Operant Conditioning to Increase Ankle Control or Decrease ... · 48 Ankle clonus is common after...

TA EMG and SOL H-reflex Training Improves Function in SCI

1

Operant Conditioning to Increase Ankle Control or Decrease Reflex Excitability Improves 1 Reflex Modulation and Walking Function in Chronic Spinal Cord Injury 2 3 4 Kathleen J. Manella1, Kathryn E. Roach1, and Edelle C. Field-Fote1, 2 5 6 1Department of Physical Therapy and 2The Miami Project to Cure Paralysis, Miller School of 7 Medicine, University of Miami, Miami, FL USA 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 Address correspondence to: E. C. Field-Fote, Department of Physical Therapy, Miller School of 42 Medicine, University of Miami, 5915 Ponce de Leon Blvd, 5th Floor, Coral Gables, FL 33146 43 44

45

Articles in PresS. J Neurophysiol (March 6, 2013). doi:10.1152/jn.01039.2011

Copyright © 2013 by the American Physiological Society.


2

ABSTRACT 46 47 Ankle clonus is common after spinal cord injury (SCI) and is attributed to loss of supra-48

spinally mediated inhibition of soleus stretch reflexes and maladaptive reorganization of spinal 49

reflex pathways. The maladaptive reorganization underlying ankle clonus is associated with 50

other abnormalities such as coactivation and reciprocal facilitation of tibialis anterior (TA) and 51

soleus (SOL), which contribute to impaired walking ability in individuals with motor-incomplete 52

SCI. Operant conditioning can increase muscle activation and decrease stretch reflexes in 53

individuals with SCI. We compared two operant conditioning-based interventions in individuals 54

with ankle clonus and impaired walking ability due to SCI. Training included either TA EMG 55

activation (TA↑) to enhance supraspinal drive or SOL H-reflex suppression (SOL↓) to modulate 56

reflex pathways at the spinal cord level. We measured clonus duration, plantar flexor (PF) reflex 57

threshold angle, timed toe tapping, dorsiflexion (DF) active range of motion (ROM), lower 58

extremity motor scores (LEMS), walking foot clearance, speed and distance, SOL H-reflex 59

amplitude modulation as an index of reciprocal inhibition, presynaptic inhibition, and low-60

frequency depression, and SOL/TA clonus coactivation,. TA↑ decreased PF reflex threshold 61

angle (-4.33°) and DF active ROM angle (-4.32°), and increased LEMS of DF (+0.8 points), 62

training leg (+2.2 points), and non-training leg (+0.8 points), and increased walking foot 63

clearance (+ 4.8 mm) and distance (+12.09 m). SOL↓ decreased SOL/TA coactivation ratio (-64

0.21) and increased non-training leg LEMS (+1.8 points), walking speed (+0.02 m/s) and 65

distance (+6.25 m). We found increased voluntary control associated with TA↑ outcomes and 66

decreased reflex excitability associated with SOL↓ outcomes. 67

68 Key words: ankle clonus, soleus stretch reflex, reciprocal inhibition, presynaptic inhibition, 69 antagonist coactivation, plantar flexor spasticity 70

71


3

INTRODUCTION 72

In persons with spinal cord injury (SCI) (see Appendix 1 for complete list of 73

abbreviations), walking function is impaired by reduced ability to produce voluntary muscle 74

contraction and by hyperactive spinal reflex activity (hyperreflexia). In those with motor-75

incomplete SCI, the impaired voluntary muscle contraction arises primarily from corticospinal 76

tract disruption, but reduced corticomotor excitability, maladaptive cortical changes (Beekhuizen 77

and Field-Fote 2005, Thomas and Gorassini 2005) and limited motor unit rate modulation 78

(Zijdewind and Thomas 2003) also contribute to this impairment. During normal human 79

walking, control of the tibialis anterior (TA) muscle is strongly related to corticospinal input 80

(Capaday et al. 1999; Schubert et al. 1997). Evidence demonstrates that TA corticospinal 81

excitability is enhanced with dorsiflexion imagery during imagined walking in non-disabled 82

individuals (Bakker et al. 2008), functional electrical stimulation after chronic central nervous 83

system lesions (Thompson et al. 2009), and locomotor training in persons with chronic motor-84

incomplete SCI (Thomas and Gorassini 2005). 85

Beyