................ ...................... 42 .............. 47 Neural Activity in PD after Dopaminergic Therapy ................................ .................. 50 3 METHODS ................................ ................................ ................................ .............. 56 Participants ................................ ................................ ................................ ............. 56 Experimental Protocol ................................ ................................ ............................. 57 Data P rocessing ................................ ................................ ................................ ..... 58 Statistical Analyses ................................ ................................ ................................ 60 Locomotor Adaptation ................................ ................................ ...................... 60 Locomotor Adaptive Learning ................................ ................................ ........... 61 4 RESULTS ................................ ................................ ................................ ............... 65 Spatiotemporal Gait Variables during Locomotor Adaptation ................................ 65 Reactively controlled Intralimb Parameters during Locomotor Adaptation ....... 65

PAGE 6

6 Stride length asymmetry ................................ ................................ ............ 65 Stance time asymmetry ................................ ................................ ............. 66 Predictively controlled Interlimb Parameters during Locomotor Adaptation ..... 66 Step length asymmetry ................................ ................................ .............. 66 Double limb support time asymmetry ................................ ......................... 66 Kinetic Gait Variables during Locomotor Adaptation ................................ ............... 67 AP GRF impulses ................................ ................................ ............................. 68 Sagittal ankle kinetics ................................ ................................ ....................... 69 Spatiotemporal Gait Variables during Locomotor Adaptive Learning ...................... 70 Reactively controlled Intralimb Parameters during De adaptation .................... 7 0 Stride length asymmetry ................................ ................................ ............ 70 Stance time asymmetry ................................ ................................ ............. 70 Predictively controlled Interlimb Parameters during De adaptation .................. 71 Step length asymmetry ................................ ................................ .............. 71 Double limb support time asymmetry ................................ ......................... 71 Associations between Step Length Asymmetry during Locomotor Adaptation and De adaptation ................................ ................................ ....... 71 Orientation of the Limbs at Heel strikes during De adaptation ......................... 72 Reactively controlled Intralimb Parameters during Savings ............................. 73 Stride length asymmetry ................................ ................................ ............ 73 Stance time asymmetry ................................ ................................ ............. 73 Predic tively controlled Interlimb Parameters during Savings ............................ 73 Step length asymmetry ................................ ................................ .............. 73 Double limb support time asymmetry ................................ ......................... 73 Kinetic Gait Variables during Locomotor Adaptive Learning ................................ ... 74 AP GRF Impulses during De adaptation ................................ .......................... 75 Sagittal Ankle Kinetics during De adaptation ................................ .................... 75 AP GRF Impulses during Savings ................................ ................................ .... 75 Sagittal Ankle Kinetics during Savings ................................ ............................. 76 5 DISCUSSION ................................ ................................ ................................ ....... 128 Effects of Levodopa on Locomotor Adaptive Learning ................................ ......... 128 Biomechanics of Reduced Step Length Aftereffects during De adaptation .......... 135 Neural Mechanisms Potentially Underlying Diminished Aftereffect Storage OFF Meds ................................ ................................ ................................ .................. 139 Clinical Implications ................................ ................................ .............................. 148 Limitations ................................ ................................ ................................ ............. 150 Conclusion ................................ ................................ ................................ ............ 150 APPENDIX: STATISTICAL ANALYSES TABLES ................................ ...................... 152 LIST OF REFERENCES ................................ ................................ ............................. 183 BIOGRAPH ICAL SKETCH ................................ ................................ .......................... 196

PAGE 7

7 LIST OF TABLES Table page 3 1 Participant characteristics and demographic information. ................................ .. 64 A 1 Multivariate results of MANOVA analyzing effects of walking condition and medication on spatiotemp oral gait characteristics during locomotor adaptation (Specific Aim 1). ................................ ................................ .............. 152 A 2 Univariate results of MANOVA analyzing effects of walking condition and medication on spatiotemporal gait characteristics during locomotor adaptation (Specific Aim 1). ................................ ................................ .............. 153 A 3 Pairwise comparisons of MANOVA analyzing effects of walking condition on spatiotemporal gait characteristics during locomotor adaptation (Specific Aim 1). ................................ ................................ ................................ ..................... 154 A 4 Multivariate results of MANOVA analyzing effects of walking condition and medication on AP GRF impulses during locomotor adaptation (Specific Aim 2). ................................ ................................ ................................ ..................... 155 A 5 Univariate results of MANOVA analyzing effects of walking condition and medication on AP GRF impulses during locomotor adaptation (Specific Aim 2). ................................ ................................ ................................ ..................... 157 A 6 Pairwise comparisons of MANOVA analyzing effects of walking condition on AP GRF impulses during locomotor adaptation (Specific Aim 2). ..................... 159 A 7 Multivariate results of MANOVA analyzing effects of walking condition and medication on sagittal ankle kinetics during locomotor adaptation (Specific Aim 2). ................................ ................................ ................................ .............. 161 A 8 Univariate results of MANOVA analyzing effects of walking condition and medication on sagittal ankle kinetics during locomotor adaptation (Specific Aim 2). ................................ ................................ ................................ .............. 162 A 9 Pairwise comparisons of MANOVA analyzing effects of walking condition on sagittal ankle kinetics during locomotor adaptation (Specific Aim 2). ............... 163 A 10 Multivariate results of MANOVA analyzing effects of walking condition and medication on spatiotemporal gait characteristics during the de adaptation phase of locomotor adaptive learning (Specific Aim 1). ................................ .... 164 A 11 Univariate results of MANOVA analyzing effects of walking condition and medication on spatiotemporal gait characteristics during the de adaptation phase of locomotor adaptive learning (Specific Aim 1). ................................ .... 165

PAGE 8

8 A 12 ANOVA results for linear regression analyzing BASELINE, EARLY, and LATE step length asymmetry values as predictors of POST TIED step length asymmetry OFF meds (Specific Aim 1). ................................ ........................... 166 A 13 Linear regressi on results analyzing BASELINE, EARLY, and LATE step length asymmetry values as predictors of POST TIED step length asymmetry OFF meds (Specific Aim 1). ................................ ................................ .............. 166 A 14 ANOVA results for linear regression analyzing BASELINE, EARLY, and LATE step length asymmetry values as predictors of POST TIED step length asymmetry ON meds (Specific Aim 1). ................................ ............................. 167 A 15 Linear regression results analyzing BASELINE, EARLY, and LATE step length asymmetry values as predictors of POST TIED step length asymmetry ON meds (Spec ific Aim 1). ................................ ................................ ............... 167 A 16 Paired t test results analyzing sagittal hip, knee, and ankle angles at heel strikes during POST TIED betwe en medicated states (Specific Aim 1). ........... 168 A 17 Paired t test results analyzing fast and slow limb step lengths between BASELINE an d POST TIED in both medicated states (Specific Aim 1). .......... 169 A 18 Multivariate results of MANOVA analyzing effects of walking condition and medication on spatiotemporal gait characteristics during the re adaptation phase of locomotor adaptive learning relative to EARLY (Specific Aim 1). ....... 170 A 19 Univariate results of MANOVA analyzing effects of walking condition and medication on spatiotemporal gait characteristics during the re adaptation phase of locomotor adaptive learning relative to EARLY (Specific Aim 1). ....... 171 A 20 Multivariate results of MANOVA analyzing effects of walking condition and medication on spatiotemporal gait charact eristics during the re adaptation phase of locomotor adaptive learning relative to LATE (Specific Aim 1). ......... 172 A 21 Univariate resu lts of MANOVA analyzing effects of walking condition and medication on spatiotemporal gait characteristics during the re adaptation phase of locomotor adaptive learning relative to LATE (Specific Aim 1). ......... 173 A 22 Multivariate results of MANOVA analyzing effects of walking condition and medication on AP GRF impulses during the de adaptation phase of locomotor adaptive learning (Specific Aim 2). ................................ ................... 174 A 23 Multivariate results of MANOVA analyzing effects of walking condition and medication on sagittal ankle kinetics during the de adaptation phase of locomotor adaptive learning (Specific Aim 2). ................................ ................... 176

PAGE 9

9 A 24 Multivariate results of MANOVA analyzing effects of walking condition and medication on AP GRF impulses during the re adaptation phase of locomotor adaptive learning relative to EARLY (Specific Aim 2). ................................ ...... 177 A 25 Multivariate results of MANOVA analyzing effects of walking condition and medication on sagittal ankle kinetics during the re adaptation phase of locomotor adaptive learning relative to EARLY (Specific Aim 2). ..................... 179 A 26 Multivariate results of MANOVA analyzing effects of walking condition and medication on AP GRF impulses during the re adaptation phase of locomotor adaptive learning relative to LATE (Specific Aim 2). ................................ ......... 180 A 27 Multivariate results of MANOVA anal yzing effects of walking condition and medication on sagittal ankle kinetics during the re adaptation phase of locomotor adaptive learning relative to LATE (Specific Aim 2). ........................ 182

PAGE 10

10 LIST OF FIGURES Figure page 2 1 A generalized diagram of basic circuitry of the basal ganglia and related structures ................................ ................................ ................................ ............ 54 2 2 A Bertec instrumented split belt treadmill ................................ ........................... 55 3 1 An outline of the protocol used to investigate locomotor ada ptation and adaptive learning ................................ ................................ ................................ 62 3 2 Example of AP GRF curve (during conventional walking) partitioned into braking (black) and propulsive (gray) phases (Specific Aim 2) ........................... 63 3 3 Example of a sagittal ankle power curve (during conventional walking Specific Aim 2) ................................ ................................ ................................ .... 63 4 1 Mean stride length asymmetry during all walk ing conditions (Specific Aim 1) .... 77 4 2 Mean stance time asymmetry during all walking conditions (Specific Aim 1) ..... 78 4 3 Mean step length asymmetry during all walkin g conditions (Specific Aim 1) ...... 79 4 4 Mean double limb support time asymmetry during all walking conditions (Specific Aim 1) ................................ ................................ ................................ .. 80 4 5 Mean stride length asymmetry during locomotor adaptation (Specific Aim 1) .... 81 4 6 Mean stance time asymmetry during locomotor adaptation (Specific Aim 1) ..... 82 4 7 Mean step length asymmetry during locomotor adaptation (Specific Aim 1) ...... 83 4 8 Mean double limb support time asymmetry during locomotor adaptation (Specific Aim 1) ................................ ................................ ................................ .. 84 4 9 Ensemble AP GRF curves during BASELINE OFF meds (Specific Aim 2) ........ 85 4 1 0 Ensemble AP GRF curves during BASELINE ON meds (Specific Aim 2) .......... 85 4 1 1 Ensemble AP GRF curves during LATE OFF meds (Specific Aim 2) ................. 86 4 1 2 Ensemble AP GRF curves during LATE ON meds (Specific Aim 2) ................... 86 4 1 3 Ensemble sagittal ankle power curves during BASELINE OFF meds (Specific Aim 2) ................................ ................................ ................................ ................. 87 4 1 4 Ensemble sagittal ankle power curves during BA SELINE ON meds (Specific Aim 2) ................................ ................................ ................................ ................. 87

PAGE 11

11 4 1 5 Ensemble sagittal ankle power curves during LATE OFF meds (Specific Aim 2) ................................ ................................ ................................ ........................ 88 4 1 6 Ensemble sagittal ankle power curves during LATE ON meds (Specific Aim 2) ................................ ................................ ................................ ........................ 88 4 1 7 Mean braking AP GRF impulses during all walking conditions (Specific Aim 2) ................................ ................................ ................................ ........................ 89 4 18 Mean propulsive AP GRF impulses during all walking conditions (Specific Aim 2) ................................ ................................ ................................ ................. 90 4 19 Mean propulsive AP GRF impulses of the portion of the propulsive GRF prior to double limb support during all walking conditions (Specific Aim 2) ................. 91 4 2 0 Mean propulsive AP GRF impulses of the portion of the propulsive GRF during double support during all walkin g conditions (Specific Aim 2) ................. 92 4 2 1 Peak A2 ankle power during all walking conditions (Specific Aim 2) .................. 93 4 2 2 A2 ankle work during all walking conditions (Specific Aim 2) ............................. 94 4 2 3 Mean braking AP GRF impulses during locomotor adaptation (Specific Aim 2) ................................ ................................ ................................ ........................ 95 4 2 4 Mean propulsive AP GRF impulses during locomotor adaptation (Specific Aim 2) ................................ ................................ ................................ ................. 96 4 2 5 Mean propulsive AP GRF impulses of the portion of the propulsive GRF prior to double limb support during locomotor adaptation (Specific Aim 2) ................. 97 4 2 6 Mean propulsive AP GRF impulses of the portion of the propulsive GRF during double limb support during locomotor adaptation (Specific Aim 2) .......... 98 4 2 7 Peak A2 ankle power during locomotor adaptation (Specific Aim 2) .................. 99 4 28 A2 ankle work during locomotor adaptation (Specific Aim 2) ............................ 100 4 29 Mean stride len gth asymmetry aftereffects during de adaptation (Specific Aim 1) ................................ ................................ ................................ ...................... 101 4 3 0 Mean stance time asymmetry aftereffects during de adaptation (Specific Aim 1) ................................ ................................ ................................ ...................... 102 4 3 1 Mean step length asymmetry aftereffects during de adaptation (Specific Aim 1) ................................ ................................ ................................ ...................... 103 4 3 2 Mean double limb support time asymmetry aftereffects during de adaptation (Specific Aim 1) ................................ ................................ ................................ 104

PAGE 12

12 4 3 3 Association between EARLY step length asymmetry and POST TIED step length asymmetry in the OFF meds state (Specific Aim 1) ............................... 105 4 3 4 Association between LATE step length asymmetry and POST TIED step length asymmetry in the OFF meds state (Specific Aim 1) ............................... 105 4 3 5 Association between EARLY step length asymmetry and POST TIED step length asymmetry in the ON meds state (Specific Aim 1) ................................ 106 4 3 6 Association between LATE step length asymmetry and POST TIED step length asymmetry in the ON meds state (Specific Aim 1) ................................ 106 4 3 7 Ensemble orientation of the limbs at fast limb heel strike (left) and slow limb heel strike (right) during POST TIED while OFF meds (Specific Aim 1) ........... 107 4 38 Ensemble orientation of the limbs at fast limb heel strike (left) and slow limb heel strike (right) during POST TIED whil e ON meds (Specific Aim 1) ............ 107 4 39 Sagittal hip and knee angles during POST TIED in both the ON and OFF meds states at heel stri kes (Specific Aim 1) ................................ ..................... 108 4 4 0 Fast limb and slow limb step lengths during POST TIED in both the OFF (top) and ON (bottom) meds states at heel strikes (Specific Aim 1) ................. 109 4 4 1 Mean stride length asymmetry savings during re adaptation (Specific Aim 1) 110 4 4 2 Mean stance time asymmetry savings during re adaptation (Specific Aim 1) ... 111 4 4 3 Mean step length asymmetry savings during re adaptation (Specific Aim 1) ... 112 4 4 4 Mean double limb support time asymmetry savings during re adaptation (Specific Aim 1) ................................ ................................ ................................ 113 4 4 5 Ensemble AP GRF curves during POST TIED OFF meds (Specific Aim 2) ..... 114 4 4 6 Ensemble AP GRF curves during POST TIED ON meds (Spec ific Aim 2) ...... 114 4 4 7 Ensemble sagittal ankle power curves during POST TIED OFF meds (Specific Aim 2) ................................ ................................ ................................ 115 4 48 Ensemble sagittal ankle power curves during POST TIED ON meds (Specific Aim 2) ................................ ................................ ................................ ............... 115 4 49 Mean braking AP GRF impulses aftereffects during de adaptation (S pecific Aim 2) ................................ ................................ ................................ ............... 116 4 5 0 Mean propulsive AP GRF impulses aftereffects of the propulsive GRF during de adaptation (Specific Aim 2) ................................ ................................ .......... 117

PAGE 13

13 4 5 1 Mean propulsive AP GRF impulses aftereffects of the portion of the propulsive GRF prior to double limb support during de adaptation (Specific Aim 2) ................................ ................................ ................................ ............... 118 4 5 2 Mean propulsive AP GRF impulses aftereffects of the portion of the propulsive GRF during double limb support during de adaptation (Specific Aim 2) ................................ ................................ ................................ ............... 119 4 5 3 Peak A2 ankle power aftereffects during de adaptation (Specific Aim 2) ......... 120 4 5 4 A2 ankle work aftereffects during de adaptation (Specific Aim 2) ..................... 121 4 5 5 Mean braking AP GRF impulses savings during re adaptation (Specific Aim 2) ................................ ................................ ................................ ...................... 122 4 5 6 Mean propulsive AP GRF impulses savings during re adaptation (Specific Aim 2) ................................ ................................ ................................ ............... 123 4 5 7 Mean propulsive AP GRF impulses savings of the portion of the propulsive GRF prior to double limb support during re adaptation (Specific Aim 2) ........... 124 4 58 Mean propulsive AP GRF impulses savings of the portion of the propulsive GRF during double limb support during re adaptation (Specific Aim 2) ............ 125 4 59 Peak A2 ankle power savings during re adaptation (Specific Aim 2) ................ 126 4 6 0 A2 ankle work savings during re adaptation (Specific Aim 2) ........................... 127

PAGE 14

14 LIST OF ABBREVIATIONS AP GRF Anterior posterior ground reaction force CPG Central pattern generator GP E Globus pallidus externus GPi Globus pallidus internus H&Y Hoehn and Yahr stage MANOVA Multivariate analysis of variance MLR Midbrain locomotor region OFF meds After 12 hour withdrawal of dopaminergic medication ON meds While optimally medicated PD PIGD Postural instability and gait difficulty PPN Pedunculopontine nucleus SBT Split belt treadmill SNpc Substantia nigra pars compacta STN Subthalamic nucleus UPDRS

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 EFFECTS OF DOPAMINERGIC THERAPY ON LOCOMOTOR ADAPTATION AND ADAPTIVE LEARNING I By Ryan T. Roemmich August 2013 Chair: Chris Hass Major: Health and Human Performance Human gait must be highly adaptable to accommodate constantly changing environments. However, we are only beginning to understand how humans adapt and store gait patterns. (PD) exhibit a variety of locomotor deficits, in c luding decreased gait velocity and the presence of gait asymmetry. While some treatments induce modest gait improvements, no interventions exist which restore parkinsonian gait to functionality similar to that of healthy adults. Thus, understanding how p ersons with PD adapt and store gait patterns and what neural mechanisms facilitate these processes could have significant impact on gait rehabilitation in PD. The purposes of this study we re to investigate : 1) locomotor adaptation and adaptive learning i n persons with PD in their native, unmedicated state and 2) how d opaminergic therapy may affect abilit ies to adapt and store new gait patterns in persons with PD. Ten participants with idiopathic PD who were being treated with stable doses of orally admi nistered dopaminergic therapy participated in this study. All participants performed two randomized sessions of testing on separate days: once while optimally medicated (ON meds) and once after withdrawal from dopaminergic medication for at

PAGE 16

16 least 12 hour s (OFF meds) During each session, w e investigated locomotor adaptation as the participants walked on a split belt treadmill (SBT) for ten minutes while the more affected leg walked at 1.0 m/s and the less affected leg walk ed at .50 m/ s We then assessed locomotor adaptive learning by observing: 1) the magnitude of the aftereffects during de adaptation ( once the belts returned to tied speeds immediately following SBT walking ) and 2) the savings during re adaptation (as the participants performed the same SBT walking task a second time after washout of aftereffects following the initial SBT task) When OFF meds, persons with PD exhibited significant decreases in step length aftereffects such that significantly less step length asymmetry was observed during de adaptation However, both locomotor adaptation and savings were relatively unaffected by d opamine. These findings suggest that dopaminergic pathways influence the storage of aftereffects but do not influence locomotor adaptation or savings. Thus, it appears important that persons with PD should be optimally medicated if walking on the SBT in a rehabilitation setting

PAGE 17

17 CHAPTER 1 INTRODUCTION Background disease (PD) is one of the most prevalent movement disorders in the adult population, affecting approximately 1% of all adults over the age of 60 and up to 3% of adults 80 years of age or older (Tanner and Goldman 1996). While PD is an idiopathic disorder of unknown origin, motor symptoms observed in persons with PD result from degeneration of dopaminergic neurons within the basal ganglia and are most commonly treated by orally administered dopaminergic medication. The cardinal motor features of PD tremor, rigidity, bradykinesia, and postural instability typically present unilaterally during initial stages of the disease then worsen bilaterally as the disease progress es and dopaminergic degeneration of the basal ganglia becomes more severe (Lee et al. 1995). Oral administration of levodopa is considered the standard treatment to counteract the dopaminergic dysfunction in PD and ameliorate parkinsonian motor deficits; as such, levodopa is typically the first therapy prescribed relatively soon after diagnosis for a large majority of persons with PD. Though dopaminergic therapy is effective in reducing the severity of motor dysfunction in PD, the dosage must be increase d over time as the disease progresses and parkinsonian motor features become less responsive to medication. These motor features often manifest as debilitating disturbances in gait and balance, as persons with PD walk with a slow, shuffling gait (Knutsson 1972) and experience falls at a frequency several times that of healthy older adults (Pickering et al. 2007). While dopaminergic treatment relieves some parkinsonian motor dysfunction, impaired dynamic stability and increased incidence of falling remai n significant contributors to lack of independence and

PAGE 18

18 decreased quality of life in persons with PD (Schrag Jahanshahi, and Quinn 2000). Thus, a better understanding of how dopaminergic therapy influences complex locomotor movements may provide a more c omplete insight into the treatment of gait deficits in persons with PD. In order to maintain dynamic stability during locomotion, human gait must be readily adapted in response to constantly changing internal and external environments. Recently, split be lt treadmills (SBT) have been used to investigate locomotor adaptation in various populations (Morton and Bastian 2006; Dietz et al. 1995; Reisman et al. 2009 ; Choi et al., 2009 ). The two SBT belts can be decoupled such that one leg can be made to walk faster than the other. During adaptation to walking on the SBT, intralimb spatiotemporal gait variables ( e.g., stride length and stance time) can be reactively altered by feedback mechanisms likely derived from the spinal cord (Rossignol et al. 1999; Pe arson, 1995) while interlimb parameters ( e.g., step length and double limb support time) appear to require cerebellar modulated predictive, feedforward control to more gradually adapt (Morton and Bastian 2006). In order to retain an adapted gait pattern, depression of cerebellar inhibition over the motor cortex is thought to be necessary. Locomotor adaptive learning, defined as the stored ability to predict a locomotor perturbation and adapt gait parameters, is thus facilitated by cerebellar actuated cha nges in cortical activity (Jayaram et al. 2011). As the cerebellum and basal ganglia both influence multiple areas of the cortex, interactions between these structures are important to modulate cortical activity. Despite the significant role of the basa l ganglia in motor planning and motor execution, these subcortical nuclei have been largely ignored in investigations of neural control of locomotor adaptation.

PAGE 19

19 Though mobility deficits related to the inability to adapt movement during transitional period s such as gait initiation, rising from a chair, and turning have been shown to be associated with an increased fear of falling in persons with PD (Rahman et al. 2011), relatively little is known about the ability of persons with PD to adapt steady state g ait. When optimally medicated, persons with PD adapt to SBT walking with limited stride length adaptation a pervasive step length asymmetry, and restriction in st ride frequencies (Roemmich et al. 2013 (in p ress ); Dietz et al. 1995), perhaps due to fau lty motor program selection by the basal ganglia. However, further study into the abilities to adapt gait in the native, unmedicated state may provide further insight into the effects of PD on locomotor adaptation and adaptive learning. For instance, the basal ganglia have previously been shown to play a role in adaptive motor learning (Marinelli et al., 2009; Bdard and Sanes, 2011; Leow et al., 2012). Further, hyperactivity of the cerebellum has been frequently observed in persons with PD during moveme nt when off medication (Yu et al. 2007; Wu and Hallett 2005) and thus may restrict the ability to store learned SBT walking patterns (Jayaram et al. 2011) As previous research has yet to investigate unmedicated persons with PD exactly how the neurological deficits native to PD affect the ability to adapt and store newly learned gait patterns remains poorly understood. Certainly, a better understanding of the dopaminergic influence on locomotor adaptation in persons with PD could provide importa nt insight into both the functional effects of parkinsonian neurological deficits on locomotor control and shed light on the neural control of locomotor adaptation in general. Therefore, the purposes of this study are to investigate 1) locomotor adaptatio n and adaptive learning persons with PD adapt

PAGE 20

20 gait in a native, unmedicated state and 2) how dopaminergic therapy may affect the ability to adapt and store new gait patterns in persons with PD. We propose that this study will provide significant informati on about the role of the basal ganglia in locomotor adaptation and locomotor adaptive learning processes as well as a better understanding of how dopaminergic therapy influences locomotor control in persons with PD. Further, we postulate that intensified basal ganglia dysfunction and potentially compensatory mechanisms of the cerebellum in the off medication state may provide further insight into how gait is adapted and new gait patterns are stored in persons with PD, advancing our understanding as to the potential value of the SBT as a rehabilitative device for gait deficits in persons with PD. Specific Aims and Central Hypotheses Specific Aim 1 To investigate the effects of dopaminergic medication on the ability of persons with PD to adapt spatiotemporal gait parameters during SBT walking store aftereffects in these parameters during de adaptation, and exhibit savings in these parameters during re adaptation Central Hypothesis 1 While adaptation will be unaffected by levodopa, the storage of aftereffects and savings of adapted spatiotemporal gait parameters during de adaptation and re adaptation, respectively, will be reduced in the unmedicated state. Specific Aim 2 To investigate the effects of dopami nergic medication on the abilities of persons with PD to adapt gait kinetics during SBT walking, store these kinetic parameters as

PAGE 21

21 aftereffects during de adaptation and reproduce these kinetic parameters as savings during re adaptation Central Hypothesis 2 Asymmetry in lower extremity kinetics will underlie asymmetry in spatiotemporal gait parameters during all treadmill walking tasks. Thus, gait kinetics during adaptation will be unaffected by levodopa while the aftereffect s and savings of adapted gait kinetics during de adaptation re adaptation respectively, will be reduced in the unmedicated state.

PAGE 22

22 CHAPTER 2 REVIEW OF LITERATURE isease disorder which affects approximately 1 1.5 million persons in the United States alone the progressive degeneration of dopaminergic neurons within the substantia nigra of the basal ganglia ( Hassler, 1938 ; Fearnley and Lees, 1991). Dysfuncti on of the basal ganglia and their associated neural networks contributes to the four cardinal motor features of PD: bradykinesia, postural instability, tremor, and rigidity. And while resting e of PD, motor function is affected and impaired across a wide variety of tasks. Though PD motor deficits are becoming increasingly well understood, persons with PD are a heterogeneous population and thus these deficits are not consistent across all perso ns affected by the disease. For instance, persons with PD are often classified into one of two primary subtypes based on the presentation of certain motor features: tremor dominant or postural instability and gait difficulty (PIGD) dominant (Jankovic et al. 1990). Despite the heterogeneity in parkinsonian phenotypes, parkinsonian motor deficits impose significant effects on independence and quality of life in nearly all persons affected by PD (Schrag, Jahanshahi, and Quinn, 2000). Neuropathology of Park isease Research spanning several decades has begun to outline and clarify patterns of neuropathology in persons with PD and describe causative relationships between subcortical dysfunction and parkinsonian motor features. Specifically, the basal ganglia

PAGE 23

23 are a multifunctional group of subcortical nuclei the striatum, pallidum, subthalamic nucleus, and substantia nigra which maintain a vast series of inputs and outputs in order to transmit signals throughout the brain. Pivotal research in the 1 980s significantly advanced the understanding of basal ganglia circuitry, outlining not only connections within subcortical structures but also suggesting interactions between the basal ganglia and multiple areas of the cortex. Alexander, DeLong, and Stri ck (1986) expanded upon traditional views which proposed that basal ganglia output consisted primarily of connections to the primary motor cortex (Kemp and Powell, 1971; Evarts and Wise, 1984) and were among the first to suggest the organization of functio nally distinct basal ganglia circuitry, postulating on connections between the basal ganglia and multiple brain areas with various functions. Accordingly, the basal ganglia are now understood to be vital components of several neural loops which facilitate signal transmission to a variety of cerebral structures including but not limited to the cortex, midbrain, thalamus, and cerebellum which influence motor function. The interactions between the cortex and subcortical structures which make up the basal ganglia motor loops have been well established. In the most basic representation of motor interactions between the basal ganglia and cortex, the cortex sends excitatory, glutamatergic signals to both the striatum (primarily the putamen) and the subthalami c nucleus (STN). The substantia nigra pars compacta (SNpc) provides both excitatory and inhibitory influence over the striatum via dopaminergic transmission to D1 and D2 striatal receptors, respectively. The striatum then sends inhibitory, GABAergic sign als to both the internal and external segments of the globus pallidus. The external segment of the globus pallidus (GPe) exerts an inhibitory GABAergic

PAGE 24

24 influence primarily over the STN. Opposite the inhibitory transmissions of the striatum, the STN modul ates the activity of the internal segment of the globus pallidus (GPi) by sending an excitatory glutamatergic signal to the GPi. As the primary output nucleus of the basal ganglia, the GPi then sends an inhibitory GABAergic signal to the thalamus, which i n turn relays an excitatory glutamatergic signal back to the cortex. The theoretical framework of functional segregation of interactions between the basal ganglia and cortical structures eventually led to the so called rate model of parkinsonism, which has become one of the most widely accepted (though it should be noted, not universally accepted) pathophysiologic models of basal ganglia function in PD. The modern rate model is derived largely from the identification three primary pathways through which th e basal ganglia, thalamus, and cortex interact in the healthy brain to execute and coordinate voluntary movements (including gait): the hyperdirect, direct, and indirect pathways (Nambu, Tokuno, and Takada, 2002; Albin, Young, and Penney, 1989; Alexander and Crutcher, 1990). When voluntarily initiating movement, these three pathways are essential in coordinating the movement through a highly specific selection of available motor programs (Mink, 1996). First, the hyperdirect the system through excitation of the STN by the cortex, which in turn excites the GPi and ultimately decreases thalamic excitation of the cortex. It has been theorized that this pathway is crucial in suppressing both intended and competing motor programs by inhibiting activ ity of the motor cortex (Nambu et al., 2000). Subsequently, the direct pathway occurs as the cortex sends an excitatory signal to the striatum, which then inhibits the GPi and eventually escalates thalamic excitation of the cortex. The direct pathway is thus thought to enhance the desired

PAGE 25

25 motor program for execution of the selected action (Alexander and Crutcher, 1990; DeLong, 1990). Finally, the indirect pathway inhibits competing motor programs in order to efficiently execute the des ired movement as the striatum inhibits activity of the GPe, in turn increasing the excitatory influence of the STN over the GPi and, ultimately, decreasing the thalamic excitation of the cortex (Mink, 1993). In persons with PD, degeneration of dopaminergic neurons within the SNpc disrupts the normal function of motor program selection by the basal ganglia circuitry (Hassler, 1938; Fearnley and Lees, 1991), resulting in improper enhancement of desired motor programs by the direct pathway and ineffective inhi bition of competing programs by the indirect pathway (Albin, Penney, and Young, 1989; DeLong, 1990). These irregularities in PD motor control manifest in motor patterns which are characterized by tremor, rigidity, bradykinesia, and postural instability. Accordingly, dopaminergic dysfunction often has debilitat ing consequences on gait. Parkinsonian gait is typically characterized by slowed walking speeds accompanied by short, shuffling steps and a stooped posture (Knutsson, 1972). And while the basal ga nglia likely play some role in gait deficits observed in PD, locomotion is influenced not only directly by the aforementioned basal ganglia interactions with the thalamus and cortex but also perhaps predominantly through reciprocal circuitry connecting the basal ganglia and the midbrain locomotor region (MLR) (Shute and Lewis 1967; Lee, Rinne, and Marsden, 2000). Consequently, in addition to basal ganglia structures, the pedunculopontine nucleus (PPN) of the MLR has recently become a primary target for tr eatments aimed at locomotor improvement (Mazzone, et al., 2005; Plaha et al. 2005; Tykocki et al. 2011).

PAGE 26

26 The PPN is composed primarily of cholinergic neurons which have been shown to be reduced in some persons with PD (Hirsch et al. 1987), particularl y those who experience frequent falling and balance deficits (Karachi et al. 2010). Like the basal ganglia, the PPN appears to exhibit connections to various motor regions throughout the central nervous system. In addition to reciprocal connections with the basal ganglia, the PPN also interacts with important locomotor control regions in the cerebellum (Yeo et al., 2012) and the spinal cord (Arava muth an et al. 2007; Pierantozzi et al. 2008) to influence the initiation, speed manipulation, and termination of gait (Lee Rinne, and Marsden 2000). Further, recent research on the se interactions has produced intriguing results which suggest that dysfunction of the PPN may be a major contributor to the freezing of gait phenomenon frequently observed in persons with PD (Schweder et al. 2010; Lewis and Barker, 200 9 ). Accordingly, th e effects of PPN stimulation on freezing of gait have become a highly emphasized area of research in the treatment of parkinsonian locomotor deficits (Thevathasan et al. 2012; Thevathasan et al. 2011; Wilcox et al., 2011; Ferraye et al., 2010), though th e results to date are equivocal. As the cerebellum is theorized to share connections with the basal ganglia, PPN, and cortical structures within the frontal lobe, it is not surprising that abnormal cerebellar function has been suggested to also contribute to some of the motor changes which occur in persons with PD. Hyperactivity of the cerebellum has been observed in persons with PD when performing upper extremity movements (Wu and Hallett 2005; Yu et al. 2007; Cerasa et al. 2006) and walking (Hanakawa et al. 1999). This cerebellar hyperactivity may act as necessary compensation for motor dysfunction of

PAGE 27

27 the basal ganglia; however, the effects of cerebellar hyperactivity on locomotor function in PD are largely unknown. Thus, locomotor disturbance in per sons with PD is a multifarious phenomenon. Dysfunction within the basal ganglia results in disruption of signal transmission and activation patterns throughout a variety of important associations with neural networks governing gait. Among these are netwo rks which influence spinal control of locomotion through brainstem structures such as the MLR and those which influence cerebellar structures facilitating adaptation in the walking pattern. Moreover, how gait and on line changes in gait are controlled in PD remains a complicated question warranting further research. Neural C ontrol of H uman L ocomotion Human gait is a neurologically complex task which is influenced and coordinated by multiple motor pathways within the central nervous system. Locomotion in humans is thought to be under both spinal and supraspinal control. In quadrupedal animals, a role for the regulation of stepping by spinal central pattern generators (CPGs) has been suggested as decerebrate cats retain the ability to walk in a straight li ne on a treadmill (Rossignol et al. 1999; Forssberg 1980). The role of CPGs in humans is less established, though the ability of patients with spinal transections to walk with partial body weight support implies a certain level of spinal control over hum an gait (Dietz, Colombo, and Jensen, 1994; Dietz, 2003). Moreover, recent research regarding muscle information into specific motor modules. These modules are thought to coordina te activation of multiple muscles simultaneously in order to simplify neuromuscular control of gait (Ivanenko, Poppele, and Lacquaniti, 2004). The neural networks governing the

PAGE 28

28 modules also appear to be susceptible to disruption through neurological insul t (Clark et al., 2010). In humans, supraspinal control of gait is better understood. The cortico basal thalamic loops outlined above modulate cortical signals prior to reciprocal transmission to locomotor centers in the MLR and cerebellum (Jahn et al. 20 08). In turn, the MLR is thought to mediate activity of the CPGs in the spinal cord (Pierantozzi et al. 2008) and likely influences gait initiation in humans, as damage to the PPN and other nuclei in the MLR have resulted in freezing of gait episodes sim ilar to those ob served in persons with PD (Ma s d eu et al. 1994; Kuo, Kenney, and Jankovic, 2008; Hathout and Bhidayasiri 2005). Further, as mentioned above, stimulation of the PPN has demonstrated therapeutic potential for reducing freezing of gait in pe rsons with PD (Thevathasan et al. 2011; Thevathasan et al. 2012; Ferraye et al. 2010; Wilcox et al. 2011; Mazzone et al. 2005). In addition to the cortico basal thalamic loops and the MLR, the cerebellum exhibits significant influence over the control of gait. The ponto medullary reticular formation is a primary site of interaction between descending signals from the MLR and cerebellum and ascending information coded within the spinal cord (Jahn et al. 2008). Here, descending motor information from the MLR is integrated with sensory feedback from the spinal cord and subsequently transmitted to the cerebellum. The cerebellum then processes and alters these signals on line to regulate, coordinate, and adapt gait (Morton and Bastian 2004; Morton and Bastian 2007). The medial zone of the cerebellum is typically associated with control of extensor tone to maintain upright posture and rhythmic timing of alternating flexor and extensor movements during gait

PAGE 29

29 (Sprague and Chambers 1953; Orlovsky, 19 72). The role of the intermediate zone in control of locomotion in humans is less clear, though it appears to assist in providing consistency and accuracy to limb kinematics (Chambers and Sprague 1955; Yu and Eidelberg 1983). Finally, the lateral zone of the cerebellum plays a significant role in adapting gait patterns to difficult environmental demands or novel contexts (Schwartz Ebner, and Bloedel 1987; Chambers and Sprague 1955; Morton and Bastian 2006). In summary, cortically and spinally gener ated signals descend and ascend, respectively, through a series of networks involving the basal ganglia, thalamus, MLR, cerebellum, and ponto medullary reticular formation in order to efficiently initiate, coordinate, modify, and terminate gait in humans. Therefore, abnormal activity at any location within this vast series of loops can alter the activity of other structures throughout the network and ultimately cause significant dysfunction of the entire locomotor system. Moreover, it is imperative that w e best understand not only how pathology within individual structures affects locomotor performance, but also how compensation for dysfunctional components influences human gait patterns. Locomotor A daptation Within traditional motor adaptation paradigms, adaptive learning is assessed by the degree of the aftereffect observed during a de adaptation task and the amount of savings upon a second exposure to the adaptation task (i.e. a re adaptation task) For instance, consider the commonly investigated visuo motor adaptation task dur in g which a participant throws a ball at a stationary target while wearing prism glasses which laterally displace the field of vision. A review by Kornheiser (1976) describes the task as follows. Initially, the field of view is d isplaced by the prisms such that the thrower does not have an accurate perception as to where the target is located on the wall.

PAGE 30

30 Thus, as the thrower initially throw s the ball at the target, sensorimotor calibration is disrupted such that the thrower will miss the target laterally by a certain distance similar to the amount of lateral displacement of the field of view induced by the glasses, given that the throw would otherwise have been accurate. However, given sufficient practice, eventually the thrower accurately hit despite the displacement of the visual field. This period is considered the accuracy thro ugh trial and error practice. Following motor adaptation, adaptive learning can be assessed through the storage of aftereffects and savings (Smith et al., 2006). Once the glasses are removed and the thrower is again asked to hit the target (i.e. the thrower will initially miss the target in the opposite direction. This portion of the task is indicative as to what degree the thrower stored the adapted movement pattern upon re e xposure to a familiar task (i.e. throwing a ball at a stationary target without visuomotor perturbation). Eventually, performance returns to baseline and the thrower is able to hit the target with a baseline degree of accuracy. Then, savings of the adapt ed pattern can be assessed by having the thrower again wear the prism glasses and repeat the throwing task. If the thrower demonstrates savings of the task, the adaptation rate will be faster during the re adaptation task and the thrower will be able to a ccurately hit the target earlier in the re adaptation process as compared to the initial adaptation Recent research has begun to investigate adaptation during steady state locomotion in humans using similar ly structured paradigms. The general protocol ty pically follows the same baseline/adaptation/de adaptation/re adaptation template.

PAGE 31

31 These experiments are important in advancing the understanding as to how h umans precisely coordinate the activities of many muscles and body segments to control everyday mo vements such as walking. Th e s e movements are coordinated in the face of an environment that is constantly changing in complexity. Indeed, adaptation of human walking patterns in response to changes in internal and external environments is essential for s afe, efficient community ambulation. Locomotor adaptation is a process during which changes in locomotor output are stabilized over time by the central forward predictive motor actions and sensorimotor feedback (Purz ner et al. 2007; Cavaco et al. 2011; Bdard and Sanes 2011; Bares et al. 2007; Bares et al. 2010; Jayaram et al. 2011; Morton and Bastian 2006). Locomotor adaptation has previously been induced within a laboratory setting using a variety of different methods. Changes in the visual field or visual stimuli, often induced by virtual reality, have been shown to influence gait adaptation (Buekers et al. 1999; Mukherjee et al. 2011; Sheik Nainar and Kaber 2007) and thus the use of virtual reali ty techniques in gait rehabilitation has become widely studied in a variety of populations (Baram and Miller 2006; Kim et al. 2009; Yang et al. 2008; Mirelman et al. 2010; Griffin et al. 2011; Darter and Wilken 2011). Gait adaptation can also be ind uced by directly altering the mechanical demands of the task. For instance, gait adaptation has been studied in response to unfamiliar destabilizing walking surfaces (Cappellini et al. 2010; Gates et al. 2012), when walking with additional resistance un ilaterally applied to a lower limb (Savin Tseng, and Morton 2010), and with the assistance of a robotic exoskeleton (Lewis and Ferris 2011). A multitude of treadmill

PAGE 32

32 based paradigms have also been introduced to study locomotor adaptation and the storag e of adaptation aftereffects after complex aerobic exercise, including podokinetic treadmills (Earhart et al. 2001), laterally oscillating treadmills (Brady, Peters, and Bloomberg, 2009), unilateral stepping (Kahn and Hornby 2009), and split belt treadmi lls (Reisman et al. 2005). Of these, split belt treadmills have garnered significant interest in clinical and neurophysiological research due to the robust adaptation and aftereffects induced by the asymmetric walking patterns and the potential for gait symmetry restoration in clinical settings. Split belt treadmills (SBT) allow for creation of complex asymmetric walking patterns and have recently been used to facilitate understanding of locomotor adaptation in various populations (Morton and Bastian 20 06; Dietz et al. 1995; Reisman et al. 2009). The two SBT belts can be decoupled such that one leg walks often observed after lower extremity injury. During adaptati on to the SBT, intralimb spatiotemporal variables ( e.g. stride length and stance time) are reactively altered by feedback mechanisms likely derived from the spinal cord (Rossignol et al. 1999; Pearson, 1995) while interlimb parameters ( e.g. step length an d double limb support time) require cerebellar modulated predictive, feed forward control to more gradually adapt (Morton and Bastian 2006). As the cerebellum exerts a tonic inhibitory influence over the cortex, previous research has suggested that depre ssion of cerebellar inhibition over the motor cortex is required to facilitate the storage of aftereffects following SBT walking (Jayaram et al. 2011).

PAGE 33

33 Recent research by Reisman and colleagues has begun to describe in greater detail how the central nervo us system utilizes these feedback and feedforward processes to regulate and control the ability to adapt gait patterns (Reisman et al., 2005) When the speed of the belts is decoupled, spatiotemporal gait parameters undergo a variety of different changes and display several distinct patterns of asymmetry. In terms of temporal variables, total stride time is symmetric throughout the SBT walking session, even during the initial adaptation to SBT walking. That is, the limbs maintain a 1:1 stepping cadence. However, the limb walking on the faster belt more time in stance. Stance time is thought to be a reactively controlled parameter governed by feedback mechanisms as it becomes asymmetric almost immediately and stabilizes asymmetrically throughout the duration of the SBT walking session. While stanc e time is an intralimb parameter (i.e. only dependent upon the movement of one limb), d ouble limb support time is an interlimb parameter which depends on the time during which both limbs are in contact with the ground. During SBT walking, the double limb support period from slow limb heel strike to fast limb toe off fast initially short er than that from the heel strike of the fast limb to the slow limb toe off slow H owever, over time, the two double limb support periods of the gait cycle become symmetric as the central nervous system begins to predict the SBT walking pattern and adjust the temporal coordination of the limbs accordingly Thus, double limb support time is thought to be a predictively contr olled parameter governed by feedforward mechanisms.

PAGE 34

34 Spatial gait parameters are adjusted in conjunction with temporal variables (Reisman et al. 2005). Stride length is adapted in similar fashion to stance time such that stride length is immediately lon ger in the fast limb as compared to the slow limb and this asymmetry persists throughout the adaptation process Thus as it is a rapidly adapting intralimb parameter, stride length is also thought to be reactively controlled by feedback mechanisms Conv ersely, s tep length adaptation mirrors the double limb support time pattern, demonstrating asymmetry during initial adaptation but eventually becoming symmetrical in a predictively controlled fashion. While it is true that step length relies on the positi on of the two feet at heel strike of the leading limb, it is not entirely correct to characterize step length as a purely spatial parameter. For instance, shifts in the mid stance angle around which the limb rotates (center of oscillation) or in the tempo ral phasing of the limbs (limb phasing) can induce changes in step length (Malone Bastian, and Torres Oviedo 2012). Thus, it is important that both the spatial (center of oscillation) and temporal (limb phasing) characteristics of step length be taken i nto account when analyzing the step length asymmetry induced by the SBT. That said, w hen c oupled with the adaptation patterns of temporal parameters, the p revious findings by Reisman and colleagues suggest that intralimb parameters (stride length, stance time ) stabilize asymmetrically under reactive, feedback driven control while interlimb parameters (step length, double limb support time) are initially asymmetric but eventually stabilize symmetrically after being predictively modulated by feedforward mech anisms (Reisman et al. 2005) So in sum, the interlimb gait parameters behave similarly to the throwing performance in the prism throwing task described above; they are initially asymmetric but gradually adapt to symmetry throughout the adaptation task,

PAGE 35

35 demonstrate robust aftereffects upon return to conventional walking during de adaptation and exhibit savings upon re adaptation to a second SBT walking bout. The spatiotemporal adaptations during SBT walking manifest from changes in kinematic and kinetic gait profiles. The sagittal joint angle profiles during fast and slow conventional walking are preserved in the joint angle profiles of the fast and slow limbs during SBT walking, respectively. That is, the kinematics of both limbs during fast conventional walking are similar to the kinematics of the fast limb during SBT and the kinematics of both limbs during slow conventional walking are similar to the kinematics of the slow limb during SBT (Re isman et al. 2005). This results in asymmetric sagittal joint angle patterns, predominantly at the ankle, which contribute to the spatiotemporal adaptations which accommodate SBT walking. The kinetics of SBT walking are also significantly asymmetric. Pr evious work by our lab has demonstrated the fast limb produced higher propulsive and braking antero posterior ground reaction force (AP GRF) impulses, higher ankle moment impulses during the propulsive and braking phases, and lower knee moment impulses dur ing the propulsive phase when compared to the slow limb during SBT walking (Roemmich Stegemller, and Hass 2012). We also observed that the knee moment impulse was significantly higher in the slow limb during braking than in the fast limb. The ankle mo ment impulses in the fast limb were negatively associated with the difference in magnitude of the belt speeds during SBT walking, indicating that a greater difference in belt speeds may lessen the contribution of the ankle to the overall braking and propul sive joint moments during SBT.

PAGE 36

36 The spatiotemporal, kinetic, and kinematic changes observed during SBT walking over a prolonged period of time result in the storage of aftereffects of an altered locomotor pattern to both conventional treadmill walking (Reis man et al. 2005) and overground walking (Reisman et al. 2007). Predictively controlled spatiotemporal gait parameters are asymmetric during conventional walking immediately following bouts of SBT walking while the reactively modulated parameters are rel atively symmetrical. However, the spatiotemporal gait parameters observed during conventional walking ( i.e. aftereffects ) display an asymmetric pattern which is opposite the asymmetry demonstrated during SBT walking. For instance, immediately following S BT adaptation, the fast limb takes a significantly longer step than the slow limb during conventional walking despite taking a significantly shorter step during the SBT walking. A similar aftereffect is observed in double limb support time, which becomes longer in the fast limb immediately following SBT walking. Very little asymmetry is observed in stride length or stance time during the de adaptation period These findings suggest that, after SBT walking, reactively controlled interlimb spatiotemporal g ait parameters (i.e. stride length and stance time ) demonstrate little aftereffect during subsequent conventional walking while predictively controlled intralimb parameters (i.e. step length and double limb support time) are stored and transferred into nor mally symmetric conventional gait patterns (Reisman et al. 2005). However, very little is known about how gait kinetics are stored upon exposure to conventional gait following SBT walking. These results provided promising information regarding the potent ial for locomotor adaptation protocols in the rehabilitation and restoration of symmetrical gait patterns in pathological populations characterized by gait asymmetry. In a series of

PAGE 37

37 manuscripts, Reisman and colleagues demonstrated that SBT training could restore gait symmetry in persons post stroke (Reisman et al. 2005; Reisman et al., 2007; Reisman et al., 2009; Reisman et al. 2010 ; Reisman et al., 2013 ). Persons post stroke often exhibit shorter step lengths with the non paretic limb due to decreased propulsion of the center of mass by the paretic limb (Balasubramanian et al. 2007; Allen Kautz, and Neptune 2011). Reisman and colleagues observed that persons post stroke adapted to an acute bout of SBT walking in similar fashion to healthy adults and consequently stored aftereffects to a greater degree than healthy participants, which could temporarily restore step length symmetry in overground gait (Reisman et al. 2009). By providing chronic training of the paretic limb (taking the longer step) on the slow belt, these investigators were able to partially restore step length symmetry on a more permanent timescale in one person post stroke (Reisman et al. 2010). The authors have concluded that the presence of adapted gait parameters and subsequent aftereffects are largely due to error correction mechanisms through which the central nervous parameters) in the back to symmetry. This error correction hypothesis suggests that, once the stimulus inducing the asymmetry (i.e. the SBT) is removed, the central nerv ous system asymmetry in predictively controlled gait parameters during the aftereffect observed in conventional walking following SBT exercise (whereas reactively controlled parameters are immediately returned to symmetry in similar fashion to which they immediately stabilize asymmetrically during SBT walking). O ur findings which suggest that the

PAGE 38

38 kinetics are distinctly modified during SBT walking suggest that from a mechanic al standpoint, the SBT seems to induce an entirely new gait pattern Whether these kinetic patterns are learned and demonstrate aftereffects during de adaptation to conventional treadmill walking or savings during re adaptation is unknown Thus, the biom echanical mechanisms which underlie the aftereffects observed in conventional gait immediately following SBT walking and savings during re adaptation remain somewhat unclear. In any regard, the SBT certainly appears to possess the potential to be a valuabl e tool in the rehabilitation of asymmetric gait patterns and research has begun to focus on the optimization of SBT based training protocols in order to maximize patient locomotor improvements and address patient specific deficits. These studies have focu sed on optimization of both the ability to learn the SBT g ait pattern and the ability to store these adapted gait parameters to conventional walking. Investigating the ability to learn the SBT gait pattern, Malone and Bastian demonstrated that conscious c orrection of step length asymmetry leads to fa ster adaptation but reduces the aftereffects (Malone and Bastian 2010). Conversely, distraction of step length correction slows adaptation but increases the aftereffects. These same investigators later demon strated that both after SBT exercise with conventional treadmill walking) and repeated switching between SBT and conventional walking have significant effects on the abi lity to learn the adaptation task and recreate it in later sessions (Malone Vasudevan, and Bastian 2011). Groups of healthy adults that underwent SBT walking protocols which only required one adaptation period (that is, one switch from conventional walk ing to SBT

PAGE 39

39 walking) or had the SBT gait pattern washed out afterward by conventional walking took longer to re adapt to the same SBT walking conditions the following day as compared to groups who were required to either switch between SBT walking and conve ntional walking multiple times within the same session or who did not perform a washout trial after SBT walking. Interestingly, the duration of the SBT training did not have a significant effect on the ability to re adapt to the SBT during the follow up s ession. Further, it has been shown that the spatial (center of oscillation) and temporal (limb phasing) characteristics of step length can be adapted separately, potentially allowing for interventions to target either spatial or temporal deficits individu ally (Malone Bastian, and Torres Oviedo 2012). The aftereffects of SBT walking during overground walking are particularly important from a clinical perspective as they show potential to alter gait asymmetry in pathological populations (Reisman et al., 2 009). T hus factors which enhance or limit SBT aftereffects have become targets of recent research. Torres Oviedo and Bastian demonstrated that gradual adaptation to SBT walking led to greater overground aftereffects while more abrupt adaptation led to g reater learning of the SBT walking pattern (Torres Oviedo and Bastian 2012). The authors interpreted these results to suggest that inducement of a series of small errors over time may result in larger aftereffects by allowing the central nervous system t o attribute these errors to a commonly performed task (in this case, walking). On the other hand, large errors may not induce the same degree of aftereffect as the central nervous system may attribute these errors to the external perturbation (in this cas e, the SBT) rather than the task itself. Torres Oviedo and Bastian also provided insight into the role of sensory cues

PAGE 40

40 (specifically vision) on locomotor adaptation and aftereffects during subsequent overground walking (Torres Oviedo and Bastian 2010). When participants closed their eyes during both the SBT walking task and during the subsequent overground gait trials, they demonstrated markedly larger aftereffects compared to participants who left their eyes open during both tasks. The investigators th en dissociated the effects of removing visual perception during either the SBT walking task or during the overground gait trials. Their results demonstrated that closing the eyes while walking on the SBT and opening them during the gait trials resulted in greater carryover of temporally adapted gait characteristics while keeping the eyes open during SBT walking and closing them during the gait trials resulted in greater aftereffects of spatially adapted gait parameters. Collectively, the work by Bastian a nd colleagues suggests that storage of adapted gait parameters to overground gait seems to be independent of the duration of the adaptive training but instead depends on a multitude of other factors. These include but are not necessarily limited to the se lection of proper belt speed configurations based on the spatiotemporal asymmetry demonstrated by the patient, the magnitude and gradual inducement of the adaptation, and the limiting of sensory cues. Though limited, research on the clinical applications of locomotor adaptation has also investigated paradigms outside of the SBT literature. Kahn and Hornby reported partial restoration of step length asymmetry in persons post stroke after chronic treadmill training during which non paretic limb stepped on t he treadmill while the paretic limb stood stationary to the side (Kahn and Hornby 2009). Unpublished data from our lab has shown that unilateral stepping may be an effective alternative to SBT walking if the patient does not have access to an SBT, as uni lateral stepping induces

PAGE 41

41 step length aftereffects to overground walking in healthy adults. However, the overground aftereffects are lesser in magnitude than those observed after SBT walking (Huynh et al., in review, 2013) Further, Hong and Earhart demonstrated that persons with PD retain after rotation effects after acute bouts of walking on a rotating platform (Hong and Earhart 200 8 ). The neurological mechanisms which underlie locomotor adaptation (particularly with regard to SBT walking) remai n somewhat unclear but are becoming better understood. It has become apparent that separate mechanisms control both the reactively and predictively modulated gait parameters. The reactively modulated, fast adapting parameters are thought to be at least partially governed by feedback mechanisms of the central pattern generators of the spinal cord. Forssberg observed that locomotor adaptation on a SBT is preserved in spinalized cats (Forssberg, 1980) and further research has demonstrated a strong role of the spinal cord in reactively controlling locomotion through proprioceptive and sensory feedback in various animal models (Pearson, 1995; Rossignol et al. 1999). Though the role of the spinal cord remains relatively unclear in humans, various populations characterized by higher level neurological dysfunction including stroke (Reisman et al. 2007), PD (Dietz et al. 1995), hemispherectomy (Choi et al. 2009), and cerebellar dysfunction (Morton and Bastian 2006) retain the ability to reactively adapt to SBT walking. However, despite the well established and significant role of the basal ganglia in planning and executing motor actions, the influence of these nuclei on locomotor adaptation has not been studied.

PAGE 42

42 Alternatively, the predictively modulated parameters are generally accepted to adapt under feedforward control of the cerebellum. Morton and Bastian demonstrated that adaptation of predictively controlled gait parameters ( i.e. step length and double limb support time) is absent in persons with ce rebellar damage (Morton and Bastian 2006). The cerebellar participants in this study retained the ability to adapt stride length and stance time but were unable to eventually stabilize step length and double limb support time symmetrically to the same de gree as control participants. Further, Jayaram and colleagues have recently utilized paired pulse transcranial magnetic stimulation (first over the cerebellum and subsequently over the primary motor cortex) to show that depression of cerebellar inhibition over the motor cortex is proportional to the ability to learn and store SBT gait patterns (Jayaram et al. 2011). In a subsequent study, Jayaram and colleagues also determined that direct anodal stimulation of the cerebellum accelerated locomotor adaptat ion to SBT walking (Jayaram et al. 2012). This finding was also observed during anodal cerebellar stimulation during an upper extremity visuomotor task (Galea et al., 2011). Thus, there is robust evidence to suggest that predictive, feedforward control of motor adaptation is facilitated in large part by the cerebellum through modulation of its tonic inhibitory influence over the motor cortex. However, locomotor adaptation and locomotor adaptive learning have not been studied in populations which demonst rate cerebellar hyperactivity. Locomotor D D isease Persons with PD (particularly those classified as PIGD) experience a wide variety of debilitating gait disturbances resulting from neurological deficits intrinsic to the basal gang lia and consequent disruption of signal transmission to structures downstream from these nuclei. Parkinsonian gait is classically characterized by slow, shuffling steps and

PAGE 43

43 stooped posture (Knutsson, 1972) and, in some patients, episodic freezing of gait (Fahn, 1995). However, more recent research has begun to uncover a multitude of gait deficits which expand upon the archetypal slow, stooped PD gait pattern and provide greater detail as to how gait is impaired in this population. Excessive variability i n gait performance has been associated with falling in both prospective and retrospective studies of healthy older adults and persons with neurological disorders, including PD (Brach et al. 2005; Hausdorff Rios, and Edelberg 2001; Nakamura Meguro, and Sasaki 1996; Hausdorff, 2009). Perhaps not surprisingly, persons with PD also exhibit higher variability in stepping while initiating gait when compared to their healthy peers (Roemmich et al., 2012) These findings are further indicators of dynamic ins tability as variability during gait initiation has als o been related to falls in elderly populations (Mbourou, Lajoie, and Teasdale, 2003) As gait has been shown to be more variable even in persons with de novo PD (Baltadjieva et al. 2006), inconsistenc y of stepping during both steady state gait and gait initiation likely contributes to the increased fall incidence in persons with PD throughout the course of the disease. The increased gait variability observed in persons with PD is indicative of the ina bility of this population to rhythmically coordinate and regulate both spatial and temporal gait characteristics. Thus, the decreased ability to rhythmically coordinate spatiotemporal gait parameters bilaterally may affect the ability to adapt the gait pat tern to SBT walking in persons with PD. Further complicating gait performance in persons with PD is the presence of abnormally high asymmetry in gait and postural control. Previous research on gait asymmetry in PD has largely been limited to analyses of s patiotemporal gait

PAGE 44

44 parameters. For instance, Nanhoe Mahabier and colleagues noted significant step time asymmetry in PD (Nanhoe Mahabier et al. 2011) and Yogev and colleagues observed significant asym metry in swing time (Yogev et al. 2007 ). Consistent with the findings demonstrating that parkinsonian symptoms typically present unilaterally during early stages of the disease before progressing bilaterally (Lee et al. 1995), swing time was asymmetric even in de novo patients (Baltadjieva et al. 2006). It also does not appear that dopaminergic therapy restores symmetry in gait in PD as Miller and colleagues observed asymmetry in muscle firing patterns in the distal lower extremity musculature (tibialis anterior and medial gastrocnemius) even when the par ticipants were optimally medicated (Miller et al. 1996). Expanding on these findings, Plotnik and colleagues observed that persons who experience freezing of gait have larger swing time asymmetry than those who do not and dopaminergic medication has litt le effect on gait asymmetry in both freezers and non freezers (Plotnik et al. 2005). Further research has begun to attempt to uncover the mechanisms which underlie these asymmetric gait patterns, though this has proven difficult. Previous findings seem to indicate that gait asymmetry is not necessarily related to asymmetry in ot her parkinsonian motor characteristics, as prior studies suggest neither upper extremity motor asymmetry nor asymmetry in UPDRS ) motor scores are correlated with gait asymmetry in PD (Plotnik et al. 2005; Yogev e t al. 2007 ). However, investigators did observe a negative correlation between gait speed and swing time asymmetry while also noting a positive relationship between swing time asymmetry and stride to stri de gait variability (Yogev et al. 2007 ). It shoul d also be noted that dynamic postural control asymmetry is evident in some persons with PD (Geurts et al. 2011).

PAGE 45

45 Therefore, it seems that asymmetry in gait accompanies global gait dysfunction in PD, though not necessarily global motor dysfunction in othe r systems such as upper extremity motor control. Presently, treatments aimed at restoring gait symmetry have been underdeveloped, as a prominent aspect of gait dysfunction in PD has been largely ignored by previous exercise based and surgical intervention studies. As persons with PD walk asymmetrically (i.e. they either cannot obtain gait symmetry without perturbation/training or default to an asymmetric walking pattern), adaptation which requires a change in gait symmetry, though it may be facilitated di fferently in this population, could prove beneficial in improving global gait function in persons with PD. As previous findings have demonstrated that PD significantly impairs steady state gait and the basal ganglia play important roles in various types of motor planning and execution, implications of basal ganglia disease on locomotor adaptation and savings of adapted gait parameters are largely unknown. Only one previous study by Dietz and colleagues has examined SBT walking in PD (Dietz et al. 1995). Their results demonstrated that persons with PD reactively adapted over a similar number of strides when compared to controls. However, this study included only a one minute adaptation period and did not include assess aftereffects or savings to assess th e degree of learning. Thus, in order to investigate both the influence of the basal ganglia and cerebellum on adaptation and locomotor adaptive learning, a longer locomotor adaptat ion period and the assessment of aftereffects and savings are required. Wor k from our lab currently in review evaluated the ability of persons with PD to adapt, save and store SBT gait patterns while on medication. Our results revealed that persons with PD adapt to a SBT walking task differently than both healthy older and

PAGE 46

46 youn ger adults while the healthy older and younger adults adapted similarly. All three groups adapted stride length immediately such that stride length asymmetry during the first five strides of adaptation was significantly different from baseline; however, a daptation appeared to be restricted in PD such that stride length asymmetry during the last five strides of a ten minute adaptation period was not significantly different from the first five strides. The PD group also demonstrated a significant step lengt h asymmetry throughout the tasks during which step length is typically relatively symmetric. Thus, the savings and aftereffects of the adapted parameters in terms of step length asymmetry relative to baseline were not significantly different than healthy young or older adults, but rather a certain degree of step length asymmetry was pervasive among all SBT tasks These results were in consistent with previous studies on upper extremity and visuomotor adaptation which have shown that savings and aftereffec ts following novel skill acquisition are diminished in persons with PD ( Bdard and Sanes 2011; Fernandez Ruiz et al. 2003; Leow et al. 20 12 ; Isaias et al. 2011; Smiley Oyen, Lowry, and Emerson, 2006; Mochizuki Kawai et al. 2004) A key difference is that many of these studies were performed while the participants were unmedicated We postulated that diminished savings and aftereffects of adapted motor patterns observed in PD may result from deficits intrinsic to the basal ganglia and compensatory adj ustments by neural components which share connections with the basal ganglia. For instance, adaptation and re adaptation may be altered in PD due to the impaired ability of the basal ganglia to preferentially select the adapted motor program previously ap plied to the novel task (Leow et al. 2012). In addition, hyperactivity of the cerebellum may play a role in limiting the savings and aftereffects of the SBT gait pattern to conventional

PAGE 47

47 treadmill walking, as Jayaram and colleagues showed that depression of cerebellum inhibition over the motor cortex is required to facilitate locomotor adaptive learning on a SBT (Jayaram et al., 2011) Taken together, impaired motor program selection may be compounded by the notion that persons with PD may not preserve th e capability to depress cerebellar activity to the same extent as neurologically healthy adults, and thus locomotor adaptive learning may be impaired in this population. Effects of D opaminergic T herapy on L D isease Cotzias and col leagues introduced chronic dopaminergic therapy as an effective treatment for the motor symptoms of parkinsonism over forty years ago (Cotzias, Van Woert, and Schiffer, 1967; Cotzias, Papavasiliou, and Gellene, 1969) and levodopa remains the standard for t reatment of PD motor deficits. Since its inception as a PD treatment in the late 1960s, levodopa has become the gold standard for parkinsonian therapy. Accordingly, a large majority of persons with PD are prescribed levodopa to alleviate motor dysfunctio n soon after diagnosis. Levodopa is administered orally to persons with PD to supplement the dysfunctional transmission of dopamine to the striatum from the substantia nigra (Hornykiewicz, 1966). This therapy significantly reduces rigidity, bradykinesia, and tremor and thus has widely been considered the gold standard of PD treatment since the 1960s (Birkmayer and Hornykiewicz 1961). Not surprisingly, levodopa has also shown to be effective in ameliorating many gait deficits in persons with PD. Continuo us gait features seem to be most responsive. Some spatiotemporal gait parameters such as gait velocity, stride length, and duration of the support phases initially improve with medication (Blin et al. 1991; Krystkowiak et al. 2003). Stride duration is unaffected by levodopa (Blin et al. 1991), which is expected as cadence is typically relatively preserved in PD (Morris et al. 1996).

PAGE 48

48 Persons with PD do begin to modify stride duration and cadence during intentional modulations of gait speed, as even wh limit manipulation of stride length in this population ( Morris et al. 199 4 ). Schaafsma and colleagues demonstrated that stride to stride variability decreases in the medicated state and thus it appears stepp ing becomes more consistent with dopaminergic therapy (Schaafsma et al. 2003). These spatiotemporal gait improvements likely result largely from amelioration of whole body rigidity and bradykinesia after dopaminergic intake. For instance, levodopa restor es activation patterns of the distal lower extremity musculature during locomotion to levels similar to controls (Cioni et al. 1997; Caliandro et al. 2011). Morris and colleagues demonstrated that the range of motion during gait at the hip, knee, and an kle joints also improves after levodopa intake (Morris et al. 2005). These authors suggest that dopaminergic therapy may provide relief from the rigid and bradykinetic effects of PD by normalizing the centrally mediated motor set which is disrupted by PD Dopaminergic therapy is also not without more consistent complications and drawbacks. Among the primary shortcomings of levodopa treatment are the (Rascol, 2000; Fahn, 1976; Marsden and Parkes 1976). Dyskinesias are involuntary movements which often appear after prolonged duration of levodopa therapy. The term completely and unpredictably st ops providing any anti gradually deteriorating toward the end of a dose.

PAGE 49

49 It has been suggested that three main factors contribute to t he onset of levodopa induced dyskinesias: disease duration, duration of levodopa therapy, and the treatment dosage being administered (Carta and Bezard 2011). The DATATOP study demonstrated that approximately 9% of persons with PD presented with dyskine sias after six months of exposure to levodopa, while this number grew to 17% after 12 months and 26% after 18 months ( Parkinson Study Group 1996). Ahlskog and Muenter suggested that 90% of persons with PD experience dyskinesias after ten years of exposur e to levodopa therapy (Ahslkog and Muenter 2001). To further demonstrate that not only duration but also the dose of levodopa is important to the development of levodopa that less th an 4% of persons taking 300 mg/day or less experienced dyskinesias while 16.5% of persons taking 600 mg/day reported episodes (Fahn et al. 2004). duration of treatment. For in stance, the DATATOP study suggests that 23% of persons ted by duration of levodopa therapy as only 2% of medicated persons with PD experience these episodes after six months of treatment, increasing to 3% and 5% after 12 and 18 months, respectively ( Parkinson Study Group 1996). Similarly, the ELLDOPA trial s howed that dose of levodopa had a relatively large et al. 2004). Thus, the decisions as to when to start de novo patients on medication and at what dosages become highly influ ential on the symptomatology of the patient and very

PAGE 50

50 difficult for the physician. Accordingly, even after diagnosis, persons with PD and their physicians often elect to withhold treatment until the severity of symptoms increases in order to delay the onse t of negative levodopa related motor side effects. Neural A ctivity in PD after D opaminergic T herapy Levodopa is administered to supplement the dysfunctional dopaminergic transmission from the SNpc to the striatum as a combined D1 D2 receptor agonist. Tha t is, levodopa is neither D1 nor D2 receptor specific but rather compensates for the depletion of both dopaminergic neurotransmitters within the basal ganglia. The increased reception of dopamine within the striatum then influences the function of the ent ire cortico basal thalamic loops and, ultimately, motor networks throughout the central nervous system. By and large, previous neuroimaging studies have described partial restoration of activation patterns of cortico basal thalamic motor loops to levels ne arer to those observed in healthy older adults after dopaminergic treatment (Jubault et al. 2009; Jahanshahi et al. 2010; Kraft et al. 2009; Ng et al. 2010; Palmer et al. 2009). Within these loops, cortical activity during motor tasks in PD is depend ent upon striatal contributions to the task. For instance, Monchi and colleagues demonstrated that the cortex is hypoactive in unmedicated persons with PD during tasks which include significant striatal involvement (such as externally triggered or self in itiated movements). On the contrary, the cortex is hyperactive in PD when the task does not require striatal activation (such as performance of automatic sequences) (Monchi et al. 2007). Martinu and colleagues then followed up on these findings with a study which reported that levodopa increased cortical activity during the externally triggered and self initiated tasks while hyperactivity of the cortex during the control tasks was unaffected (Martinu

PAGE 51

51 et al. 2012). Thus, these findings suggest that lev odopa may globally increase activation of the cortical motor loops when they are hypoactive in PD; however, levodopa has little effect on cortical hyperactivity observed in this population. In addition to the cortex, subcortical structures also exhibit cha nges in activation patterns during motor tasks after dopaminergic intake. Employing a bimanual force production task and BOLD fMRI, Kraft and colleagues described significant increases in the activity of the putamen and thalamus after levodopa treatment. When on medication, the activity levels in these structures became similar to those observed in healthy older adults (Kraft et al. 2009). Palmer and colleagues observed a similar normalizing effect of levodopa on the effective connectivity and temporal dynamics of signaling patterns from the striatum and thalamus to cortical and cerebellar areas (Palmer et al. 2009). However, while improved, the connectivity and temporal patterns of subcortical and cortical activity were still mildly abnormal in PD ev en after treatment. Jahanshahi and colleagues also noted an increase in connectivity between the striatum and prefrontal cortex during a motor timing task (Jahanshahi et al. 2010). It appears as though dopaminergic therapy has restorative effects on the activation and connectivity of subcortical structures, although a certain degree of subcortical dysfunction remains even after treatment. Along the same lines, much of the research on cerebellar function after levodopa administration suggests that cerebel lar activity is at least partially normalized by levodopa intake. The cerebellum is hyperactive in persons with PD when off medication, perhaps as compensation for the dysfunctional basal ganglia (Yu et al. 2007). Lewis and colleagues suggest that cereb ellar hyperactivity decreases to normal

PAGE 52

52 levels after levodopa (Lewis et al. 2007) while Payoux and colleagues noted persistent hyperactivity in the cerebellum after treatment (Payoux et al. 2010). Several other studies have utilized various imaging tech niques to demonstrate that connectivity of the cerebellum in relation to subcortical structures is normalized in PD after dopaminergic intake (Jahanshahi et al. 2010; Wu et al. 2009; Palmer et al. 2009). Therefore, these results seem to fall in line wi th the imaging research on partial normalization of cortical and basal ganglia function after levodopa administration. In sum, levodopa appears to restore cortical, subcortical, and cerebellar activation and connectivity in persons with PD to levels more s imilar to healthy adults. However, abnormalities in neurological function remain even wh ile these persons are optimally medicated. As the neural mechanisms underlying locomotor adaptation are becoming increasingly well understood, a protocol utilizing a locomotor adaptation task may both shed light on the functional effects of abnormal activity in the motor pathways of the parkinsonian brain and help to clarify how these pathways change with levodopa intake. Specifically, by investigating persons with PD in the off medicated state, we will explore the functional locomotor effects of the altered basal ganglia and cerebellar function which results from PD and how these may change after medication. This literature review outlines an important gap in the lite rature regarding how the neurological dysfunction resulting from PD affects the ability to adapt and store new gait patterns. Investigation of SBT walking in persons with PD both on and off medication could provide valuable insight as to how parkinsonian neural deficits affect gait adaptation and how dopaminergic treatment impacts the abilit ies to adapt and store new gait patterns Since cerebellar hyperactivity and basal ganglia dysfunction are

PAGE 53

53 significantly lessened with dopaminergic medication, the ef fects of the neural deficits of the native, unmedicated PD state on locomotor adaptability and locomotor adaptive learning in PD remain unknown. These findings could have important implications in both clinical and basic science by providing information a s to whether or not SBT based exercise protocols may be useful in ameliorating PD gait deficits and providing further insight into the neurological dysfunction of PD and its effect on locomotor performance.

PAGE 54

54 Figure 2 1. A generalized diagram of basic circuitry of the basal ganglia and related structures .. Black arrows represent inhibitory GABAergic transmissions, white arrows represent excitatory glutamatergic (or in the case of PPN, cholinergic) transmissions, and gray arrows represent dopaminergic tra nsmissions. Gray boxes represent nuclei intrinsic to the basal ganglia. The striatum is composed of the putamen and the caudate nucleus. Abbreviations: SNpc, substantia nigra pars compacta; SNpr, substantia nigra pars reticulata; GPe, globus pallidus exter nus; GPi, globus pallidus internus; STN, subthalamic nucleus; PPN, pedunculopontine nucleus.

PAGE 55

55 Figure 2 2. A Bertec instrumented split belt treadmill (Bertec Corporation, Columbus, OH)

PAGE 56

56 CHAPTER 3 METHODS Participants T en persons with PD w ere recruited for the study (Table 1) Participants had neither walked on a SBT nor experienced any lower extremity orthopedic injury for at least one year prior to participation. All were being treated with stable doses of orally administered levodopa ther apy. Four participants were also taking a dopamine agonist. All participants provide d written informed consent before participating in the study as approved by the University Institutional Review Board. The inclusion criteria include: 1. Idiopathic PD diagn osed by a movement disorders specialist. The initial diagnosis was based on the presence of at least two of three cardinal motor signs of PD: bradykinesia, resting tremor and rigidity, and a demonstrated good response to levodopa medication therapy. A good response to levodopa was defined as a 30% improvement in parkinsonian motor signs. This motor score improvement was (UPDRS) motor examination subscore, following the administration of levodopa (1.5 tim es their typical dose) during their screening neurological examination. Such an inclusion criteria was necessary to exclude patients with Parkinson's plus syndromes (such as progressive supranuclear palsy, multiple system atrophy, striato nigral degenerati on, corticobasal degeneration, and Lewy body disease). 2. Hoehn and Yahr (H&Y) stage II or worse when in the off medication state. 3. Age between 40 85 years. Exclusion criteria for all participants include: 1. Failure to meet the inclusion criteria. 2. Loss of visi on, peripheral neuropathy, vestibular dysfunction, or those taking medications affecting balance or alertness /attention. 3. Presence of active unstable medical or psychiatric conditions, diabetes, or any orthopedic condition that would preclude their ability to participate in the exercises. 4. Presence of active or unstable/untreated cardiovascular disease.

PAGE 57

57 5. Presence of any recent changes in mental and/or physical condition that might affect gait and balance Experimental P rotocol Participants perform ed the same tasks on two separate days of testing no longer than one week apart. During the first visit, the participants perform ed the tasks either while in their optimally medicated state ( ON meds) or after having withdrawn from taking any anti parkinsonian medicat ion for at least twelve hours ( OFF meds). During the subsequent visit, the participants perform ed the same tasks again in the opposite medicated state. The order in which the participants performed the tasks in the on meds and off meds states was randomi zed such that five participants were on meds during the first visit and five participants were off meds during the first visit Prior to testing on b oth days, the participants complete d the motor portion (Section III) of the ting Scale (UPDRS) while being video recor ded. These videos were scored by a single independent movement dis orders trained neurologist who wa s blinded to the medicated state of each participant. During the testing sessions, thirty five pass ive retroreflec tive markers were attached to the body in accordance with the Vicon Plug in Gait full body marker system. Kinematic data were collected using a 7 camera motion capture system (120 Hz; Vicon, Oxford, UK) while participants walk ed on an instrumented SBT (Be rtec Corporation, Colu mbus, OH). Kinetic data were collected at 1200 Hz. Participants first walk ed on the SBT while both belts move d together at a self selected comfortable speed for five minutes to accommodate to walking on the treadmill. All participants held onto the handrails for the duration of all treadmill sessions. Participants were then accommodated to the slow and fast walk ing speeds as they then walk ed for two

PAGE 58

58 BASELINE), followed by two minutes with both belts moving togethe ), and then again for two minutes with both belts m oving at the slow speed as a washout period (Reisman et al. 2005). After the washout period, the treadmill was brought to a stop and then sped up such t reported less affected leg moved at .50 m/s while the belt u nder the more affected leg mo ved at 1.0 m/s. These speeds were selected as we aimed to select a speed which could be tolerated by persons with PD while unmedicated and this specific speed combination has previously been shown to be tolerable by persons po st stroke with relatively severe gait impairment (Reisman et al. 2007). The participants walk ed under these conditions for ten minutes This was considered the adaptation portion of the experiment (ADAPT). Then the belt moving at 1.0 m/s w as slowed to .50 m/s for five minutes to assess the aftereffects of the SBT gait parameters stored during de adaptation to conventional treadmill walking (POST TIED ). Following the five minute POST TIED condition, the participants again walk ed on the SBT with the bel ts in the same 1.0/.50 m/s configuration as performed during ADAPT for five more minutes This was considered the re adaptation portion of the experiment to assess savings (RE ADAPT). Finally, the participants were given a five minute cool down with both belts moving at .50 m/s for five minutes to washout the adapted gait parameters before leaving the lab Th e same protocol w as performed by each participant in the opposite medicated state on the second visit (Figure 3 1 ) Data P rocessing Heel strikes and toe offs were manually labeled in Vicon software based on marker velocity profiles. Marker data were filtered using a 4 th order low pass

PAGE 59

59 Butterworth filter with a cutoff frequency of 10 Hz. Stride length was defined as the distance trave led by the ankle marker along the walking axis from heel strike to toe off (Reisman et al. 2005). Stance time was defined as the percent of the gait cycle between heel strike and subsequent toe off of the same limb. Step length wa s defined as the distan ce between the ankle markers along the walking axis at heel strike. There are two double limb support periods during each gait cycle, defined as 1) slow double limb support, or the time from fast leg heel strike to slow leg toe off and 2) fast double limb support, or the time from slow leg heel strike to fast leg toe off. Sagittal joint angles and joint moments were calculated using the Vicon Plug in Gait model. Sagittal ankle power w as calculated as the product of the ankle sagittal angular velo city and moment. The braking and propulsive phases of each gait cycle were determined by the portions of the gait cycle during which the AP GRFs were negative and positive, respectively. All AP GRFs, joint moments, and ankle powers were normalized in mag cycle. We then calculated the AP GRF braking and propulsive impulses (the areas under the braking and propulsive portions of the AP GRF curve s respectively) as well as the sa gittal joint moment braking and propulsive impulses (the areas under the braking and propulsive portions of the sagittal joint moment curves, respectively). An example as to how the AP GRFs were partitioned into braking and propulsive phases is shown on F igure 3 2 Propulsive impulses were partitioned into the portions of these impulses occurring prior to double limb support (i.e. the portion of the propulsive impulse that could contribute to contralateral step length) and after double limb support (i.e. the portion of the propulsive impulse contributing to body weight transfer). The A2

PAGE 60

60 phase of the ankle power profile (Judge Davis, and Ounpuu 1996) was designated as the portion of the power curves during which power moved from a negative value (i.e. po wer absorption) to a positive value (i.e. power generation) during late stance. An example as to how the A2 portion of the ankle power curve was defined is shown in Figure 3 3. This phase was selected because of its remarkably strong association with ste p length (Judge, Davis, and Ounpuu, 1996). We also calculated the work done by the ankle during A2 by integrating this area of the ankle joint power curve. We define d symmetry in each spatiotemporal gait parameter using the following asymmetry index: Asym metry = (fast leg parameter slow leg parameter)/(fast leg parameter + slow leg parameter) Spatiotemporal asymmetry data and kinetic data w ere averaged over the first and last five strides of ADAPT (EARLY and LATE, respectively to assess adaptation ) as well as over the first five strides of POST TIED (to assess aftereffects during de adaptation ) and RE ADAPT (to assess savings during re adaptation ) Baseline values were determined by averaging data across the first 30 seconds of BASELINE Statistica l A nalyses Four MANOVAs were performed to analyze differences in the spatiotemporal and kinetic gait parameters among various walking condition s and between medicated states L evels of significance all analyses were < .05. Locomotor A daptation A 3x2 ( walking condition x medicated state ) MANOVA was performed to analyze the spatiotemporal and kinetic gait variables among walking condition s involved in

PAGE 61

61 locomotor adaptation (BASELINE, EARLY, LATE) and between medicated states (ON meds, OFF meds). Lo comotor A daptive L earning A 2x2 ( walking condition x medicated state) MANOVA was performed to analyze the effect of medication on the storage of spatiotemporal and kinetic gait aftereffects during conventional treadmill walking immediately following SBT by comparing between walking condition s (BASELINE, POST TIED) and between medicated states (ON meds, OFF meds) Two further 2x2 MANOVAs were performed to analyze the effect of medication on the savings of the spatiotemporal and kinetic gait variables during re adaptation after washout of the first SBT exposure. One of these 2x2 MANOVA s was performed to analyze differences in the gait parameters during the first five strides of the first and second exposures to the SBT gait pattern by comparing between walking condition s ( EARLY, RE ADAPT) and between medicated states (ON meds, OFF meds) The other 2x2 MANOVA was performed to analyze differences in gait parameters between the last five strides of the first exposure and the first five strides of the second exposure by comparing between walking condition s (LATE, RE ADAPT) and between medicated states (ON meds, OFF meds). We also performed linear regressions to assess step length asymmetry values during BASELINE, EARLY, and LATE as predictors of step length asymmetry during POST TIED in both the ON meds and OFF meds states Paired t tests were performed to compare the step lengths of the fast and slow limbs during BASELINE and POST TIED. Further paired t tests were performed to compare lower extremity sagittal joint angles between limbs at heel strikes during POST TIED.

PAGE 62

62 Figure 3 1 An outline of the protocol used to investigate locomotor adaptation and adaptive learning. Solid black horizont al lines indicate that the belts were moving at the same speed. The solid horizontal double lines represent the more affected limb while the dashed horizontal double lines represent the less affected limb. White boxes indicate the time intervals when kin ematic and kinetic data wer 0.50 m/s walking speed.

PAGE 63

63 Figure 3 2. Example of AP GRF curve (during conventional walking) partitioned into braking (black) and propulsive (gray) phases (Specific Aim 2). The braking and propulsive impulses were calculated as the area under the black and gray portions of the curve, respectively. Figure 3 3. Example of a sagittal ankle power curve (during conventional walking Specific Aim 2). Peak A2 ankle power was calculated as the peak of the gray shaded portion of the curve. A2 ankle work was calculated as the area of the gray shaded portion of the c urve.

PAGE 64

64 Table 3 1. Participant characteristics and demographic information. Participant Sex Age (yrs) Height (cm) Mass (kg) Disease duration (months) OFF meds UPDRS motor score ON meds UPDRS motor score OFF meds H&Y ON meds H&Y 1 M 65 177.2 64.1 60 40 38 2.5 2.5 2 F 71 154.9 79.5 12 46 41 3 3 3 M 69 174.0 76.9 72 39 42 2.5 2.5 4 M 76 190.4 74.4 N/A 58 46 3 3 5 M 65 164.5 73.4 48 40 39 2.5 2 6 M 66 175.9 67.3 60 45 39 2.5 2 7 F 49 172.7 95.9 60 24 23 2 2 8 M 72 170.2 72.0 120 41 37 2.5 2.5 9 M 70 165.1 70.5 108 32 34 3 2.5 10 M 65 180.3 84.3 48 31 28 2.5 2.5 Mean 66.8 172.5 75.8 65.3 39.6 36.7 2.6 2.5 (SD) 7.3 9.7 9.1 32.4 9.3 6.8 0.3 0.4 Disease duration indicates time since initial diagnosis. UPDRS Unified Disease Rating Scale, H&Y Hoehn & Yahr stage.

PAGE 65

65 CHAPTER 4 RESULTS Spatiotemporal G ait V ariables d uring L ocomotor A daptation Asymmetry in stride length, stance time, step length, and double l imb support time is shown across all walking condit ions in both medicat ion conditions ( Figures 4 1 through 4 4 ) Comparisons among walking condition s relevant to locomotor adaptation (BASELINE vs. EARLY vs. LATE) and between medicated (ON vs. OFF) states as well as interactions between walking condition s and medicated states will be further outlined and discussed in the following sections. Reactively controlled I ntralimb P arameters d uring L ocomotor A daptation All participants were able to successfully adapt to the SBT walking task regardless of medication status. They were also able to maintain a 1:1 cadence between the limbs across all walking condition s (i.e. never did any participant step twice with the same limb before taking a subsequent step with the contralateral limb). W e observed a significant main effect of walking condition on stride length asymmetry and stanc e time asymmetry during locomotor adaptation Stride length asymmetry S tride length asymmetry significantly increased immediately from BASELINE to EARLY adaptation (p<.05, Figure 4 5 ). The participants also significantly increased stride length asymmetry further from EARLY to LATE (p<.05). We did not observe a significant main effect of medication on stride length asymmetry nor did we observe a significant walking condition x medication interaction for stride length asymmetry.

PAGE 66

66 Stance time asymmetry Stance time asymmetry also significantly increased immediately from BASELINE to EARLY (p<.05, Figure 4 6 ). Stance time asymmetry did not change significantly from EARLY to LATE. We did not observe a significant main effect of medication on stance time asymmetry nor did we observe a significant walking condition x medication interaction for stance time asymmetry. Predictively controlled I nterlimb P arameters d uring L ocomotor A daptation W e also observed a main effect of walking condition on step length asymmetry and double limb s upport time asymmetry during locomotor adaptation (p<.05). Step length asymmetry Step length asymmetry significantly increased from BASELINE t o EARLY (p<.05 Figure 4 7 ). By LATE, step length symmetry was restored to values not significantly different from BASELINE. We did not observe a main effect of medication on step length asymmetry nor did we observe a significant walking condition x medi cation interaction Double limb support time asymmetry Similar to step length asymmetry, double limb support time asymmetry significantly increased from BASELINE to EARLY (p<.05, Figure 4 8 ). By LATE, double limb support time symmetry was restored to valu es not significantly different from BASELINE. We did not observe a main effect of medication on double limb support time asymmetry nor did we observe a significant walking condition x medication interaction In summary, we observed patterns of spatiotempo ral gait asymmetry during locomotor adaptation (i.e. from BASELINE to EARLY to LATE during initial exposure to

PAGE 67

67 the SBT walking task) which were similar to those previously reported by Reisman and colleagues in healthy adults and adults post stroke ( Reisman et al., 2005 ; Reisman et al., 2007 ). Once the belt speeds became asymmetric, t he reactively controlled intralimb parameters ( e.g., stride length and stance time) became asymmetric immediately and remained asymmetric throughout locomoto r adaptation. The predictively controlled interlimb parameters ( e.g., step length and double limb support time) were initially asymmetric during EARLY but gradually adapted such that, by LATE, asymmetry in these parameters was similar to BASELINE. Consistent with Central H ypothesis 1, w e did not observe any effect of dopaminergic medication on the abilities to adapt reactively controlled or predictively controlled parameters d uring locomotor ad aptation Kinetic G ait V ariables d uring L ocomotor A daptation Ensemble AP GRF curves (mean standard deviation) for BASELINE (OFF meds Figure 4 9 O N meds Figure 4 1 0 ) and LATE (OFF meds Figure 4 1 1 ON meds Figure 4 12 ) are shown to indicate how t he AP GRF profiles change from conventional wa lking to SBT walking Similar ly, ensemble ankle power curves are also shown (mean standard deviation) for BASELINE (OFF meds Figure 4 1 3 ON meds Figure 4 1 4 ) and LATE (OFF meds Figure 4 1 5 ON meds Figure 4 1 6 ). Mean AP GRF impulses as well as mean sagittal peak A2 ankle p owers and mean A2 ankle work values are shown during all walking condition s in both medicated states in Figures 4 1 7 through 4 2 2 C omparisons among walking condition s relevant to locomotor adaptation (BASELINE vs. EARLY vs. LATE) and between medicat ed st ates as well as interactions between walking condition s and medicated states will be further outlined and discussed in the following sections.

PAGE 68

68 AP GRF impulses We observed significant main effects of walking condition on fast limb braking impulse, fast limb propulsive impulse, fast limb propulsive impulse prior to double limb support, fast limb propulsive impulse after double limb support, slow braking impulse, and slow propulsive impulse after double limb support (all p<.05). We did not observe significant main effects of walking condition on either slow propulsive impulse or slow propulsive impulse prior to double support. The fast limb braking impulses significantly increased (i.e. became more negative) from BASELINE to EARLY (p<.05 Figure 4 2 3 ) and further increased fr om EARLY to LATE (p<.05) The fast limb propulsive impulses also increased significantly during the locomotor adaptation phase, but only from BAS E LINE to LATE (p<.05, Figure 4 2 4 ). The portion of the fast limb propulsive impulse which occurred prior to double support was not significantly different between BASELINE and EARLY or LATE (Figure 4 2 5 ). However, the portion of the fast limb propulsive impulse which occurred later during double support significantly increased from BASELINE t o LATE (p<.05, Figure 4 2 6 ) but not from BASELINE to EARLY (p=.083). The slow limb braking impulses significantly increased (i.e. became more negative) from BASE L INE to EARLY (p<. 05, Figure 4 2 3 ) but then significantly decreased from EARLY to LATE (p<.05) such that, during LATE, they were not significantly different from BASELINE. We did not observe a main effect of walking condition on the slow limb propulsive impulses (Figure 4 2 4 ) or the portion of the slow limb propulsive impulses prior to double suppo rt (Figure 4 2 5 ). However, the portion of the slow limb propulsive impulses during double support significantly decreased from BASELINE to EARLY (p<.05, Figure 4 2 6 ) and remained similar from EARLY to LATE.

PAGE 69

69 Sagittal ankle kinetics We observed significant main effects of walking condition on fast limb peak A2 power, fast limb A2 work, and slow limb peak A2 power (all p<.05). We did not observe a significant main effect of walking condition on slow limb A2 work. The fast limb peak A2 power significantly i ncreased from BASELINE to EARLY (p<.05, Figure 4 2 7 ) and remained similar in magnitude from EARLY to LATE. Similarly, A2 ankle work increased in a similar pattern from BASELINE to EARLY (p<.05, Figure 4 28 ) but not from EARLY to LATE. We did not observe any differences between BASELINE, EARLY, and LATE in slow limb peak A2 power (Figure 4 2 7 ) or A2 ankle work (Figure 4 28 ). In summary, many of the AP GRF and sagittal ankle kinetic measures changed asymmetrically from BASELINE to EARLY and/or LATE particularly within the fast limb Braking impulses became significantly larger during EARLY in both the fast and slow limbs (relative to BASELINE), though only in the slow limb did the braking impulses return to values similar to BASELINE by LATE whereas those in the fast limb remained significantly higher Propulsive impulses in the fast limb also increased significantly from BASELINE to LATE, due primarily to increased force production during the double limb support portion of the propulsive ph ase. On the contrary, force production during the double limb support portion of the propulsive phase decreased from BASELINE to EARLY in the slow limb. Fast limb peak A2 ankle power and A2 ankle work also increased immediately upon exposure to the SBT t ask during EARLY as compared to BASELINE but did not change from EARLY to LATE. Slow limb peak A2 ankle power and A2 ankle work did not change from BASELINE to EARLY or LATE. Consistent with Central Hypothesis 2, w e did not observe any effect of dopamine rgic medication on the

PAGE 70

70 abilities to adapt gait kinetics during the locomotor adaptation phase of the experiment. Thus, when considered with the spatiotemporal results reported above, the effect of dopamine on the ability to adapt gait appears to be minima l. Spatiotemporal G ait V ariables d uring L ocomotor A daptive L earning Comparisons of stride length asymmetry, stance time asymmetry, step length asymmetry, and double limb support time asymmetry among walking condition s relevant to two phases of locomotor ad aptive learning storage of aftereffects during de adaptation (BASELINE vs. POST TIED) and savings during re adaptation (EARLY vs. RE ADAPT and LATE vs. RE ADAPT) and between medicated states as well as interactions between walking condition s and medica ted states will be further outlined and discussed in the following sections. Reactively controlled I ntralimb P arameters d uring D e adaptation Stride length asymmetry Stride length asymmetry was significantly higher during POST TIED than during BASELINE (p< .05 Figure 4 29 ). However, we did not observe a significant main effect of medication on stride length asymmetry nor did we observe a significant walking condition x medication interaction Stance time asymmetry Similar to stride length asymmetry, stance time asymmetry was significantly higher during POST TIED than during BASELINE (p<.05, Figure 4 3 0 ) We also did not observe a significant main effect of medication on stride length asymmetry nor did we observe a significant walking condition x medication interaction.

PAGE 71

71 Predictively controlled I nterlimb P arameters d uring D e adaptation Step length asymmetry Step length asymmetry was significantly higher during POST TIED than during BASELINE (p<.05 Figure 4 3 1 ). We also observed a significant main effect of medication as well as a significant walking condition x medication interaction for step length asymmetry (both p<.05) indicating that the change in step length asymmetry from BASELINE to POST TIED was significantly larger in the ON meds state as compared to the OFF meds state. Double limb support time asymmetry Double limb support time asymmetry was also significantly higher during POST TIED than during BASELINE (p<.05 Figure 4 3 2 ). However, unlike step length asymmetry, we did observed neither a signifi cant main effect of medication nor a significant walking condition x medication interaction for double limb support time asymmetry. Associations b etween S tep L ength A symmetry d uring L ocomotor A daptation and D e adaptation In the OFF meds state, we observed that step length asymmetry during LATE was a significant predictor of step length asymmetry during POST =.091, p= .745) and EARLY = .397, p =.083) were not (Figures 4 3 3 and 4 3 4 ) In the ON meds state, step length .258, p=.222), and LATE TIED (Figures 4 3 5 and 4 3 6 )

PAGE 72

72 Orientation of the Limbs at H e el strikes d uring D e adaptation Ensemble orientation s of the limbs at heel strike are sh own in the OFF meds (Figure 4 3 7 ) and ON meds (Figure 4 3 8 ) states. During POST TIED, the participants exhibited significantly greater hip flexion (p<.05) and a trend toward significantly greater knee extension (p=.074) in the fast limb at fast limb heel strike when ON meds c ompared to OFF meds (Figure 4 39 ). We also observed significantly greater knee extension in the fast limb at slow limb heel strike when the partic ipants were ON meds (Figure 4 39 ) While the slow limb took a significantly shorter step during POST TIED as compared to BASELINE in both the ON meds and OFF meds states, the fast limb only took a significantly longer step during POST TIED as compared to BASELINE when ON meds (p<.05, Figure 4 4 0 ). In summary, consistent with Central Hypothesis 1, dopamine significantly increased the ability of the persons with PD to store the step length aftereffect during conventional walking immediately following SBT wa lking (i.e. de adaptation) Persons with PD demonstrated greater hip flexion and knee extension at fast limb heel strike during POST TIED when ON meds, which led to significantly greater fast limb step lengths and, ultimately, greater step length asymmetr y We did not observe a significant effect of medication on aftereffects of a predictively controlled temporal parameter, double limb support time asymmetry. Further, dopamine did not affect the ability of persons with PD to de adapt reactively controlle d parameters when de adapting from SBT walking to conventional walking.

PAGE 73

73 Reactively controlled I ntralimb P arameters d uring S avings Stride length asymmetry Stride length asymmetry was significantly higher during RE ADAPT than during EARLY and significantly lower during RE ADAPT than during LATE ( both p<.05 Figure 4 4 1 ). However, we did not observe a significant main effect of medication on stride length asymmetry nor did we observe a significant walking condition x medication interaction. Stance time asymm etry Similar to stride length asymmetry, stance time asymmetry was significantly higher during RE ADAPT than during EARLY (p<.05 Figure 4 4 2 ). We did not observe a significant difference between stance time asymmetry during RE ADAPT and LATE (p=.081) W e also did not observe a significant main effect of medication on stride length asymmetry nor did we observe a significant walking condition x medication interaction. Predictively controlled I nterlimb P arameters d uring S avings Step length asymmetry Step le ngth asymmetry was significantly lower during RE ADAPT than during EARLY and significantly higher during RE ADAPT than during LATE (both p<.05 Figure 4 4 3 ) We also did not observe a significant main effect of medication on step length asymmetry nor did we observe a significant walking condition x medication interaction. Double limb support time asymmetry Similar to step length asymmetry, double limb support time asymmetry was significantly lower during RE ADAPT than during EARLY and significantly higher during RE ADAPT than during LATE (both p<.05 Figure 4 4 4 ). We also did not observe a

PAGE 74

74 significant main effect of medication on double limb support time asymmetry nor did we observe a significant walking condition x medication interaction. In summary, cont rary to Central Hypothesis 1, we did not observe any effect of dopaminer gic medication on savings of the adapted locomotor pattern Savings were apparent in the persons with PD as the interlimb asymmetry (in step length and double limb support time) was s ignificantly lower during RE ADAPT as compared to EARLY, indicating that the adapted gait pattern was more quickly achieved during the second exposure to SBT walking. However, savings were not complete as we observed significantly increased asymmetry during RE ADAPT as compared to LATE. We did not observe signi ficant effects of medication on savings in e ither the EARLY vs. RE ADAPT comparisons or the LATE vs. RE ADAPT comparisons. Kinetic G ait V ariables d uring L ocomotor A daptive L earning Comparisons of AP GRF impulses, sagittal peak A2 ankle powers, and sagittal A2 ankle work production among walking cond ition s relevant to two phases of locomotor adaptive learning de adaptation (BASELINE vs. POST TIED) and savings (EARLY vs. RE ADAPT and LATE vs. RE ADAPT) and between medicated states as well as interactions between walking condition s and medicated sta tes will be further outlined and discussed in the following sections. We postulated that kinetic aftereffects during de adaptation may underlie the step length aftereffects observed during POST TIED, and thus may also be affected by dopaminergic medicatio n. Ensemble AP GRF impulse curves and sagittal ankle power curves during POST TIED in both the ON meds and OFF meds are shown in Figures 4 4 5 through 4 4 8

PAGE 75

75 AP GRF I mpulses d uring D e adaptation We observed neither significant main effects of walking condit ion or medication nor any significant walking condition x medication interactions for any of the AP GRF impulse variables (Figure 4 49 through 4 5 2 ) AP GRF impulses were not significantly different between BASELINE and POST TIED nor did medication have a ny effect on AP GRF production during these walking condition s. Sagittal A nkle K inetics d uring D e adaptation Similar to the AP GRF impulses during de adaptation we observed neither significant main effects of walking condition or medication nor any signif icant walking condition x medication interactions for peak A2 power or A2 work production (Figures 4 5 3 and 4 5 4 ) Thus, sagittal ankle kinetics were not significantly different between BASELINE and POST TIED nor did medication have any effect on the sagittal ankle kinetics during these walking condition s. In summary, contrary to Central Hypothesis 2 the kinetics w hich have been previously shown to be strongly associated with changes in step length during conventional gait were not significantly different during BASELINE and POST TIED. Both AP GRFs and sagittal ankle kinetics were similar during conventional walkin g both before and immediately following SBT walking. Thus, dopaminergic medication enhances the ability to store greater step length aftereffects during POST TIED but these asymmetric step length patterns do not result from changes in AP GRFs or sagittal ankle kinetics. AP GRF I mpulses d uring S avings Similar to the findings we observed when comparing AP GRF impulses during de adaptation (i.e. BASELINE vs. POST TIED), we observed neither significant main

PAGE 76

76 effects of walking condition or medication nor signif icant walking condition x medication interactions on AP GRF impulses when comparing both EARLY vs. RE ADAPT as well as LATE vs. RE ADAPT (Figure 4 5 5 through 4 58 ) Thus, AP GRF impulses during RE ADAPT were not significantly different from those observed during EARLY and LATE. Further, medication did not significantly affect AP GRF production during RE ADAPT. Sagittal A nkle K inetics d uring S avings We also observed neither significant main effects of walking condition or medication nor significant walking condition x medication interactions on peak A2 power or A2 work production (Figures 4 5 9 and 4 6 0 ) Thus, sagittal ankle kinetics during RE ADAPT were not significantly different from those observed during EARLY and LATE. Further, medication did not sig nificantly affect AP GRF production during RE ADAPT. In summary, contrary to Central Hypothesis 2, we did not observe any significant effect of dopaminergic medication on the savings of AP GRFs or sagittal ankle kinetics. Persons with PD were able to re a dapt to the SBT gait pattern immediately upon a second exposure similarly regardless as to whether they were ON meds or OFF meds. Thus, dopamine does not appear to affect savings of adapted spatiotemporal gait characteristics or force production during ga it.

PAGE 77

77 Figure 4 1. Mean s tride length asymmetry during all walking conditions ( Specific Aim 1) We observed a significant main effect of walking condition but no significant differences between medicated states. Error bars indicate standard error.

PAGE 78

78 Figure 4 2. Mean s tance time asymmetry during al l walking conditions ( Specific Aim 1 ) We observed a significant main effect of walking condition but no significant differences between medicated states. Error bars indicate standard error.

PAGE 79

79 Figure 4 3. Mean s tep length asymmetry during al l walking conditions ( Specific Aim 1 ) We observed a significant main effect of walking condition as well as a significant decrease in step length asymmetry during POST TIED in the OFF meds state as co mpared to ON meds. Error bars indicate standard error.

PAGE 80

80 Figure 4 4. Mean d ouble limb support time asymmetry during all walking conditions ( Specific Aim 1 ) We observed a significant main effect of walking condition but no significant differences betwe en medicated states. Error bars indicate standard error.

PAGE 81

81 Figure 4 5. Mean stride length asymmetry during locomotor adaptation (Specific Aim 1). We observed a significant main effect of walking condition but no significant differences between medicated states. Error bars indicate standard error.

PAGE 82

82 Figure 4 6 Mean stance time asymmetry during locomotor adaptation (Specific Aim 1). We observed a significant main effect of walking condition but no sign ificant differences between medicated states. Error bars indicate standard error.

PAGE 83

83 Figure 4 7 Mean step length asymmetry during locomotor adaptation (Specific Aim 1). We observed a significant main effect of walking condition but no significant diff erences between medicated states. Error bars indicate standard error.

PAGE 84

84 Figure 4 8 Mean double limb support time asymmetry during locomotor adaptation (Specific Aim 1). We observed a significant main effect of walking condition but no significant dif ferences between medicated states. Error bars indicate standard error.

PAGE 85

85 Figure 4 9 Ensemble AP GRF curves during BASELINE OFF meds (Specific Aim 2). Dark solid and dashed lines indicate the mean of the fast and slow limbs, respectively, across all participants. Shaded areas indicate standard deviation. Figure 4 10 Ensemble AP GRF curves during BASELINE ON meds (Specific Aim 2). Dark solid and dashed lines indicate the mean of the fast and slow limbs, respectively, across all parti cipants. Shaded areas indicate standard deviation.

PAGE 86

86 Figure 4 11 Ensemble AP GRF curves during LATE OFF meds (Specific Aim 2). Dark solid and dashed lines indicate the mean of the fast and slow limbs, respectively, across all participants. Shaded are as indicate standard deviation. Figure 4 1 2 Ensemble AP GRF curves during LATE ON meds (Specific Aim 2). Dark solid and dashed lines indicate the mean of the fast and slow limbs, respectively, across all participants. Shaded areas indicate standard deviation.

PAGE 87

87 Figure 4 1 3 Ensemble sagittal ankle power curves during BASELINE OFF meds (Specific Aim 2). Dark solid and dashed lines indicate the mean of the fast and slow limbs, respectively, across all participants. Shaded areas indicate standard d eviation. Figure 4 14 Ensemble sagittal ankle power curves during BASELINE ON meds (Specific Aim 2). Dark solid and dashed lines indicate the mean of the fast and slow limbs, respectively, across all participants. Shaded areas indicate standard devi ation.

PAGE 88

88 Figure 4 15 Ensemble sagittal ankle power curves during LATE OFF meds (Specific Aim 2). Dark solid and dashed lines indicate the mean of the fast and slow limbs, respectively, across all participants. Shaded areas indicate standard deviation. Figure 4 16 Ensemble sagittal ankle power curves during LATE ON meds (Specific Aim 2). Dark solid and dashed lines indicate the mean of the fast and slow limbs, respectively, across all participants. Shaded areas indicate standard deviation.

PAGE 89

89 Fig ure 4 1 7 Mean braking AP GRF impulses during all walking conditions (Specific Aim 2). We observed significant main effects of walking condition on both the fast and slow braking impulses but no significant differences between medicated states. Error ba rs indicate standard error.

PAGE 90

90 Figure 4 18 Mean propulsive AP GRF impulses during all walking conditions (Specific Aim 2). We observed significant main effects of walking condition on both the fast and slow propulsive impulses but no significant differ ences between medicated states. Error bars indicate standard error.

PAGE 91

91 Figure 4 19 Mean propulsive AP GRF impulses of the portion of the propulsive GRF prior to double limb support during all walking conditions (Specific Aim 2). We observed a significa nt main effect of walking condition on the fast propulsive impulses prior to double support but neither a significant effect of walking condition on the slow propulsive impulses prior to double support nor significant differences between medicated states. Error bars indicate standard error.

PAGE 92

92 Figure 4 20 Mean propulsive AP GRF impulses of the portion of the propulsive GRF during double support during all walking conditions (Specific Aim 2). We observed significant main effects of walking condition on b oth the fast and slow propulsive impulses during double support but no significant differences between medicated states. Error bars indicate standard error.

PAGE 93

93 Fi gure 4 21 Peak A2 ankle power during all walking conditions (Specific Aim 2). We observed significant main effects of walking condition on both slow and fast peak A2 ankle power but no significant differences between medicated states. Error bars indicate standard error.

PAGE 94

94 Figure 4 2 2 A2 ankle work during all walking condi tions (Specific Aim 2). We observed a significant main effect of walking condition on fast A2 ankle work but found neither a significant main effect of walking condition on slow A2 ankle work nor significant differences between medicated states. Error ba rs indicate standard error.

PAGE 95

95 Figure 4 23 Mean braking AP GRF impulses during locomotor adaptation (Specific Aim 2). We observed a significant main effect of walking condition but no significant differences between medicated states. Error bars indicate standard error.

PAGE 96

96 Figure 4 2 4 Mean propulsive AP GRF impulses during locomotor adaptation (Specific Aim 2). We observed a significant main effect of walking condition but no significant differences between medicated states. Error bars indicate standard error.

PAGE 97

97 Figure 4 2 5 Mean propulsive AP GRF impulses of the portion of the propulsive GRF prior to double limb support during locomotor adaptation (Specific Aim 2). We observed a significant main effect of walking condition on the fast propulsive impulses prior to double s upport but no significant differences between medicated states. Error bars indicate standard error.

PAGE 98

98 Figure 4 2 6 Mean propulsive AP GRF impulses of the portion of the propulsive GRF during double limb support during locomotor adaptat ion (Specific Aim 2). We observed a significant main effect of walking condition on the fast propulsive impulses prior to double support but no significant differences between medicated states. Error bars indicate standard error.

PAGE 99

99 Figure 4 2 7 Peak A2 ankle power during locomotor adaptation (Specific Aim 2). We observed significant main effects of walking condition on fast peak A2 ankle power but no significant differences between medicated states. Error bars indicate standard error.

PAGE 100

100 Figure 4 28 A2 ankle work during locomotor adaptation (Specific Aim 2). We observed significant main effects of walking condition on fast peak A2 ankle work but no significant differences between medicated states. Error bars indicate standard error.

PAGE 101

101 Figure 4 29 Mean stride length asymmetry aftereffects during de adaptation (Specific Aim 1). We observed a significant main effect of walking condition but no significant differences between medicated states. Error bars indicate standard error.

PAGE 102

102 Figure 4 3 0 Mean stance time asymmetry aftereffects during de adaptation (Specific Aim 1). We observed a significant main effect of walking condition but no significant differences between medicated states. Error bars indicate standard error.

PAGE 103

103 Figure 4 3 1 Mean step length asymmetry aftereffects during de adaptation (Specific Aim 1). We observed a significant main effect of walking condition a significant main effect of medication, and a significant walking condition x medication interaction. indicat es p<.05. Error bars indicate standard error.

PAGE 104

104 Figure 4 3 2 Mean double limb support time asymmetry aftereffects during de adaptation (Specific Aim 1). We observed a significant main effect of walking condition but no significant differences between medicated states. Error bars indicate standard error.

PAGE 105

105 Figure 4 3 3 Association between EARLY step length asymmetry and POST TIED step length asymmetry in the OFF meds state (Specific Aim 1) We did not ob serve a significant association nor was EARLY step length asymmetry a significant predictor of POST TIED step length asymmetry (OFF meds) when included in a linear regression model. Figure 4 3 4 Association between LATE step length asymmetry and POST TIED step length asymmetry in the OFF meds state (Specific Aim 1) We observed a significant association and LATE step length asymmetry was the lone significant predictor of POST TIED step length asymmetry (OFF meds) when included in a linear regress ion model (p<.05)

PAGE 106

106 Figure 4 3 5 Association between EARLY step length asymmetry and POST TIED step length asymmetry in the ON meds state (Specific Aim 1) We did not observe a significant association nor was EARLY step length asymmetry a significant predictor of POST TIED step length asymmetry (ON meds) when included in a linear regression model. Figure 4 3 6 Association between LATE step length asymme try and POST TIED step length asymmetry in the ON meds state (Specific Aim 1) We did not observe a significant association nor was LATE step length asymmetry a significant predictor of POST TIED step length asymmetry (ON meds) when included in a linear r egression model.

PAGE 107

107 Figure 4 3 7 Ensemble orientation of the limbs at fast limb heel strike (left) and slow limb heel strike (right) during POST TIED while OFF meds (Specific Aim 1). The red limb represents the fast limb while the black limb represents t he slow limb. The relative difference between the anterior posterior positioning of the feet represents the step length asymmetry observed during POST TIED. Figure 4 38 Ensemble orientation of the limbs at fast limb heel strike (left) and slow limb h eel strike (right) during POST TIED while ON meds (Specific Aim 1). The red limb represents the fast limb while the black limb represents the slow limb. The relative difference between the anterior posterior positioning of the feet represents the step le ngth asymmetry observed during POST TIED.

PAGE 108

108 Figure 4 39 Sagittal hip and knee angles during POST TIED in both the ON and OFF meds states at heel strikes (Specific Aim 1) The participants demonstrated significantly more hip flexion in the fast limb at fast limb heel strike and trends toward increased knee extension in the fast limb at both fast limb and slow limb heel strikes when ON meds. indicates p<.05, + indicates p<.08.

PAGE 109

109 Figure 4 4 0 Fast limb and slow limb step lengths during POST TIED in both the O FF (top) and O N (bottom) meds states at heel strikes (Specific Aim 1) The participants demonstrated significantly shorter slow limb step lengths during POST TIED as compared to BASELINE in both the OFF and ON meds states However, only in the ON meds state did the participants exhibit fast limb step lengths during POST TIED which were significantly longer than during BASELINE. indicates p<.05.

PAGE 110

110 Figure 4 4 1 Mean stride length asymmetry savings during re adaptation (Specific Aim 1). We observed a significant main effect of walking condition but no significant differences between medicated states. Error bars indicate standard error.

PAGE 111

111 Figure 4 4 2 Mean stance time asymmetry savings during re adaptation (Specific Aim 1). We observed a significant main effect of walking condition but no significant differences between medicated states. Error bars indicate standard error.

PAGE 112

112 Figure 4 4 3 Mean step length asymmetry savings during re adaptation (Specific Aim 1). We observed a significant main effect of walking condition but no significant differences between medicated states. Error bars indicate standard error.

PAGE 113

113 Fig ure 4 4 4 Mean double limb support time asymmetry savings during re adaptation (Specific Aim 1). We observed a significant main effect of walking condition but no significant differences between medicated states. Error bars indicate standard error.

PAGE 114

114 Figure 4 4 5 Ensemble AP GRF curves during POST TIED OFF meds (Specific Aim 2). Dark solid and dashed lines indicate the mean of the fast and slow limbs, respectively, across all participants. Shaded areas indicate standard deviation. Figure 4 4 6 Ensemble AP GRF curves during POST TIED ON meds (Specific Aim 2). Dark solid and dashed lines indicate the mean of the fast and slow limbs, respectively, across all participants. Shaded areas indicate standard deviation.

PAGE 115

115 Figure 4 4 7 Ensemble sagi ttal ankle power curves during POST TIED OFF meds (Specific Aim 2). Dark solid and dashed lines indicate the mean of the fast and slow limbs, respectively, across all participants. Shaded areas indicate standard deviation. Figure 4 48 Ensemble sagit tal ankle power curves during POST TIED ON meds (Specific Aim 2). Dark solid and dashed lines indicate the mean of the fast and slow limbs, respectively, across all participants. Shaded areas indicate standard deviation.

PAGE 116

116 Figure 4 49 Mean braking AP GRF impulses aftereffects during de adaptation (Specific Aim 2). We observed neither significant effects of walking condition nor medication Error bars indicate standard error.

PAGE 117

117 Figure 4 50 Mean propulsive AP GRF impulses aftereffects of the propulsive GRF during de adaptation (Specific Aim 2). We observed neither significant effects of walking condition nor medication. Error bars indicate standard error.

PAGE 118

118 Figure 4 5 1 Mean propulsive AP GRF impulses aftereffects of the portion of the propulsive GRF prior to double limb support during de adaptation (Specific Aim 2). We observed neither significant effects of walking condition nor medication. Error bars indicate stan dard error.

PAGE 119

119 Figure 4 5 2 Mean propulsive AP GRF impulses aftereffects of the portion of the propulsive GRF during double limb support during de adaptation (Specific Aim 2). We observed neither significant effects of walking condition nor medication. Error bars indicate standard error.

PAGE 120

120 Figure 4 5 3 Peak A2 ankle power aftereffects during de adaptation (Specific Aim 2). We observed neither significant effects of walking condition nor medication. Error bars indicate standard error.

PAGE 121

121 Figure 4 5 4 A2 ankle work aftereffects during de adaptation (Specific Aim 2). We observed neither significant effects of walking condition nor medication. Error bars indicate standard error.

PAGE 122

122 Figure 4 5 5 Mean braking AP GRF impulses savings during re adaptation (Specific Aim 2). We observed neither significant effects of walking condition nor medication when comparing both EARLY and LATE to RE ADAPT Error bars indicate standard error.

PAGE 123

123 Figure 4 5 6 Mean propulsive AP GRF impulses savings during re adaptation (Specific Aim 2). We observed neither significant effects of walking condition nor medication when comparing both EARLY and LATE to RE ADAPT. Error bars indicate standard error.

PAGE 124

124 Figure 4 5 7 Mean propulsive AP GRF impulses savings of the portion of the propulsive GRF prior to double limb support during re adaptation (Specific Aim 2). We observed neither significant effects of walking condition nor medication when comparing both EARLY and L ATE to RE ADAPT. Error bars indicate standard error.

PAGE 125

125 Figure 4 58 Mean propulsive AP GRF impulses savings of the portion of the propulsive GRF during double limb support during re adaptation (Specific Aim 2). We observed neither significant effects of walking condition nor medication when comparing both EARLY and LAT E to RE ADAPT. Error bars indicate standard error.

PAGE 126

126 Figure 4 59 Peak A2 ankle power savings during re adaptation (Specific Aim 2). We observed neither significant effects of walking condition nor medication when comparing both EARLY and LATE to RE ADAPT. Error bars indicate standard error.

PAGE 127

127 Figure 4 6 0 A2 ankle work savings during re adaptation (Specific Aim 2). We observed neither significant effects of walking condit ion nor medication when comparing both EARLY and LATE to RE ADAPT. Error bars indicate standard error.

PAGE 128

128 CHAPTER 5 DISCUSSION T he purposes of th e present study we re to investigate 1) locomotor adaptation and adaptive learning persons with PD adapt gait in a native, unmedicated state and 2) how dopaminergic therapy affect s locomotor adaptation and adaptive learning in persons with PD. When OFF m eds, persons with PD exhibited the ability to adapt gait to SBT walking, stored aftereffects dur ing de adaptation, and demonstrated savings upon re adaptation. W e observed a significant effect of dopamine rgic medication on the ability to store step length aftereffects during de adaptation after SBT walking. Indeed, s tep length asymmetry during POST TIED was diminished when the participants were OFF meds as compared to ON meds. Conversely, we observed no effect of dopamine on the savings of the SBT gait pattern during re adaptation Further, locomotor adaptation was unaffected by levodopa. The ada ptive motor learning perspectives, biomechanics, potential underlying neural mechanisms, and clinical relevance of these findings will be further discussed in the following sections. Effects of L evodopa on L ocomotor A daptive L earning During SBT walking pat terns in which one limb walks faster than the other simultaneously, interlimb gait parameters (step length, double support time) are initially asymmetric (Reisman et al., 2005) Over time, as the central nervous system begins to predict the SBT perturbation and accordingly coordinates the movements of the lower extremities, the interlimb parameters gradually approach symmetry Then, after the belts of the SBT are returned to symmetric speeds, the adapted gait pattern i s temporarily stored as an aftereffect in healthy individuals such that a degree of asymmetry (which is opposite to the initial asymmetry exhibited during the adaptation

PAGE 129

129 task) is demonstrated in the interlimb gait parameters during conventional walking. F inally, the belts can again be split in the same manner as during the initial adaptation task. During this re adaptation period, if savings of the initially adapted gait pattern are present, the interlimb parameters of the SBT gait pattern will be more sy mmetric than during the initial bout of adaptation. If complete savings have occurred, the interlimb parameters will be similar to those observed at the end of the initial adaptation task. Thus, locomotor adaptation is assessed by the change in interlimb parameters during the initial bout of SBT walking while locomotor adaptive learning is assessed through storage of aftereffects and savings The degree of aftereffect storage during de adaptation is quantified by the magnitude of the asymmetry present i n the interlimb gait parameters during conventional treadmill walking (i.e. de adaptation) immediately following the SBT adaptation task (Reisman et al., 2005) Savings is assessed by the amount of asymmetry present in the interlimb gait parameters during initial strides of re adaptation during a second bout of SBT walking following washout of the first bout Thus, larger values of step length or double limb support time asymmetry during POST TIED in the present study indicate d a larger aftereffect of the SBT walking pattern during de adaptation Conver sely, smaller values during RE ADAPT indicate d that savings had occurred as the symmetrical state gradually achieved over time throughout the first adaptation task was partial ly reproduced upon re adaptation to the SBT walking pattern. In the current study, we observed a significant effect of dopamine on the ability to store step length aftereffects from SBT walking to conventional walking such that step length asymmetry during de adaptation was di minished when the participants were OFF meds.

PAGE 130

130 In contrast we observed no effect of dopamine on the savings of the SBT gait pattern during re adaptation The findings indicating reduced aftereffect of step length during POST TIED following SBT walking are consistent with previous research performed in other adaptive motor learning paradigms in persons with PD. For instance, Stern and colleagues (1988) employed a prism reaching task and observed unaffected adaptation pat terns and rates but diminished spatial aftereffect during de adaptation (i.e. the reach of the PD participants was significantly more accurate upon removal of the prism glasses). Similarly, Fernandez Ruiz and colleagues (2003) utilized a prism throwing ad aptation task to investigate visuomotor adaptive learning in persons with different basal ganglia disorders (PD ON meds investigators also observed normal visuomotor adaptation patterns and rates but a reduced aftereff ect in the throwing task upon removal of the prism glasses in both the PD and H s e stud ies w ere limited to examination of visuomotor adaptive learning and thus the investigators suggested that the basal ganglia play an import ant role in the storage of visuomotor aftereffects. Further, t hese stud ies did not assess savings of the visuomotor adaptation task by subjecting the participants to a second bout of prism adaptation. Another study by Isaias and colleagues demonstrated t hat persons with PD exhibit reduced interlimb transfer following a visuomotor adaptation task (2011). Similar to the findings by Fernandez Ruiz and colleagues (2003) and Stern and colleages (1988) these investigators observed that adaptation rates to the visuomotor adaptation task were similar between persons with PD and healthy older adults. However, when performing the same task a

PAGE 131

131 second time with the opposite hand, persons with PD showed diminished improvement in performance compared to the healthy ad ults. Considering our findings as well as the f indings from these previous studies, we suggest that dopaminergic pathways may play a role not only in the storage of visuomotor aftereffects following visuomotor adaptation but also in the storage of step le ngth aftereffects following locomotor ad aptation Thus, dopamine appears to be important to facilitate storage of multiple types of aftereffects during various modes of adaptive motor learning. As basal ganglia dysfunction is intensified after withdrawal of dopaminergic medication in persons with PD, our findings also support the notion of an association between basal ganglia function and storage of adapted motor function upon de adaptation (such as return to conventional walking or throwing a ball in the absence of prism glasses after motor adaptation ) The neural mechanisms potentially underlying these deficits will be further discussed in greater detail in later section within this document. Another interesting finding from this study was the observation that the amount of step length asymmetry during LATE is a significant predictor of step length asymmetry during POST TIED, but only when the persons with PD were OFF meds. This indicates that the degree to which the participants, when OFF meds were able to restore their step lengths to symmetry by the end of locomotor adaptation was significantly associated with the storage of step length aftereffects during de adaptation As step length asymmetry during LATE was not significantly affected by dopaminergic medication, we feel confident that the reduced aftereffect observed OFF meds is not due to incomplete adaptation. In fact, it is perhaps more surprising that step length asymmetry during LATE adaptation was not significantly associated with step length

PAGE 132

132 asymmetry during POST TIED when ON meds. Previous research by Fernandez Ruiz and Diaz (1999) demonstrated that degree of adaptation (in terms of the change in performance from early adaptation to late adaptation) was strongly associated with t he magnitude of the aftereffect observed following a prism throwing task. While it is also possible that there was not enough variability in step length asymmetry between participants in either LATE or POST TIED to observe significant associations between these conditions with only ten participants the standard deviations of step length asymmetry during these conditions were similar between the ON meds and OFF meds states. Thus, the between subject variability appeared to be similar in both states. More over, w hy adaptation does not significantly predict the magnitude of aftereffects during de adaptation while persons with PD are ON meds requires further research. In the present study, eight of the ten participants exhibited greater step length asymmetry during POST TIED while ON meds as compared to OFF meds. It is important to note that the two participants that demonstrated greater step length asymmetry while OFF meds during POST TIED exhibited certain features of either their disease progression or the ir performance on the SBT which may explain these findings. First, one of these participants was the youngest participant included in the current study and presented the leas t severe case of PD (as quantified by motor UPDRS scores) As there is an establ ished learning effect between sessions during locomotor adaptation tasks (Malone et al., 2011), it is likely that the aftereffect diminished in this participant when ON meds due to the fact that the participant was first tested in the OFF meds state. Thus the reduction in step length asymmetry during POST TIED when ON meds may have been due to the learning effect in combin ation with potentially

PAGE 133

133 insignificant basal ganglia dysfunction. Importantly, the four other participants that were tested first in the OFF meds state with relatively more severe PD still exhibited greater step length asymmetry during POST TIED while ON meds. This indicates that the learning effect was not so great as to washout the effects of levodopa, except potentially in our m ildes t case. The second participant exhibiting greater step length asymmetry while OFF meds than ON meds was tested first in the ON meds state and took two fast limb step s which were markedly shorter than the other three among the first five strides of POST TIED seemingly in order to maintain balance upon surprise to the SBT perturbation Removing from the analysis the degree of step length asymmetry during POST TIED would have been greater in the ON meds state than in the OFF meds state in this participant as well. However, these steps were included in the analysis to preserve the integrity of the data. Thus, th e effect of dopamine rgic medication (and potentially basal ganglia dysfunction) on step length aftereffects during conventional following SBT walking appears to be robust. While we observed significantly diminished step length asymmetry during de adaptatio n (i.e. during POST TIED) we did not observe any effect of medication on savings of the SBT gait pattern upon re adaptation to the task (i.e. during RE ADAPT). These findings are in contrast to previous studies of various upper extremity visuomotor tasks which demonstrated significant involvement of the basal ganglia in savings of adapted movements (Marinelli et al., 2009; Bdard and Sanes, 2011; Leow et al., 2012). We suggest that these differences potentially result due to the differing nature and task demands of adaptive locomotor and visuomotor tasks as well as the varying degree of overlearning provided during the initial adaptation task. Previous research

PAGE 134

134 suggests that overlearning during the initial adaptation task increases the amount of savings demonstrated during subsequent re adaptation tasks of similar nature (Joiner and Smith, 2008) During motor adaptation tasks, given sufficient practice, performance will eventually be adapted to a specific asymptote (in locomotor adaptation tasks this as ymptote is baseline symmetry of interlimb spatiotemporal gait parameters) such that performance remains relatively constant and little further adaptation occurs. Overlearning is quantified by the amount of practice given after performance has reached this asymptote (Driskell et al., 1992). In the previous study by Marinelli and colleagues (Marinelli et al., 2009), it appears that overlearning was minimized as performance on the adaptation task only briefly reached asymptotic baseline values. In the study by B dard and Sanes ( Bdard and Sanes, 2011), no overlearning was observed as performance during adaptation never reached asymptote. Only Leow and colleagues (Leow et al., 2012) allowed for overlearning but even then only appeared to allow for 10 15 tria ls of practice once the performance asymptote was reached. Conversely, in the present study, participants walked on the SBT for ten minutes during the adaptation task and had typically restored symmetry in the interlimb gait parameters by the second minut e. Thus, participants typically walked for at least eight minutes once they had reached a baseline level of interlimb (i.e. either step length or double limb the PD part icipants in this study performed roughly 300 state. It should be considered that walking is a relatively automatic task while the visuomotor tasks likely require a greater degree of cognitive and attentional resources.

PAGE 135

135 Acc ordingly, it is likely not sufficient to directly relate performance on a single step during SBT walking to performance of a single discretely performed visuomotor adaptation trial. While it may be unlikely that continuous, automatic processes such as ste pping evoke the same processes of learning and consolidation as the discrete tasks used in the visuomotor literature our results may suggest that basal ganglia related deficits in savings during adaptive motor learning may be overcome by increased overlea rning. Indeed, previous research suggests that persons with PD do not retain newly learned motor tasks in the absence of sufficient practice as well as older adults without subcortical dysfunction (Mochizuki Kawai et al., 2004). However, th e role of overlearning in reducing parkinsonian deficits in savings requires further research within both the locomotor and visuomotor adaptation paradigms. Biomechanics of R educed S tep L ength A ftereffects d uring D e adaptation As a primary finding of this study w as that step length aftereffects were diminished during de adaptation, it is important to consider how step length can be changed from a biomechanical perspective. That is, what movement patterns of the lower limbs are changing to alter the positioning of the feet at the heel strike of each limb? Step length can be altered by adjusting the kinematics and kinetics of gait in a variety of ways. In terms of the kinematics, step length is defined by the distance along the anterior posterior axis in position of the feet at lead limb heel strike. Thus, when all other kinetics and kinematics are held constant, increase s in hip flexion and knee extension in the lead (or stepping) limb will result in greater for ward positioning of the leading foot. This would ul timately lead to a longer step with the leading limb. Our results indicate that during the first five strides of POST TIED persons with PD exhibited significantly greater hip flexion and knee extension at the heel strike of the fast

PAGE 136

136 limb while ON meds a s compared to OFF meds While the slow limb took a significantly shorter step during POST TIED as compared to BASELINE in both the ON meds and OFF meds states, only while ON meds did the persons with PD take a significantly longer step with the fast limb during POST TIED as compared to BASELINE. Thus, while OFF meds step length asymmetry during POST TIED was limited by decreased hip flexion and knee extension of the fast limb. That is, the fast limb foot was placed closer to the trunk (in the anterior posterior direction) at heel strike while OFF meds which ultimately limited length of fast limb step during POST TIED such that it was similar to BASELINE. It stands to reason that, due to the increased rigidity observed in this study and others when persons with PD are OFF meds (Hornykiewicz, 1973) perhaps the fast limb step length was limited during POST TIED because the participants simply could not generate significantly more hip flexion and knee extension with the more affected (fast ) limb than that which was observed during BASELINE. However, in addition to the BASELINE tied belt data collected at the slow speed, we also collected baseline gait data in which the participants walked with both limbs moving at the fast speed (not repor ted in this document) During this fast speed baseline, the participants demonstrated the ability to generate steps which were markedly longer than those observed during POST TIED, even with the more affected limb. This suggests that even while OFF meds, persons with PD maintained the ability to produce longer steps than were taken during POST TIED. Thus, the inability to produce a similarly long fast limb step during POST TIED in the OFF and ON meds states (due to rigidity or other

PAGE 137

137 parkinsonian motor fe atures) was likely not a factor in the diminished aftereffect observed OFF meds. P revious research has also suggested that changes in step length while walking on the treadmill can be described by changes in either spatial (center of oscillation) or tempor al (limb phasing) characteristics of the oscillation of the limbs (Malone, 2010). T he center s of oscillation in each limb are asymmetric during de adaptation as the limb angle at mid stance around which the fast limb oscillates is significantly more flexe d than the mid stance limb angle around w hich the slow limb oscillates (i.e. the fast limb foot is generally placed further in front of the trunk throughout the gait cycle as compared to the slow limb foot). Limb angle phasing is also shifted such that fast limb is phase lagged relative to the slow limb. In the current study, d uring POST TIED the center of oscillation difference was greater in six of the ten PD participants while the limb phasing lag was greater in four of the ten PD participants when ON meds as compared to OFF meds. These findings indicate that the greater step length asymmetry observed during POST TIED when the participants were ON meds resulted from increased storage of spatial characteristics of the SBT gait pattern in six particip ants and increased storage of temporal characteristics of the SBT gait pattern in four participants. As two of the ten participants demonstrated greater step length asymmetry while OFF meds as compared to ON meds during POST TIED, the six/four ratio menti oned above indicates that some participants demonstrated increased storage of both spatial and temporal characteristics of the SBT pattern. Thus, dopaminergic medication does not appear to enhance the storage of either specifically spatial or temporal cha racteristics of adapted gait patterns Rather, dopamine appears to

PAGE 138

138 enhance the storage of altered foot placement patterns which some participants achieved by storing either spatial or temporal (or both) components of the adapted gait pattern. Step length can also modified by the stance phase kinetics of the contralateral limb. That is, the left leg can be made to take a longer step if the right leg generates greater propulsion during late stance to drive the body forward (Balasubramanian et al., 2007) Specifically, Judge and colleagues (1996) demonstrated that ankle plantarflexor power generation during the push off phase of late stance is the strong est kinetic predictor of step length Further evidence for the contribution of contralateral limb propul sion and ankle plantarflex ion mechanics to step length has been recently demonstrated in post stroke literature. For example, Balasubramanian and colleagues (2007) showed that persons post stroke exhibiting the greatest impairment in paretic limb propuls ive force generation also exhibited the greatest degree of step length asymmetry (though in addition to reduced non paretic step length, this was also thought to be due in part to compensatory lengthening of the paretic step). Allen and colleagues (2011) observed that persons post stroke generating significantly longer steps with the paretic limb (as compared to the non paretic limb) produced significantly diminished propulsive forces and ankle plantarflexor moments during p ush off with the paretic limb. Further, multiple groups have demonstrated that ankle plantarflexor spasticity is an important determinant for step length asymmetry (Hsu et al., 2003; Lin et al., 2006) We observed that, while SBT walking, the fast limb generates significantly higher peak ankle plantarflexor power during push off as well as greater AP GRFs and ankle

PAGE 139

139 plantarflexor work throughout the propulsive phase of gait. We hypothesized that an aftereffect ma y exist during de adaptation suc h that the slow limb overcorrects for these patterns of asymmetric force and power production (similar to how the nervous system overcorrects for errors in interlimb spatiotemporal gait parameters during de adaptation) and thus the slow limb would generate greater push off during de adaptation. This idea of greater push off contributions from the slow limb ankle plantarflexors could explain the increase in fast limb step length during POST TIED when ON meds. However, we observed that the propulsive GRFs, peak ankle plantarflexor powers, and propulsive ankle plantarflexor work returned to baseline values immediately such that they were not significantly different between BASELINE and POST TIED Of particular interest is that the portion of the propulsive i mpulse prior to double limb support (i.e. the portion of the propulsive impulse which could theoretically drive the center of mass forward prior to contralateral heel strike, thus influencing contralateral step length) also showed very little aftereffect. These findings further the notion that dopamine influences the storage of adapted kinematics and that the kinetics adapt reactively (similar to intralimb parameters) during locomotor adaptation and adaptive learning Thus, after an acute bout of SBT walk ing, the increased step length asymmetry observed during POST TIED when persons with PD are ON meds results from a kinematic aftereffect related to the placement of the feet rather than a kinetic aftereffect of asymmetric force or torque production. Neur al M echanisms P otentially U nderlying D iminished A ftereffect S torage OFF M eds Previous research has indicated that multiple different brain structures and neural networks are important for various phases of the locomotor adaptation and locomotor

PAGE 140

140 adaptive learning processes. The cerebellum plays important roles in both adaptation and adaptive learning processes (Morton and Bastian, 2006; Jayaram et al., 2011; Jayaram et al., 2012). Initial adaptation appears to be largely spinally mediated. The primary motor cortex also plays a role in storing aftereffects (Galea et al., 2011). To our knowledge, this is the first study to suggest that dopaminergic pathways also play a role in the storage of aftereffects during locomotor adaptive learning. Insigh ts into the role of spinal control of initial locomotor adaptation have come largely from the study of decerebrate cats. It has long been known that spinalized cats retain the ability to walk on a treadmill and can even respond to external perturbations d uring gait (see Rossignol et al., 1996 for review). Further, these cats appear to retain the ability to adapt gait over time as observed by improvements in walking patterns after chronic locomotor training (Barbeau and Rossignol, 1987). Yanagihara and co lleages (1993) expanded on these findings within an acute setting by observ ing that decerebrate cats retain the ability to adapt gait to SBT walking. While spinal control of gait in bipedal species (such as humans) is not as well understood as spinal cont rol of gait in quadrupeds, it has been hypothesized that spinal structures control adaptation of the reactively modulated, intralimb parameters during initial locomotor adaptation (Reisman et al., 2005). Evidence supporting this hypothesis has begun to mo unt as several studies suggest that damage to or dysfunction of cortical (Reisman et al., 2007) subcortical (Roemmich et al., 2013, in revision) and cerebellar (Morton and Bastian, 2006) structures does not impair initial adaptation of reactively modulat ed gait parameters through feedback control. Thus, when pairing these findings with those describing relatively unaffected visuomotor adaptation (Marinelli et al., 2009; Bdard

PAGE 141

141 and Sanes, 2011; Leow et al., 2012) and locomotor adaptation to a podokinetic treadmill (Hong, Perlmutter, and Earhart, 2007) in PD it was not particularly surprising that we did not observe any effect of levodopa on reactively modulated parameters during locomotor adaptation. The basal ganglia do not appear to play a significant role in feedback control of gait adaptation Conversely, dysfunction in supraspinal structures does have some impact on feedforward control of gait adaptation. The cerebellum has long been known to play a significant role in trial and error adaptation of movement (see Thach et al., 1992 for review). More recently, it has been shown that anodal cerebellar stimulation increases the rate of adaptation during SBT walking (Jayaram et al., 2012) and p ersons with cerebellar damage also experience disruption in f eedforward, predictive control of gait adaptation (Morton and Bastian, 2006) suggesting a role of the cerebellum in modulation of predictively controlled interlimb gait parameters When walking on a SBT, persons with cerebellar damage did not reach the s ymmetric asymptote in interlimb parameters that is typically reached within the first few minutes of SBT walking in healthy adults. Further, very little step length and double limb support time aftereffects were observed during de adaptation. The latter findings are similar in pattern to those we observed in the participants with PD when OFF meds. However, given that step length asymmetry and double limb support asymmetry were not significantly different between medicated states during LATE adaptation, i t is unlikely that hypoactivation of the cerebellum is a viable mechanism underlying the aftereffect impairme nts in the unmedicated PD state That is, unlike the cerebellar patients, unmedicated persons with PD were still able to reach a relatively symmet ric asymptote

PAGE 142

142 by LATE adaptation. Though we observed diminished aftereffect storage, predictively modulated feedforward control of gait adaptation remains intact in persons with PD. Several neural mechanisms could potentially underlie the diminished storage of step length aftereffects observed in the current study when persons with PD were OFF meds. First, as previously discussed, cerebellar dysfunction has been shown to lead to dimi nished aftereffects during de adaptation after SBT walking (Morton and Bastian, 2006), though it is unlikely that cerebellar dysfunction played a role in our findings for reasons described above. Hyperactivity of the cerebellum and/or basal ganglia dysfun ction could have also played a role in the reduced aftereffects observed OFF meds. These mechanisms will be discussed in greater detail within the following paragraphs. Beyond locomotor adaptation, t here are specific changes which occur in cerebellar acti vity during locomotor adaptive learning (i.e. the storage of aftereffects during de adaptation and savings during re adaptation) Multiple studies of perturbed locomotion in decerebrate cats have demonstrated increases in complex spike discharge patterns within climbing afferent fibers to the cerebellar Purkinje cells in response to the perturbations (Matsukawa and Udo, 1985; Kim, Wang, and Ebner, 1987). In another study of decerebrate cats, Yanigihara and Udo (1994) observed increased complex spike firi ng rates in these climbing fibers during SBT walking, concluding that climbing fiber discharges indicate error signals even in the control of locomotion. A cellular phenomenon known as long term depression (LTD) which indicates the reduction of synaptic transmission from parallel fibers to cerebellar Purkinje cells, occurs when parallel and climbing fibers are activated simultaneously

PAGE 143

143 w hile both converging onto the same Purkinje cell (Ito, 1989). Thus, it was hypothesized that the increase in complex spi ke firing rates in climbing fibers leads to LTD which is necessary to facilitate locomotor adaptation and potentially locomotor adaptive learning. Indeed, LTD has been shown to be vital to loco motor adaptation in decerebrate cats Nitric oxide has been shown to induce LTD (Lev Ram et al., 1995), and thus when nitric oxide was deprived in decerebrate cats, locomotor adaptation to SBT walking was abolished (Yanigihara and Kondo, 1996). Recent work by Jayaram and colleagues has brought investigation of the relationships between LTD and locomotor adaptive learning into the human realm. These authors observed that both the magnitude of adaptation during SBT walking (measured as the difference in step length asymmetry between the first five strides and the las t 30 seconds of the 10 minute SBT adaptation task) as well as the magnitude of the step length aftereffects following SBT walking were strongly associated with the reduction of cerebellar inhibition over the primary cortex imm ediately following the SBT tas k (Jayaram et al., 2011). As LTD decreases excitability of the Purkinje cells and reduced cerebellar inhibition is indicative of decreased excitability of the Purkinje cells, this study demonstrated that LTD of cerebellar Purkinje cells is vital to locomo tor adaptation and locomotor adaptive learning. Thus, inhibitory activity of the cerebellum over the motor cortex must be reduced in order to produce large step length aftereffects following SBT walking. This could be problematic for persons with PD as hy peractivity of the cerebellum has recently been observed during upper extremity motor tasks in PD (Wu and Hallett, 2005; Cerasa et al., 2006; Yu et al., 2007). These observations have been made

PAGE 144

144 primarily while persons with PD are OFF meds, with cerebellar activity appearing more normal while ON meds (Rascol et al., 1997) and in very early stage de novo patients (Spraker et al., 2010) It stands to reason that, if the cerebellum is hyperactive in PD when OFF meds sufficiently reducing cerebellar activity to allow for storage of large aftereffects during de adaptation could be difficult and thus impaired in the unmedicated state Our findings regarding reduced step length aftereffects during de adaptation when OFF meds are in line with these ideas. However significant research should be done before cerebellar hyperactivity can be considered to be a probable mechanism underlying decreased aftereffect storage in unmedicated persons with PD. As the magnitude of adaptation from EARLY to LATE was not different between the ON and OFF meds states, it is difficult to conclude that cerebellar hyperactivity is likely the sole mechanism underlying the results observed in the current study. Further, c erebellar hyperactivity has been previously observed during upper extremity tapping movements requiring very little interlimb coordination, as these experiments were performed while the participants underwent functional brain imaging. Thus, the neural cor relates controlling the tasks which have been previously studied are likely dissimilar to those which control locomotion and, specifically, locomotor adaptive learning. While one SPECT imaging study did observe hyperactive patterns in the cerebellar vermi s during gait in PD (Hanakawa et al., 1999), further research should explore the role of the cerebellum with regard to both steady state gait and locomotor adaptive learning in persons with PD. A third, and perhaps most likely, neural mechanism potentially underlying diminished aftereffects in PD OFF meds can be proposed by considering the role s of

PAGE 145

145 the basal ganglia and primary motor cortex in motor learning. Motor adaptation has often been described as the learning of an internal model that predicts conse quences of motor output (Huang et al., 2011). Accordingly, recent research has attempted to describe these internal models of motor adaptation. While there remains significant controversy as to whether the nervous system controls adaptation through the a ctivation or selection of these learned models (or whether these models even exist at all; Huang et al., 2011), much of the controversy surrounds mechanisms of savings while de adaptation mechanisms can be modeled quite accurately. For instance, Kojima, I wamoto, and Yoshi d a (2004) proposed a two module, gain specific model This model was able to accurately account for adaptation and de adaptation rates by assigning one module the role of increas ing movement gain and the other the role of decreasing gain Smith, Ghazizadeh, and Shadmehr (2006) recently proposed a different two module multi rate model of motor adaptation and adaptive learning. This model was composed of a fast acting module that strongly responds to sensory errors and a slow acting modul e which is less sensitive to sensory errors but indicates learning throughout the adaptation and adaptive learning processes The fast module represents movement patterns during adaptation and de adaptation with a high degree of accuracy. On the other ha nd, the authors suggest that the slow module is important to account for savings and other phenomena which occur after de adaptation that were previously unaccounted for by the gain specific model (Smith, Ghazizadeh, and Shadmehr, 2006). Interestingly, bo th the gain specific and multi rate models are able to predict the rapid rate of de adaptation as compared to the relatively slower initial adaptation. Based on the previously discussed cerebellar related impairments in adaptation and de

PAGE 146

146 adaptation the authors of the multi rate model hypothesized that the fast module is likely controlled by the cerebellum. However, recent research by Huang and colleagues has identified shortcomings of the proposed models such that these investig ators have suggested that adaptive learning may be controlled by additional so acting in parallel with internal models (Huang et al., 2011). Hu ang and colleagues (2011) suggested that adaptation and adaptive learning are susceptible to influence from model free processes such as use dependent plasticity and operant reinforcement learning. Once the performance asymptote is reached during adaptation, use dependent plasticity is induced through repeated performance of the adapted movement and perfo rmance of this particular movement is reinforced due to perceived success of adaptation to the perturbation. In sum, the authors postulated that model based learning, use dependent plasticity, and operant reinforcement learning occur in conjunction during motor adaptation and adaptive learning. It is logical then that these processes may be controlled by anatomically distinct neural processes. I t seems highly likely that model based learning d uring adaptation is cerebellar influenced based on research dem onstrating impaired motor adaptation in persons with cerebellar dysfunction (Morton and Bastian, 2006) Persons with PD, on the contrary, show deficiencies in operant reinforcement learning. In a study by Rutledge and colleagues (2009), dopaminergic medi cation was demonstrated to selectively reinforce positive outcomes during motor learning while this reinforcement was absent once the participants were OFF meds. Thus, the rewarding of movements deemed successful may be affected in PD when OFF meds. This is perhaps not surprising given the largely

PAGE 147

147 important role of dopamine in reward based learning (Schul t z, 1998). However, interestingly, this idea is in line with a suggestion proposed by Huang and colleagues regarding the mechanisms of adaptive learning (Huang et al., 2011). These authors postulated that the primary motor cortex may be important in the storage of aftereffects during de adaptation and savings during re adaptation (Huang et al., 2011) as the primary motor cortex is both directly and indirectly affected by dopaminergic signals from the midbrain and basal ganglia, respectively ( Luft and Schwarz, 2009; Alexander, DeLong, and Strick, 1986) Indeed, r ecent work in the upper extremity has dissociated the effects of the cerebellum on adapta tion from the effects of the primary motor cortex on storage of aftereffects during de adaptation (Galea et al., 2011). Direct electrical s timulation of the cerebellum facilitated faster adaptation while stimulation of the primary motor cortex facilitated larger aftereffects during de adaptation (Galea et al., 2011). Therefore, it is possible that t he basal ganglia may play a significant role in the modulation of cortical activity during de adaptation. Activity in the basal ganglia has influence over cort ical activity via the thalamus through the classically described cortico basal thalamic loops ( Alexander, DeLong, and Strick, 1986; DeLong, 1990) Using a transcranial magnetic stimulation paradigm plasticity of the primary motor corte x was found to be a ltered in unmedicated PD such that excitability of the cortex in response to electrical stimulation was reduced (Ueki et al., 2006). The observed impairment in cortical excitability was not observed in the same persons with PD upon testing in the optimall y medicated state. Thus, if excitability of the primary motor cortex is essential for the storage of aftereffects during de adaptation (Galea et al., 2011), it stands to reason that the reduced cortical excitability observed in persons with PD when OFF

PAGE 148

148 me ds could significantly impair the storage of aftereffects. Indeed, this is precisely what we observed in the present study. In summary, we have proposed several neural mechanisms which could help to explain the diminished aftereffects during de adaptation we observed when the participants were OFF meds. While cerebellar dysfunction has been previously shown to diminish the storage of aftereffects during de adaptation (Morton and Bastian, 2006), persons with PD are typically not characterized by cerebellar insult Hyperactivity in the cerebellum is frequently observed in persons with PD (Wu and Hallett, 2005; Yu et al., 2007) and inhibitory activity of the cerebellum over the primary motor cortex must be depressed to facilitate aftereffect storage during d e adaptation (Jayaram et al., 2011). However, cerebellar hyperactivity has not been previously reported in persons with PD when performing motor adaptation and adaptive learning tasks and thus the potential role of cerebellar hyperactivity in the diminish ed step length aftereffects observed in the present study is purely speculative. Finally, primary motor cortex excitability has previously been shown to be important for the storage of aftereffects during de adaptation (Galea et al., 2011) and persons wit h PD demonstrate reduced cortical excitability when OFF meds but not ON meds (Ueki et al., 2006). Moreover, altered function of the cortico basal thalamic loops seems like the most likely mechanism underlying the diminished step length aftereffects we obs erved when persons with PD OFF meds, though significant further research is needed to either confirm or reject these postulations as they are admittedly highly speculative Clinical I mplications W e have previously shown that an acute bout of SBT walking can at least temporarily alter step length asymmetry in persons with PD (Roemmich et al., in

PAGE 149

149 revision, 2013). While not all persons with PD exhibit gait asymmetry, those affected by asymmetric walking patterns may benefit from SBT training. We have observed that asymmetry. That is, they exhibit significant step length asymmetry while walking under conditions during which step lengt h is typically symmetrical in healthy adults This is observed even during late SBT walking after step length asymmetry has been initially perturbed and adapted toward symmetry (Roemmich et al., in revision, 2013) Thus, large changes in step length asym metry during POST TIED may be necessary to The findings of the present study indicate that, in order to obtain maximal change in step length asymmetry after SBT walking exercise, persons w ith PD and rehabilitation professionals should be careful to ensure that the exercise is being performed while in the optimally medicated state. Effects of asymmetric locomotor training on the SBT may be lesser in magnitude if performed before the dopamin ergic medication has taken full effect, during wearing off periods, or in the absence of medication entirely. Further, while this study does not provide direct evidence of the effects of overlearning on savings in persons with PD, our results indicate that savings is unaffected by dopamine in persons with PD who have been given afforded a rather large amount of overlearning during the adaptation portion of the paradigm. These findings were in contrast with previous studies which provided either zero ( Bdar d and Sanes, 2011) or relatively little (Marinelli et al., 2011; Leow et al., 2012) opportunity for overlearning during motor adaptation. Given these findings, we suggest that allowing overlearning during a motor adaptation task may allow for increased sa vings in persons

PAGE 150

150 with PD. This may have clinical importance in a rehabilitation setting in which the rehabilitation facility perhaps to increase the amplitude of stepping or to practice navigating a common obstacle such a step, for example and then allow the patient to reproduce these learned tasks in similar, practical everyday settings. Limitations 1. We collected data in 30 second intervals during the locomotor adaptation an d adaptive learning tasks. Unfortunately, the participants did not reach asymptote performance within the first 30 seconds of the adaptation phase (EARLY) and thus measurement of the adaptation rate was not possible. 2. All conditions were performed while pa rticipants walked on the SBT. The assessment of aftereffects during de adaptation have the most direct clinical relevance when these trials are performed during overground gait (i.e. if the SBT is to be used to restore gait symmetry during everyday locomo tion, it is important to understand the magnitude of the aftereffects of SBT on overground gait). However, due to limited space with which to collect reliable overground gait data, we opted to assess aftereffects during de adaptation while the participant s walked on the SBT. This is a commonly used procedure (Reisman et al., 2005; Reisman et al., 2007). 3. Persons with PD experience motor fluctuations such that they sometimes experience day to day differences in motor symptom severity. Since performance in the ON and OFF states was assessed on different days, the changes in performance and UPDRS scores may not be due entirely to effects of dopaminergic medication. The order of the ON and OFF meds testing sessions was randomized among participants and thus w e do not suspect day to day motor fluctuations had a significant influence on our reported outcomes. 4. The treadmill speeds tested in this study were standardized across all participants. They were not self selected nor were any scales of perceived exertion performed. Thus, the task was likely easier for some of the participants relative to others. It is unknown to what effect perceived exertion influences locomotor adaptation or adaptive learning during SBT walking. Conclusion The purposes of this study w e re to investigate : 1) locomotor adaptation and adaptive learning in persons with PD in their native, unmedicated state and 2) how d opaminergic therapy may affect abilit ies to adapt and store new gait patterns in

PAGE 151

151 persons with PD. Our results indicate tha t persons with PD retain the abilities to adapt locomotion to an SBT perturbation, to store aftereffects of the adapted SBT pattern during de adaptation, and demonstrate savings of the adapted SBT pattern upon re adaptation to a second bout of SBT walking even while OFF meds. However, dopaminergic medication enhances the magnitude of step length aftereffects observed during de adaptation while adaptation and savings during re adaptation are unaffected. These findings have import ant implications with regar d gait rehabilitation of persons with PD as well as understanding the neural processes controlling of locomotor adaptation and adaptive learning If SBT training is to be used to restore gait symmetry in persons with PD, training sessions should be perfor med while the patients are optimally medicated. Further, dopaminergic pathways appear to play a significant role in the storage of aftereffects following locomotor adaptation.

PAGE 152

152 APPENDIX STA TI STICAL ANALYSES TABLES Table A 1. Multivariate r esults of MANOVA analyzing effects of walking condition and medication on spatiotemporal gait characteristics during locomotor adaptation (Specific Aim 1). Conditions included Dependent variables Effect Parameters BASELINE, EARLY, LATE Stride length asymmetry S tance time asymmetry D ouble limb support time asymmetry S tep length asymmetry Condition F = 129.056 p < 0.001 BASELINE, EARLY, LATE Stride length asymmetry S tance time asymmetry D oub le limb support time asymmetry S tep length asymmetry Meds F = 0.410 p = 0.796 BASELINE, EARLY, LATE Stride length asymmetry S tance time asymmetry D ouble limb support time asymmetry S tep length asymmetry Condition*Meds F = 0.905 p = 0.524

PAGE 153

153 Table A 2. Univariate results of MANOVA analyzing effects of walking condition and medication on spatiotemporal gait characteristics during locomotor adaptation (Specific Aim 1). Conditions included Effect Dependent variable Parameters BASELINE, EARLY, LATE Condition Stride length asymmetry F = 787.638 p < 0.001* Stance time asymmetry F = 109.704 p < 0.001* Step length asymmetry F = 34.252 p < 0.001* Double limb support time asymmetry F = 115.037 p < 0.001* BASELINE, EARLY, LATE Meds Stride length asymmetry F = 0.021 p = 0.887 Stance time asymmetry F = 0.144 p = 0.713 Step length asymmetry F = 0.216 p = 0.653 Double limb support time asymmetry F = 0.208 p = 0.659 BASELINE, EARLY, LATE Condition*Meds Stride length asymmetry F = 0.039 p = 0.962 Stance time asymmetry F = 0.101 p = 0.904 Step length asymmetry F = 2.087 p = 0.153 Double limb support time asymmetry F = 1.467 p = 0.257

PAGE 154

154 Table A 3. Pairwise comparisons of MANOVA analyzing effects of walking condition on spatiotemporal gait characteristics during locomotor adaptation (Specific Aim 1). Dependent variable Condition 1 Condition 2 Mean Difference Standard Error p value Stride length asymmetry BASELINE EARLY 0.211 0.006 <0.001* LATE 0.246 0.006 < 0.001* EARLY BASELINE 0.211 0.006 <0.001* LATE 0.035 0.008 0.002* Stance time asymmetry BASELINE EARLY 0.115 0.009 <0.001* LATE 0.088 0.006 <0.001* EARLY BASELINE 0.115 0.009 <0.001* LATE 0.028 0.008 0.010* Step length asymmetry BASELINE EARLY 0.112 0.014 <0.001* LATE 0.016 0.015 0.305 EARLY BASELINE 0.112 0.014 <0.001* LATE 0.096 0.015 <0.001* Double limb support time asymmetry BASELINE EARLY 0.169 0.013 <0.001* LATE 0.001 0.012 0.931 EARLY BASELINE 0.169 0.013 <0.001* LATE 0.168 0.013 <0.001*

PAGE 155

155 Table A 4 Multivariate r esults of MANOVA analyzi ng effects of walking condition and medication on AP GRF impulses during locomotor adaptation (Specific Aim 2). Conditions included Dependent variables Effect Parameters BASELINE, EARLY, LATE Fast limb braking impulse F ast limb propulsive impulse F ast limb propulsive impulse prior to DS F ast limb propulsive impulse during DS S low limb braking impulse S low limb propulsive impulse S low limb propulsive impulse prior to DS S low limb propulsive impulse during DS Condition F = 14.064 p < 0.001 BASELINE, EARLY, LATE Fast limb braking impulse Fast limb propulsive impulse Fast limb propulsive impulse prior to DS Fast limb propulsive impulse during DS Slow limb braking impulse Slow limb propulsive impulse Slow limb propulsive impulse prior to DS Slow limb propulsive impulse during DS Meds F = 0. 988 p = 0. 656

PAGE 156

156 Table A 4. Continued. Conditions included Dependent variables Effect Parameters BASELINE, EARLY, LATE Fast limb braking impulse Fast limb propulsive impulse Fast limb propulsive impulse prior to DS Fast limb propulsive impulse during DS Slow limb braking impulse Slow limb propulsive impulse Slow limb propulsive impulse prior to DS Slow limb propulsive impulse during DS Condition*Meds F = 0.549 p = 0.886

PAGE 157

157 Table A 5. Univariate results of MANOVA analyzing effects of walking condition and medication on AP GRF impulses during locomotor adaptation (Specific Aim 2 ). Conditions included Effect Dependent variable Parameters BASELINE, EARLY, LATE Condition Fast limb braking impulse F = 48.954 p < 0.001* Fast limb propulsive impulse F = 8.510 p = 0.007 Fast limb propulsive impulse prior to DS F = 0.445 p = 0. 624 Fast limb propulsive impulse during DS F = 26.774 p < 0.001* Slow limb braking impulse F = 13.702 p = 0.003* Slow limb propulsive impulse F = 1.626 p = 0.228 Slow limb propulsive impulse prior to DS F = 3.352 p = 0.080 Slow limb propulsive impulse during DS F = 9.714 p = 0.002* BASELINE, EARLY, LATE Meds Fast limb braking impulse F = 0.149 p = 0.709 Fast limb propulsive impulse F = 0.419 p = 0.535 Fast limb propulsive impulse prior to DS F = 0.825 p = 0.390 Fast limb propulsive impulse during DS F = 0.068 p = 0.801 Slow limb braking impulse F = 0.035 p = 0.857 Slow limb propulsive impulse F = 1.006 p = 0.345

PAGE 158

158 Table A 5. Continued. Conditions included Effect Dependent variable Parameters Slow limb propulsive impulse prior to DS F = 0.361 p = 0.564 Slow limb propulsive impulse during DS F = 0.597 p = 0.462 BASELINE, EARLY, LATE Condition*Meds Fast limb braking impulse F = 0.111 p = 0.892 Fast limb propulsive impulse F = 0.253 p = 0.707 Fast limb propulsive impulse prior to DS F = 1.917 p = 0.182 Fast limb propulsive impulse during DS F = 0.642 p = 0.454 Slow limb braking impulse F = 0.025 p = 0.950 Slow limb propulsive impulse F = 0.613 p = 0.513 Slow limb propulsive impulse prior to DS F = 0.894 p = 0.419 Slow limb propulsive impulse during DS F = 0.100 p = 0.882

PAGE 159

159 Table A 6. Pairwise comparisons of MANOVA analyzing effects of walking condition on AP GRF impulses during locomotor adaptation (Specific Aim 2 ). Dependent variable Condition 1 Condition 2 Mean Difference Standard Error p value Fast limb braking impulse BASELINE EARLY 2.993 0.649 0.00 2 LATE 7.748 0.775 <0.001* EARLY BASELINE 2.993 0.649 0.00 2 LATE 4.755 0. 922 0.00 1 Fast limb propulsive impulse BASELINE EARLY 3.702 1.494 0.0 38 LATE 5.437 0.889 <0.001* EARLY BASELINE 3.702 1.494 0.038 LATE 1.735 1.555 0. 297 Fast limb propulsive impulse prior to DS BASELINE EARLY 0.906 1.089 0.430 LATE 0.274 0.779 0.734 EARLY BASELINE 0.906 1.089 0.430 LATE 0.633 1.057 0.566 Fast limb propulsive impulse during DS BASELINE EARLY 2.290 0.730 0.014 LATE 4.671 0.515 <0.001* EARLY BASELINE 2.290 0.730 0.014 LATE 2.381 0.651 0.00 6 Slow limb braking impulse BASELINE EARLY 3.300 0.808 0.004* LATE 0.451 0.451 0.347 EARLY BASELINE 3.300 0.808 0.004* LATE 3.751 0.990 0.005*

PAGE 160

160 Table A 6. Continued. Dependent variable Condition 1 Condition 2 Mean Difference Standard Error p value Slow limb propulsive impulse BASELINE EARLY LATE 1.873 0.800 1.622 1.062 0.282 0.473 EARLY BASELINE 1.873 1.622 0.282 LATE 2.673 1.785 0.173 Slow limb propulsive impulse prior to DS BASELINE EARLY 0.075 1.279 0.955 LATE 2.663 0.762 0.008* EARLY BASELINE 0.075 1.279 0.955 LATE 2.588 1.378 0.097 Slow limb propulsive impulse during DS BASELINE EARLY 1.823 0.498 0.006* LATE 1.833 0.499 0.006* EARLY BASELINE 1.823 0.498 0.006* LATE 0.010 0.437 0.982

PAGE 161

161 Table A 7 Multivariate r esults of MANOVA analyz ing effects of walking condition and medication on sagittal ankle kinetics during locomotor adaptation (Specific Aim 2). Conditions included Dependent variables Effect Parameters BASELINE, EARLY, LATE Fast limb peak A2 power Fast limb A2 work Slow limb peak A2 power Slow limb A2 work Condition F = 3.603 p = 0.00 7 BASELINE, EARLY, LATE Fast limb peak A2 power Fast limb A2 work Slow limb peak A2 power Slow limb A2 work Meds F = 0.120 p = 0.968 BASELINE, EARLY, LATE Fast limb peak A2 power Fast limb A2 work Slow limb peak A2 power Slow limb A2 work Condition*Meds F = 1.681 p = 0 155

PAGE 162

162 Table A 8. Univariate results of MANOVA analyzing effects of walking condition and medication on sagittal ankle kinetics during locomotor adaptation (Specific A im 2 ). Conditions included Effect Dependent variable Parameters BASELINE, EARLY, LATE Condition Fast limb peak A2 power F = 37.604 p < 0.001* Fast limb A2 work F = 27.042 p < 0.001* Slow limb peak A2 power F = 3.967 p = 0. 066 Slow limb A2 work F = 0.051 p = 0.950 BASELINE, EARLY, LATE Meds Fast limb peak A2 power F = 0.0 76 p = 0. 790 Fast limb A2 work F = 0. 006 p = 0. 939 Slow limb peak A2 power F = 0. 284 p = 0. 611 Slow limb A2 work F = 0. 581 p = 0. 471 BASELINE, EARLY, LATE Condition*Meds Fast limb peak A2 power F = 2.999 p = 0.084 Fast limb A2 work F = 0.672 p = 0.503 Slow limb peak A2 power F = 1.195 p = 0.322 Slow limb A2 work F = 1.309 p = 0.294

PAGE 163

163 Table A 9. Pairwise comparisons of MANOVA analyzing effects of walking condition on sagittal ankle kinetics during locomotor adaptation (Specific Aim 2). Dependent variable Condition 1 Condition 2 Mean Difference Standard Error p value Fast limb peak A2 power BASELI NE EARLY 0.962 0.162 0.001* LATE 0. 911 0. 109 <0.001* EARLY BASELINE 0. 962 0. 162 0.001* LATE 0.051 0.0 94 0. 604 Fast limb A2 work BASELINE EARLY 11.988 2.255 0.001* LATE 9.848 1.167 <0.001* EARLY BASELINE 11.988 2.255 0.001* LATE 2.139 1.620 0.228 Slow limb peak A2 power BASELINE EARLY 0.325 0.184 0.121 LATE 0.371 0.128 0.023* EARLY BASELINE 0.325 0.184 0.121 LATE 0.046 0.107 0.679 Slow limb A2 work BASELINE EARLY 0.108 1.230 0.933 LATE 0.225 1.064 0.839 EARLY BASELINE 0.108 1.230 0.933 LATE 0.332 0.852 0.708

PAGE 164

164 Table A 10. Multivariate results of MANOVA analyzing effects of walking condition and medication on spatiotemporal gait characteristics during the de adaptation phase of locomotor adaptive learning (Specific Aim 1). Conditions included Dependent variables Effect Parameters BASELINE, POST TIED Stride length asymmetry Stance time asymmetry Double limb support time asymmetry Step length asymmetry Condition F = 29.0 6 6 p < 0.001* BASELINE, POST TIED Stride length asymmetry Stance time asymmetry Double limb support time asymmetry Step length asymmetry Meds F = 1.031 p = 0. 463 BASELINE, POST TIED Stride length asymmetry Stance time asymmetry Double limb support time asymmetry Step length asymmetry Condition*Meds F = 5.852 p = 0. 029*

PAGE 165

165 Table A 11 Univariate results of MANOVA analyzing effects of walking condition and medication on spatiotemporal gait characteristics during the de adaptation phase of locomotor adaptive learning (Specific Aim 1). Conditions included Effect Dependent variable Parameters BASELINE, POST TIED Condition Stride length asymmetry F = 27.186 p = 0.001* Stance time asymmetry F = 16.103 p = 0.00 3 Step length asymmetry F = 41.967 p < 0.001* Double limb support time asymmetry F = 42.335 p < 0.001* BASELINE, POST TIED Meds Stride length asymmetry F = 0. 167 p = 0. 692 Stance time asymmetry F = 0. 475 p = 0. 508 Step length asymmetry F = 5.748 p = 0. 040* Double limb support time asymmetry F = 0. 159 p = 0. 700 BASELINE, POST TIED Condition*Meds Stride length asymmetry F = 0. 305 p = 0. 594 Stance time asymmetry F = 0. 040 p = 0. 845 Step length asymmetry F = 6.806 p = 0. 028* Double limb support time asymmetry F = 1. 620 p = 0. 235

PAGE 166

166 Table A 12. ANOVA results for linear regression analyzing BASELINE, EARLY, and LATE step length asymmetry values as predictors of POST TIED step length asymmetry OFF meds (Specific Aim 1). Model Sum of Squares df Mean Square F p value Regression 0.015 3 0.005 7.460 0.019* Residual 0.004 6 0.001 Total 0.019 9 Table A 13 Linear regression results analyzing BASELINE, EARLY, and LATE step length asymmetry values as predictors of POST TIED step length asymmetry OFF meds (Specific Aim 1). Predictor Unstandardized Beta Coefficients Standard Error Standardized Beta Coefficien ts t p value BASELINE 0.046 0.265 0.070 0.341 0.745 EARLY 0.397 0.191 0.447 2.079 0.083 LATE 0.793 0.183 0.987 4.328 0.005

PAGE 167

167 Table A 1 4 ANOVA results for linear regression analyzing BASELINE, EARLY, and LATE step length asymmetry values as predictors of POST TIED step length asymmetry ON meds (Specific Aim 1). Model Sum of Squares D f Mean Square F p value Regression 0.002 3 0.001 0.638 0.618 Residual 0.008 6 0.001 Total 0.010 9 Table A 15. Linear regression results analyzing BASELINE, EARLY, and LATE step length asymmetry values as predictors of POST TIED step length asymmetry ON meds (Specific Aim 1). Predictor Unstandardized Beta Coefficients Standard Error Standardized Beta Coefficients t p value BASELINE 0.281 0.435 0.315 0.647 0.542 EARLY 0.258 0.190 0.712 1.362 0.222 LATE 0.162 0.314 0.215 0.517 0.642

PAGE 168

168 Table A 16 Paired t test results analyzing sagittal hip, knee, and ankle angles at heel strikes during POST TIED between medicated states (Specific Aim 1). Dependent variable Mean Difference (OFF ON) Standard Error t p value Slow hip angle at slow heel strike 1.291 1.209 1.067 0.317 Slow knee angle at slow heel strike 1.756 1.228 1.431 0.190 Fast hip angle at slow heel strike 1.702 1.274 1.336 0.218 Fast knee angle at slow heel strike 3.181 1.444 2.203 0.059 Slow hip angle at fast heel strike 1.324 1.736 0.763 0.467 Slow knee angle at fast heel strike 2.146 1.418 1.513 0.169 Fast hip angle at fast heel strike 3.371 1.347 2.503 0.037* Fast knee angle at fast heel strike 3.990 1.923 2.074 0.072

PAGE 169

169 Table A 17 Paired t test results analyzing fast and slow limb step lengths between BASELINE and POST TIED in both medicated states (Specific Aim 1). Dependent variable Mean Difference (BASELINE POST TIED) Standard Error t p value Fast step length (OFF) 0.0130 0.0184 0.0708 0.497 Slow step length (OFF) 0.0491 0.0136 3.615 0.006* Fast step length (ON) 0.0463 0.0176 2.639 0.027* Slow step length (ON) 0.0437 0.0135 3.236 0.010*

PAGE 170

170 Table A 18. Multivariate results of MANOVA analyzing effects of walking condition and medication on spatiotemporal gait characteristics during the r e adaptation phase of locomotor adaptive learning relative to EARLY (Specific Aim 1). Conditions included Dependent vari ables Effect Parameters EARLY RE ADAPT Stride length asymmetry Stance time asymmetry Double limb support time asymmetry Step length asymmetry Condition F = 43.878 p < 0.001* EARLY, RE ADAPT Stride length asymmetry Stance time asymmetry Double limb support time asymmetry Step length asymmetry Meds F = 2.153 p = 0.192 EARLY, RE ADAPT Stride length asymmetry Stance time asymmetry Double limb support time asymmetry Step length asymmetry Condition*Meds F = 0.707 p = 0.616

PAGE 171

171 Table A 19. Univariate results of MANOVA analyzing effects of walking condition and medication on spatiotemporal gait characteristics during the re adaptation phase of locomotor adaptive learning relative to EARLY (Specific Aim 1). Conditions included Effect Depende nt variable Parameters EARLY, RE ADAPT Condition Stride length asymmetry F = 11.354 p = 0.00 8 Stance time asymmetry F = 5.899 p = 0.0 38 Step length asymmetry F = 20.237 p = 0.001* Double limb support time asymmetry F = 39.266 p < 0.001* EARLY, RE ADAPT Meds Stride length asymmetry F = 0.16 0 p = 0.6 98 Stance time asymmetry F = 1.012 p = 0. 341 Step length asymmetry F = 1.195 p = 0. 303 Double limb support time asymmetry F = 1.000 p = 0. 343 EARLY, RE ADAPT Condition*Meds Stride length asymmetry F = 1.525 p = 0. 248 Stance time asymmetry F = 0.744 p = 0. 411 Step length asymmetry F = 0.942 p = 0. 357 Double limb support time asymmetry F = 0.083 p = 0. 779

PAGE 172

172 Table A 20. Multivariate results of MANOVA analyzing effects of walking condition and medication on spatiotemporal gait characteristics during the re adaptation phase of locomotor adaptive learning relative to LATE (Specific Aim 1). Conditions included Dependent varia bles Effect Parameters LATE RE ADAPT Stride length asymmetry Stance time asymmetry Double limb support time asymmetry Step length asymmetry Condition F = 21.234 p = 0.001* LATE RE ADAPT Stride length asymmetry Stance time asymmetry Double limb support time asymmetry Step length asymmetry Meds F = 0.518 p = 0. 727 LATE RE ADAPT Stride length asymmetry Stance time asymmetry Double limb support time asymmetry Step length asymmetry Condition*Meds F = 1.768 p = 0.254

PAGE 173

173 Table A 21. Univariate results of MANOVA analyzing effects of walking condition and medication on spatiotemporal gait characteristics during the re adaptation phase of locomotor adaptive learning relative to LATE (Specific Aim 1). Conditions included Effect Dependen t variable Parameters LATE, RE ADAPT Condition Stride length asymmetry F = 11. 592 p = 0.008* Stance time asymmetry F = 3.866 p = 0.0 81 Step length asymmetry F = 5.325 p = 0.0 46 Double limb support time asymmetry F = 40.985 p < 0.001* LATE, RE ADAPT Meds Stride length asymmetry F = 0. 565 p = 0. 471 Stance time asymmetry F = 1. 327 p = 0. 279 Step length asymmetry F = 0.535 p = 0. 483 Double limb support time asymmetry F = 0.000 p = 0. 997 LATE, RE ADAPT Condition*Meds Stride length asymmetry F = 1. 265 p = 0.2 90 Stance time asymmetry F = 0. 982 p = 0. 348 Step length asymmetry F = 2.741 p = 0. 132 Double limb support time asymmetry F = 2.426 p = 0. 154

PAGE 174

174 Table A 22. Multivariate results of MANOVA analyzing effects of walking condition and medication on AP GRF impulses during the de adaptation phase of locomotor a daptive learning (Specific Aim 2 ). Conditions included Dependent variables Effect Parameters BASELINE, P OST TIED Fast limb braking impulse Fast limb propulsive impulse Fast limb propulsive impulse prior to DS Fast limb propulsive impulse during DS Slow limb braking impulse Slow limb propulsive impulse Slow limb propulsive impulse prior to DS Slow limb propul sive impulse during DS Condition F = 3.036 p = 0.418 BASELINE, POST TIED Fast limb braking impulse Fast limb propulsive impulse Fast limb propulsive impulse prior to DS Fast limb propulsive impulse during DS Slow limb braking impulse Slow limb propulsive impulse Slow limb propulsive impulse prior to DS Slow limb propulsive impulse during DS Meds F = 1.847 p = 0.517

PAGE 175

175 Table A 22. Continued. Conditions included Dependent variables Effect Parameters BASELINE, POST TIED Fast limb braking impulse Fast limb propulsive impulse Fast limb propulsive impulse prior to DS Fast limb propulsive impulse during DS Slow limb braking impulse Slow limb propulsive impulse Slow limb propulsive impulse prior to double support Slow limb propulsive impulse d uring double support Condition*Meds F = 9.319 p = 0.248

PAGE 176

176 Table A 23. Multivariate results of MANOVA analyzing effects of walking condition and medication on sagittal ankle kinetics during the de adaptation phase of locomotor a daptive learning (Specific Aim 2 ). Conditions included Dependent variables Effect Parameters BASELINE, POST TIED Fast limb peak A2 power Fast limb A2 work Slow limb peak A2 power Slow limb A2 work Condition F = 1 598 p = 0. 545 BASELINE, POST TIED Fast limb peak A2 power Fast limb A2 work Slow limb peak A2 power Slow limb A2 work Meds F = 0.331 p = 0.874 BASELINE, POST TIED Fast limb peak A2 power Fast limb A2 work Slow limb peak A2 power Slow limb A2 work Condition*Meds F = 8.267 p = 0. 262

PAGE 177

177 Table A 24. Multivariate results of MANOVA analyzing effects of walking condition and medication on AP GRF impulses during the r e adaptation phase of locomotor adaptive learning relative to EARLY (Specific Aim 2). Conditions included Dependent variables Effect Param eters EARLY, RE ADAPT Fast limb braking impulse Fast limb propulsive impulse Fast limb propulsive impulse prior to DS Fast limb propulsive impulse during DS Slow limb braking impulse Slow limb propulsive impulse Slow limb propulsive impulse prior to DS Slow limb propulsive impulse during DS Condition F = 2.438 p = 0. 460 EARLY, RE ADAPT Fast limb braking impulse Fast limb propulsive impulse Fast limb propulsive impulse prior to DS Fast limb propulsive impulse during DS Slow limb braking impulse Slow limb propulsive impulse Slow limb propulsive impulse prior to DS Slow limb propulsive impulse during DS Meds F = 4.430 p = 0. 353

PAGE 178

178 Table A 24. Continued. Conditions included Dependent variables Effect Parameters EARLY, RE ADAPT Fast limb braking impulse Fast limb propulsive impulse Fast limb propulsive impulse prior to DS Fast limb propulsive impulse during DS Slow limb braking impulse Slow limb propulsive impulse Slow limb propulsive impulse prior to DS Slow limb propulsive impulse during DS Condition*Meds F = 97.009 p = 0.078

PAGE 179

179 Table A 25. Multivariate results of MANOVA analyzing effects of walking condition and medication on sagittal ankle kinetics during the r e adaptation phase of locomotor adaptive learning relative to EARLY (Specific Aim 2). Conditions included Dependent variables Effect Parameters EARLY RE ADAPT Fast limb peak A2 power Fast limb A2 work Slow limb peak A2 power Slow limb A2 work Condition F = 181.856 p = 0. 057 EARLY RE ADAPT Fast limb peak A2 power Fast limb A2 work Slow limb peak A2 power Slow limb A2 work Meds F = 0 .460 p = 0. 816 EARLY RE ADAPT Fast limb peak A2 power Fast limb A2 work Slow limb peak A2 power Slow limb A2 work Condition*Meds F = 1.941 p = 0. 504

PAGE 180

180 Table A 26. Multivariate results of MANOVA analyzing effects of walking condition and medication on AP GRF impulses during the re adaptation phase of locomotor adaptive learning relative to LATE (Specific Aim 2). Conditions included Dependent variables Effect Parameters LATE, RE ADAPT Fast limb braking impulse Fast limb propulsive impulse Fast limb propulsive impulse prior to DS Fast limb propulsive impulse during DS Slow limb braking impulse Slow limb propulsive impulse Slow li mb propulsive impulse prior to DS Slow limb propulsive impulse during DS Condition F = 105.567 p = 0. 075 LATE, RE ADAPT Fast limb braking impulse Fast limb propulsive impulse Fast limb propulsive impulse prior to DS Fast limb propulsive impulse during DS Slow limb braking impulse Slow limb propulsive impulse Slow limb propulsive impulse prior to DS Slow limb propulsive impulse during DS Meds F = 17.222 p = 0. 184

PAGE 181

181 Table A 26. Continued. Conditions included Dependent variables Effect Parameters LATE, RE ADAPT Fast limb braking impulse Fast limb propulsive impulse Fast limb propulsive impulse prior to DS Fast limb propulsive impulse during DS Slow limb braking impulse Slow limb propulsive impulse Slow limb propulsive impulse prior to DS Slow limb propulsive impulse during DS Condition*Meds F = 0.847 p = 0.691

PAGE 182

182 Table A 27. Multivariate results of MANOVA analyzing effects of walking condition and medication on sagittal ankle kinetics during the re adaptation phase of locomotor adaptive learning relative to LATE (Specific Aim 2). Conditions included Dependent variables Effect Parameters LATE, RE ADAPT Fast limb peak A2 power Fast limb A2 work Slow limb peak A2 power Slow limb A2 work Condition F = 2.888 p = 0. 425 LATE, RE ADAPT Fast limb peak A2 power Fast limb A2 work Slow limb peak A2 power Slow limb A2 work Meds F = 0. 447 p = 0. 822 LATE, RE ADAPT Fast limb peak A2 power Fast limb A2 work Slow limb peak A2 power Slow limb A2 work Condition*Meds F = 2.612 p = 0. 444

PAGE 183

183 LIST OF REFERENCES Impact of deprenyl and tocopherol treatment on Parkinson's disease in DATATOP patients requiring levodopa. Pa rkinson Study Group. Ann Neurol 1996 ;39:37 45. Ahlskog JE, Muenter MD. Frequency of levodopa related dyskinesias and motor fluctuations as estimated from the cumulative literature. Mov Disord 2001;16:448 58. Albin RL, Young AB, Penney JB. The functional anatomy of basal ganglia disorde rs. Trends Neurosci 1989;12:366 75. Alexander GE, Crutcher MD. Functional architecture of basal ganglia circuits: neural substrates of parallel processing. Trends Neurosci 1990;13:266 71. Alexander GE, DeLong MR, Strick PL. Parallel organization of f unctionally segregated circuits linking basal ganglia and cortex. Annu Rev Neurosci 1986;9:357 81. Allen JL, Kautz SA, Neptune RR. Step length asymmetry is representative of compensatory mechanisms used in post stroke hemiparetic walking. Gait Posture 2 011;33:538 43. Aravamuthan BR, Muthusamy KA, Stein JF, Aziz TZ, Johansen Berg H. Topography of cortical and subcortical connections of the human pedunculopontine and subthalamic nuclei. Neuroimage 2007;37:694 705. Balasubramanian CK, Bowden MG, Neptu ne RR, Kautz SA. Relationship between step length asymmetry and walking performance in subjects with chronic hemiparesis. Arch Phys Med Rehabil 2007;88:43 9. Baltadjieva R, Giladi N, Gruendlinger L, Peretz C, Hausdorff JM. Marked alterations in the gait timing and rhythmicity of patients with de novo Parkinson's disease. Eur J Neurosci 2006;24:1815 20. Baram Y, Miller A. Virtual reality cues for improvement of gait in patients with multiple sclerosis. Neurology 2006;66:178 81. Barbeau H, Rossigno l S. Recovery of locomotion after chronic spinalization in the adult cat. Brain Res 1987;412:84 95. Bares M, Lungu OV, Hus rov I, Gescheidt T. Predictive motor timing performance dissociates between early diseases of the cerebellum and Parkinson's dis ease. Cerebellum 2010;9:124 35. Bares M, Lungu O, Liu T, Waechter T, Gomez CM, Ashe J. Impaired predictive motor timing in patients with cerebellar disorders. Exp Brain Res 2007;180:355 65.

PAGE 184

184 B dard P, Sanes JN. Basal ganglia dependent processes in r ecalling learned visual motor adaptations. Exp Brain Res 2011;209:385 93. B irkmayer W H ornykiewicz O. [The L 3,4 dioxyphenylalanine (DOPA) effect in Parkinson akinesia]. Wien Klin Wochenschr 1961;73:787 8. Blin O, Ferrandez AM, Pailhous J, Serratr ice G. Dopa sensitive and dopa resistant gait parameters in Parkinson's disease. J Neurol Sci 1991;103:51 4. Brach JS, Berlin JE, VanSwearingen JM, Newman AB, Studenski SA. Too much or too little step width variability is associated with a fall history in older persons who walk at or near normal gait speed. J Neuroeng Rehabil 2005;2:21. Brady RA, Peters BT, Bloomberg JJ. Strategies of healthy adults walking on a laterally oscillating treadmill. Gait Posture 2009;29:645 9. Buekers M, Montagne G, de Rugy A, Laurent M. The regulation of externally paced human locomotion in virtual reality. Neurosci Lett 1999;275:171 4. Caliandro P, Ferrarin M, Cioni M, Bentivoglio AR, Minciotti I, D'Urso PI, et al. Levodopa effect on electromyographic activation patterns of tibialis anterior muscle during walking in Parkinson's disease. Gait Posture 2011;33:436 41. Cappellini G, Ivanenko YP, Dominici N, Poppele RE, Lacquaniti F. Motor patterns during walking on a slippery walkway. J Neurophysiol 2010;103:746 60. Carta M, Bezard E. Contribution of pre synaptic mechanisms to L DOPA induced dyskinesia. Neuroscience 2011;198:245 51. Cavaco S, Anderson SW, Correia M, Magalhaes M, Pereira C, Tuna A, et al. Task specific contribution of the human striatum to perceptual motor skill learning. J Clin Exp Neuropsychol 2011;33:51 62. Cerasa A, Hagberg GE, Peppe A, Bianciardi M, Gioia MC, Costa A, et al. Functional changes in the activity of cerebellum and frontostriatal regions during externally and internally timed movement in Parkinson's disease. Brain Res Bull 2006;71:259 69. C hambers WW S prague JM Functional localization in the cerebellum. I. Organization in longitudinal cortico nuclear zones and their contribution to the control of posture, both extra pyramidal and pyramidal. J Comp Neurol 1955;103:105 29. Choi JT, Vining EP, Reisman DS, Bastian AJ. Walking flexibility after hemispherectomy: split belt treadmill adaptation and fe edback control. Brain 2009;132 :722 33. Cioni M, Richards CL, Malouin F, Bedard PJ, Lemieux R. Characteristics of the electromyographic patterns of lower limb muscles during gait in patients with

PAGE 185

185 Parkinson's disease when OFF and ON L Dopa treatment. Ital J Neurol Sci 1997;18:195 208. Clark DJ, Ting LH, Zajac FE, Neptune RR, Kautz SA. Merging of healthy motor modules predicts reduced locomotor performance and muscle coordination complexity post stroke. J Neurophysiol 2010;103:844 57. Cotzias GC, Papavasiliou PS, Gellene R. L dopa in parkinson's syndrome. N Engl J Med 1969;281:272. Cotzias GC, Van Woert MH, Schiffer LM. Aromatic amino acids and modification of parkinsonism. N Engl J Med 1967;276:374 9. Darter BJ, Wilken JM. Gait training with virtual reality based real time feedback: improving gait perfor mance following transfemoral amputation. Phys Ther 2011;91:1385 94. DeLong MR. Primate models of movement disorders of basal ganglia origin. Trends Neurosci 1990;13:281 5. Dietz V. Spinal cord pattern generators for locomotion. Clin Neurophysiol 20 03;114:1379 89. Dietz V, Colombo G, Jensen L. Locomotor activity in spinal man. Lancet 1994;344 :1260 3. Dietz V, Zijlstra W, Prokop T, Berger W. Leg muscle activation during gait in Parkinson's disease: adaptation and interlimb coordination. Electr oencephalogr Clin Neurophysiol 1995;97:408 15. Driskell JE, Willis RP, Copper C. Effect of overlearning on retention. J App Psychol 1992;77:615 22. Earhart GM, Jones GM, Horak FB, Block EW, Weber KD, Fletcher WA. Forward versus backward walking: t ransfer of podokinetic adaptation. J Neurophysiol 2001;86:1666 70. Evarts EV, Wise SP. Basal ganglia outputs and motor control. Ciba Found Symp 1984;107:83 102. Fahn S. Biochemistry of the basal ganglia. Adv Neurol 1976;14:59 89. Fahn S. The fr eezing phenomenon in parkinsonism. Adv Neurol 1995;67:53 63. Fahn S, Oakes D, Shoulson I, Kieburtz K, Rudolph A, Lang A, et al. Levodopa and the progression of Parkinson's disease. N Engl J Med 2004;351:2498 508.

PAGE 186

186 Fearnley JM, Lees AJ. Ageing and Pa rkinson's disease: substantia nigra regional selectivity. Brain 1991;114:2283 301. Fernandez Ruiz J, Diaz R, Hall Haro C, Vergara P, Mischner J, Nu ez L, et al. Normal prism adaptation but reduced after effect in basal ganglia disorders using a throwin g task. Eur J Neurosci 2003;18:689 94. Ferraye MU, Deb B, Fraix V, Goetz L, Ardouin C, Yelnik J, et al. Effects of pedunculopontine nucleus area stimulation on gait disorders in Parkinson's disease. Brain 2010;133:205 14. Forssberg H, Grillner S, Halbertsma J, Rossignol S. The locomotion of the low spinal cat. II. Interlimb coordination. Acta Physiol Scand 1980;108:283 95. Galea JM, Vazquez A, Pasricha N, de Xivry JJ, Celnik P. Dissociating the roles of the cerebellum and motor cortex during adaptive learning: the motor cortex retains what the cerebellum learns. Cereb Cortex 2011;21:1761 70. Gates DH, Dingwell JB, Scott SJ, Sinitski EH, Wilken JM. Gait characteristics of individuals with transtibial amputations walking on a destabilizing r ock surface. Gait Posture 2012;36:33 9. Geurts AC, Boonstra TA, Voermans NC, Diender MG, Weerdesteyn V, Bloem BR. Assessment of postural asymmetry in mild to moderate Parkinson's disease. Gait Posture 2011;33:143 5. Griffin HJ, Greenlaw R, Limousin P, Bhatia K, Quinn NP, Jahanshahi M. The effect of real and virtual visual cues on walking in Parkinson's disease. J Neurol 2011;258:991 1000. Hanakawa T, Katsumi Y, Fukuyama H, Honda M, Hayashi T, Kimura J, et al. Mechanisms underlying gait disturban ce in Parkinson's disease: a single photon emission computed tomography study. Brain 1999;122:1271 82. Hassler R. Zur Pathologie der Paralysis agitains und des postenzephalitischen Parkinsonismus. J Psychol Neurol 1938;48:386 476. Hathout GM, Bhida yasiri R. Midbrain ataxia: an introduction to the mesencephalic locomotor region and the pedunculopontine nucleus. AJR Am J Roentgenol 2005;184:953 6. Hausdorff JM. Gait dynamics in Parkinson's disease: common and distinct behavior among stride length, gait variability, and fractal like scaling. Chaos 2009;19:026113. Hausdorff JM, Rios DA, Edelberg HK. Gait variability and fall risk in community living older adults: a 1 year prospective study. Arch Phys Med Rehabil 2001;82:1050 6.

PAGE 187

187 Hirsch EC, Gra ybiel AM, Duyckaerts C, Javoy Agid F. Neuronal loss in the pedunculopontine tegmental nucleus in Parkinson disease and in progressive supranuclear palsy. Proc Natl Acad Sci U S A 1987;84:5976 80. Hong M, Perlmutter JS, Earhart GM. Podokinetic after ro tation in Parkinson disease. Brain Res 2007;1128:99 106. Hong M, Earhart GM. Rotating treadmill training reduces freezing in Parkinson disease: preliminary observations. Parkinsonism Relat Disord 2008;14:359 63. Hornykiewicz O. Dopamine (3 hydroxy tyramine) and brain function. Pharmacol Rev 1966;18:925 64. Hsu AL, Tang PF, Jan MH. Analysis of impairments influencing gait velocity and asymmetry of hemiplegic patients after mild to moderate stroke. Arch Phys Med Rehabil 2003;84:1185 93. Huang VS, Haith A, Mazzoni P, Krakauer JW. Rethinking motor learning and savings in adaptation paradigms: model free memory for successful actions combines with internal models. Neuron 2011;70:787 801. Isaias IU, Moisello C, Marotta G, Schiavella M, Canesi M Perfetti B, et al. Dopaminergic striatal innervation predicts interlimb transfer of a visuomotor skill. J Neurosci 2011;31:14458 62. Ito M. Long term depression. Annu Rev Neurosci 1989;12:85 102. Ivanenko YP, Poppele RE, Lacquaniti F. Five basic muscle activation patterns account for muscle activity during human locomotion. J Physiol 2004;556:267 82. Jahanshahi M, Jones CR, Zijlmans J, Katzenschlager R, Lee L, Quinn N, et al. Dopaminergic modulation of striato frontal connectivity during motor timing in Parkinson's disease. Brain 2010;133:727 45. Jahn K, Deutschl nder A, Stephan T, Kalla R, H fner K, Wagner J, et al. Supraspinal locomotor control in quadrupeds and humans. Prog Brain Res 2008;171:353 62. Jankovic J, McDermott M, Carter J, Gauthier S, Goetz C, Golbe L, et al. Variable expression of Parkinson's disease: a base line analysis of the DATATOP cohort. The Parkinson Study Group. Neurology 1990;40:1529 34. Jayaram G, Galea JM, Bastian AJ, Cel nik P. Human locomotor adaptive learning is proportional to depression of cerebellar excitability. Cereb Cortex 2011;21:1901 9. Jayaram G, Tang B, Pallegadda R, Vasudevan EV, Celnik P, Bastian A. Modulating locomotor adaptation with cerebellar stimulat ion. J Neurophysiol 2012;107:2950 7.

PAGE 188

188 Joiner WM, Smith MA. Long term retention explained by a model of short term learning in the adaptive control of reaching. J Neurophysiol 2008;100:2948 55. Jubault T, Monetta L, Strafella AP, Lafontaine AL, Monch i O. L dopa medication in Parkinson's disease restores activity in the motor cortico striatal loop but does not modify the cognitive network. PLoS One 2009;4:e6154. Judge JO, Davis RB, Ounpuu S. Step length reductions in advanced age: the role of ankle and hip kinetics. J Gerontol A Biol Sci Med Sci 1996;51:M303 12. Kahn JH, Hornby TG. Rapid and long term adaptations in gait symmetry following unilateral step training in people with hemiparesis. Phys Ther 2009;89:474 83. Karachi C, Grabli D, Ber nard FA, Tand D, Wattiez N, Belaid H, et al. Cholinergic mesencephalic neurons are involved in gait and postural disorders in Parkinson disease. J Clin Invest 2010;120:2745 54. Kemp JM, Powell TP. The connexions of the striatum and globus pallidus: sy nthesis and speculation. Philos Trans R Soc Lond B Biol Sci 1971;262:441 57. Kim JH, Jang SH, Kim CS, Jung JH, You JH. Use of virtual reality to enhance balance and ambulation in chronic stroke: a double blind, randomized controlled study. Am J Phys Med Rehabil 2009;88:693 701. Kim JH, Wang JJ, Ebner TJ. Climbing fiber afferent modulation during treadmill locomotion i n the cat. J Neurophysiol 1987;57:787 802. Knutsson E. An analysis of Parkinsonian gait. Brain 1972;95:475 86. Kojima Y, Iwamoto Y, Yoshida K. Memory of learning facilitates saccadic adaptation in the monkey. J Neurosci 2004;24:7531 9. Kornheis er AS. Adaptation to laterally displaced vision: a rev iew. Psychol Bull 1976;83:783 816. Kraft E, Loichinger W, Diepers M, Lule D, Schwarz J, Ludolph AC, et al. Levodopa induced striatal activation in Parkinson's disease: a functional MRI study. Parkin sonism Relat Disord 2009;15:558 63. Krystkowiak P, Blatt JL, Bourriez JL, Duhamel A, Perina M, Blond S, et al. Effects of subthalamic nucleus stimulation and levodopa treatment on gait abnormalities in Parkinson disease. Arch Neurol 2003;60:80 4. K uo SH, Kenney C, Jankovic J. Bilateral pedunculopontine nuclei strokes presenting as freezing of gait. Mov Disord 2008;23:616 9.

PAGE 189

189 Lee CS, Schulzer M, Mak E, Hammerstad JP, Calne S, Calne DB. Patterns of asymmetry do not change over the course of idiopat hic parkinsonism: implications for pathogenesis. Neurology 1995;45:435 9. Lee MS, Rinne JO, Marsden CD. The pedunculopontine nucleus: its role in the genesis of movement disorders. Yonsei Med J 2000;41:167 84. Leow LA, Loftus AM, Hammond GR. Impair ed savings despite intact initial learning of motor adaptation in Parkinson's disease. Exp Brain Res 2012;218:295 304. Lev Ram V, Makings LR, Keitz PF, Kao JP, Tsien RY. Long term depression in cerebellar Purkinje neurons results from coincidence of ni tric oxide and depolarization induced Ca2+ transients. Neuron 1995;15:407 15. Lewis CL, Ferris DP. Invariant hip moment pattern while walking with a robotic hip exoskeleton. J Biomech 2011;44:789 93. Lewis MM, Slagle CG, Smith AB, Truong Y, Bai P, McKeown MJ, et al. Task specific influences of Parkinson's disease on the striato thalamo cortical and cerebello thalamo cortical motor circuitries. Neuroscience 2007;147:224 35. Lewis SJ, Barker RA. A pathophysiological model of freezing of gait in Pa rkinson's disease. Parkinsonism Relat Disord 2009;15:333 8. Lin PY, Yang YR, Cheng SJ, Wang RY. The relation between ankle impairments and gait velocity and symmetry in people with stroke. Arch Phys Med Rehabil 2006;87:562 8. Luft AR, Schwarz S. Do paminergic signals in primary motor cortex. Int J Dev Neurosci 2009;27:415 21. Malone LA, Bastian AJ. Thinking about walking: effects of conscious correction versus distraction on locomotor adaptation. J Neurophysiol 2010;103:1954 62. Malone LA, Ba stian AJ, Torres Oviedo G. How does the motor system correct for errors in time and space during locomotor adaptation? J Neurophysiol 2012;108:672 83. Malone LA, Vasudevan EV, Bastian AJ. Motor adaptation training for faster relearning. J Neurosci 2011 ;31:15136 43. Marinelli L, Crupi D, Di Rocco A, Bove M, Eidelberg D, Abbruzzese G, et al. Learning and consolidation of visuo motor adaptation in Parkinson's disease. Parkinsonism Relat Disord 2009;15:6 11. Marsden CD, Parkes JD. "On off" effects i n patients with Parkinson's disease on chronic levodopa therapy. Lancet 1976;1:292 6.

PAGE 190

190 Martinu K, Degroot C, Madjar C, Strafella AP, Monchi O. Levodopa influences striatal activity but does not affect cortical hyper activity in Parkinson's disease. Eur J Neurosci 2012;35:572 83. Masdeu JC, Alampur U, Cavaliere R, Tavoulareas G. Astasia and gait failure with damage of the pontomesencephalic locomotor region. Ann Neurol 1994;35:619 21. Matsukawa K, Udo M. Responses of cerebellar Purkinje cells to m echanical perturbations during locomotion of decerebrate cats. Neurosci Res 1985;2:393 8. Mazzone P, Lozano A, Stanzione P, Galati S, Scarnati E, Peppe A, et al. Implantation of human pedunculopontine nucleus: a safe and clinically relevant target in Parkinson's disease. Neuroreport 2005;16:1877 81. Mbourou GA, Lajoie Y, Teasdale N. Step length variability at gait initiation in elderly fallers and non fallers, and young adults. Gerontology 2003;49:21 6. Miller RA, Thaut MH, McIntosh GC, Rice RR. Components of EMG symmetry and variability in parkinsonian and healthy elderly gait. Electroencephalogr Clin Neurophysiol 1996;101:1 7. Mink JW. The basal ganglia: focused selection and inhibition of compet ing motor programs. Prog Neurobiol 1996;50:381 425. Mirelman A, Patritti BL, Bonato P, Deutsch JE. Effects of virtual reality training on gait biomechanics of individuals post stroke. Gait Posture 2010;31:433 7. Mochizuki Kawai H, Kawamura M, Hasegawa Y, Mochizuki S, Oeda R, Yamanaka K, et al. Deficits in long term retention of learned motor skills in patients with cortical or subcortical degeneration. Neuropsychologia 2004;42:1858 63. Monchi O, Petrides M, Mejia Constain B, Strafella AP. Cortical activity in Parkinson's disease during executive processing depends on striatal involvement. Brain 2007;130:233 44. Morris ME, Iansek R, Matyas TA, Summers JJ. The pathogenesis of gait hypokinesia in Parkinson's disease. Brain 1994;117:1169 8 1. Morris ME, Matyas TA, Iansek R, Summers JJ. Temporal stability of gait in Parkinson's disease. Phys Ther 1996;76:763 77; discussion 78 80. Morris M, Iansek R, McGinley J, Matyas T, Huxham F. Three dimensional gait biomechanics in Parkinson's d isease: evidence for a centrally mediated amplitude regulation disorder. Mov Disord 2005;20:40 50. Morton SM, Bastian AJ. Cerebellar control of balance and locomotion. Neuroscientist 2004;10:247 59.

PAGE 191

191 Morton SM, Bastian AJ. Cerebellar contributions to locomotor adaptations during splitbelt treadmill walking. J Neurosci 2006;26:9107 16. Morton SM, Bastian AJ. Mechanisms of cerebellar gait ataxia. Cerebellum 2007;6:79 86. Mukherjee M, Siu KC, Katsavelis D, Fayad P, Stergiou N. The influence of visual perception of self motion on locomotor adaptation to unilateral limb loading. J Mot Behav 2011;43:101 11. Nakamura T, Meguro K, Sasaki H. Relationship between falls and stride length variability in senile dementia of the Alzheimer type. Gero ntology 1996;42:108 13. Nambu A, Tokuno H, Hamada I, Kita H, Imanishi M, Akazawa T, et al. Excitatory cortical inputs to pallidal neurons via the subthalamic nucleus in the monkey. J Neurophysiol 2000;84:289 300. Nambu A, Tokuno H, Takada M. Func tional significance of the cortico subthalamo pallidal 'hyperdirect' pathway. Neurosci Res 2002;43:111 7. Nanhoe Mahabier W, Snijders AH, Delval A, Weerdesteyn V, Duysens J, Overeem S, et al. Walking patterns in Parkinson's disease with and without fr eezing of gait. Neuroscience 2011;182:217 24. Ng B, Palmer S, Abugharbieh R, McKeown MJ. Focusing effects of L dopa in Parkinson's disease. Hum Brain Mapp 2010;31:88 97. Orlovsky GN. The effect of different descending systems on flexor and extens or activity during locomotion. Brain Res 1972;40:359 71. Palmer SJ, Eigenraam L, Hoque T, McCaig RG, Troiano A, McKeown MJ. Levodopa sensitive, dynamic changes in effective connectivity during simultaneous movements in Parkinson's disease. Neuroscienc e 2009;158:693 704. Payoux P, Brefel Courbon C, Ory Magne F, Regragui W, Thalamas C, Balduyck S, et al. Motor activation in multiple system atrophy and Parkinson disease: a PET study. Neurology 2010;75:1174 80. Pearson KG. Proprioceptive regulati on of locomotion. Curr Opin Neurobiol 1995;5:786 91. Pickering RM, Grimbergen YA, Rigney U, Ashburn A, Mazibrada G, Wood B, et al. A meta analysis of six prospective studies of falling in Parkinson's disease. Mov Disord 2007;22 :1892 900. Pieranto zzi M, Palmieri MG, Galati S, Stanzione P, Peppe A, Tropepi D, et al. Pedunculopontine nucleus deep brain stimulation changes spinal cord excitability in Parkinson's disease patients. J Neural Transm 2008;115:731 5.

PAGE 192

192 Plaha P, Gill SS. Bilateral deep br ain stimulation of the pedunculopontine nucleus for Parkinson's disease. Neuroreport 2005;16:1883 7. Plotnik M, Giladi N, Balash Y, Peretz C, Hausdorff JM. Is freezing of gait in Parkinson's disease related to asymmetric motor function? Ann Neurol 200 5;57:656 63. Purzner J, Paradiso GO, Cunic D, Saint Cyr JA, Hoque T, Lozano AM, et al. Involvement of the basal ganglia and cerebellar motor pathways in the preparation of self initiated and externally triggered movements in humans. J Neurosci 2007;27 :6029 36. Rahman S, Griffin HJ, Quinn NP, Jahanshahi M. On the nature of fear of falling in Parkinson's disease. Behav Neurol 2011;24:219 28. Rascol O. The pharmacological therapeutic management of levodopa induced dyskinesias in patients with Pa rkinson's disease. J Neurol 2000;247 Suppl 2:II51 7. Rascol O, Sabatini U, Fabre N, Brefel C, Loubinoux I, Celsis P, et al. The ipsilateral cerebellar hemisphere is overactive during hand movements in akinetic parkinsonian p atients. Brain 1997;120 :103 10. Reisman DS, Block HJ, Bastian AJ. Interlimb coordination during locomotion: what can be adapted and stored? J Neu rophysiol 2005;94:2403 15. Reisman DS, McLean H, Bastian AJ. Split belt treadmill training poststroke: a case study. J Neurol Ph ys Ther 2010;34:202 7. Reisman DS, McLean H, Keller J, Danks KA, Bastian AJ. Repeated split belt treadmill training improves poststroke step length asymmetry. Neurorehabil Neural Repair 2013;27:460 8. Reisman DS, Wityk R, Silver K, Bastian AJ. Locomotor adaptation on a split belt treadmill can improve walking symmetry post stroke. Brain 2007;130:1861 72. Reisman DS, Wityk R, Silver K, Bastian AJ. Split belt treadmill adaptation transfers to overground walking in persons poststroke. Neuroreh abil Neural Repair 2009;23:735 44. Roemmich RT, Nocera JR, Vallabhajosula S, Amano S, Naugle KM, Stegem ller EL, et al. Spatiotemporal variability during gait initiation in Parkinson's disease. Gait Posture 2012;36:340 3. Roemmich RT, Stegem ller EL, Hass CJ. Lower extremity sagittal joint moment production during split belt treadmill walking. J Biomech 2012;45:2817 21.

PAGE 193

193 Rossignol S, Chau C, Brustein E, B langer M, Barbeau H, Drew T. Locomotor capacities after complete and partial lesions of the spinal cord. Acta Neurobiol Exp (Wars) 1996;56:449 63. Rossignol S, Drew T, Brustein E, Jiang W. Locomotor performance and adaptation after partial or complete spinal cord lesions in the cat. Prog Brain Res 1999;123:349 65. Rutledge RB, Lazza ro SC, Lau B, Myers CE, Gluck MA, Glimcher PW. Dopaminergic drugs modulate learning rates and perseveration in Parkinson's patients in a dynamic foraging task. J Neurosci 2009;29:15104 14. Savin DN, Tseng SC, Morton SM. Bilateral adaptation during loc omotion following a unilaterally applied resistance to swing in nondisabled adults. J Neurophysiol 2010;104:3600 11. Schaafsma JD, Giladi N, Balash Y, Bartels AL, Gurevich T, Hausdorff JM. Gait dynamics in Parkinson's disease: relationship to Parkinso nian features, falls and response to levodopa. J Neurol Sci 2003;212:47 53. Schrag A, Jahanshahi M, Quinn N. What contributes to quality of life in patients with Parkinson's disease? J Neurol Neurosurg Psychiatry 2000;69:308 12. Schultz W. Predic tive reward signal of dopamine neurons. J Neurophysiol 1998;80:1 27. Schwartz AB, Ebner TJ, Bloedel JR. Responses of interposed and dentate neurons to perturbations of the locomotor cycle. Exp Brain Res 1987;67:323 38. Schweder PM, Hansen PC, Gre en AL, Quaghebeur G, Stein J, Aziz TZ. Connectivity of the pedunculopontine nucleus in parkinsonian freezing of gait. Neuroreport 2010;21:914 6. Sheik Nainar MA, Kaber DB. The utility of a virtual reality locomotion interface for studying gait behavio r. Hum Factors 2007;49:696 709. Shute CC, Lewis PR. The ascending cholinergic reticular system: neocortical, olfactory and subcortical projections. Brain 1967;90:497 520. Smiley Oyen AL, Lowry KA, Emerson QR. Learning and retention of movement se quences in Parkinson's disease. Mov Disord 2006;21:1078 87. Smith MA, Ghazizadeh A, Shadmehr R. Interacting adaptive processes with different timescales underlie short term motor learning. PLoS Biol 2006;4:e179. S prague JM C hambers WW. Regulatio n of posture in intact and decerebrate cat. I. Cerebellum, reticular formation, vestibular nuclei. J Neurophysiol 1953;16:451 63.

PAGE 194

194 Spraker MB, Prodoehl J, Corcos DM, Comella CL, Vaillancourt DE. Basal ganglia hypoactivity during grip force in drug na v e Parkinson's disease. Hum Brain Mapp 2010;31:1928 41. Stern Y, Mayeux R, Hermann A, Rosen J. Prism adaptation in Parkinson's disease. J Neurol Neurosurg Psychiatry 1988;51:1584 7. Tanner CM, Goldman SM. Epidemiology of Parkinson's disease. Neuro l Clin 1996;14:317 35. Thach WT, Goodkin HP, Keating JG. The cerebellum and the adaptive coordination of movement. Annu Rev Neurosci 1992;15:403 42. Thevathasan W, Coyne TJ, Hyam JA, Kerr G, Jenkinson N, Aziz TZ, et al. Pedunculopontine nucleus s timulation improves gait freezing in Parkinson disease. Neurosurgery 2011;69:1248 53; discussion 54. Thevathasan W, Pogosyan A, Hyam JA, Jenkinson N, Foltynie T, Limousin P, et al. Alpha oscillations in the pedunculopontine nucleus correlate with gait performance in parkinsonism. Brain 2012;135:148 60. Tykocki T, Mandat T, Nauman P. Pedunculopontine nucleus deep brain stimulation in Parkinson's disease. Arch Med Sci 2011;7:555 64. Ueki Y, Mima T, Kotb MA, Sawada H, Saiki H, Ikeda A, et al. Al tered plasticity of the human motor cortex in Parkinson's disease. Ann Neurol 2006;59:60 71. Wilcox RA, Cole MH, Wong D, Coyne T, Silburn P, Kerr G. Pedunculopontine nucleus deep brain stimulation produces sustained improvement in primary progressive freezing of gait. J Neurol Neurosurg Psychiatry 2011;82:1256 9. Wu T, Hallett M. A functional MRI study of automatic movements in patients with Parkinson's disease. Brain 2005;128:2250 9. Wu T, Wang L, Chen Y, Zhao C, Li K, Chan P. Changes of functional connectivity of the motor network in the resting state in Parkinson's disease. Neurosci Lett 2009;460:6 10. Yanagihara D, Kondo I. Nitric oxide plays a key role in adaptive control of locomotion in cat. Proc Natl Acad Sci U S A 1996;93:1329 2 7. Yanagihara D, Udo M. Climbing fiber responses in cerebellar vermal Purkinje cells during perturbed locomotion in decerebrate cats. Neurosci Res 1994;19:245 8. Yanagihara D, Udo M, Kondo I, Yoshida T. A new learning paradigm: adaptive changes in interlimb coordination during perturbed locomotion in decerebrate cats. Neurosci Res 1993;18:241 4.

PAGE 195

195 Yang YR, Tsai MP, Chuang TY, Sung WH, Wang RY. Virtual reality based training improves community ambulation in individuals with stroke: a randomize d controlled trial. Gait Posture 2008;28:201 6. Yeo SS, Lee DG, Choi BY, Chang CH, Jung YJ, Kim SH, et al. Neural connectivity of the pedunculopontine nucleus in relation to walking ability in chronic patients with intracerebral hemorrhage. Eur Neurol 2012;67:226 31. Yogev G, Plotnik M, Peretz C, Giladi N, Hausdorff JM. Gait asymmetry in patients with Parkinson's disease and elderly fallers: when does the bilateral coordination of gait require attention? Exp Brain Res 2007;177:336 46. Yu H, S ternad D, Corcos DM, Vaillancourt DE. Role of hyperactive cerebellum and motor cortex in Parkinson's disease. Neuroimage 2007;35:222 33. Yu J, Eidelberg E. Recovery of locomotor function in cats after localized cerebellar lesions. Brain Res 1983 ;273:121 31.

PAGE 196

196 BIOGRAPHICAL SKETCH Ryan Roemmich was born in York, Nebraska. He attended York High School and lat er graduated with a Bachelor of Science degree in b iological s ystems e ngineering from the University of Nebraska Lincoln in Lincoln, Nebraska While studying at the University of Nebraska Lincoln, he volunteered for two years in the Movement and Neurosciences Center at the Madonna Rehabilitation Hospital in Lincoln, Nebraska where he worked on projects involving biomechanical analysis of patho logical gait patterns. He then pursued a Ph.D. at the University of Florida in Dr. stimulatio n on gait and postural control in persons with essential tremor, and the potential of split belt treadmill walking in gait rehabilitation.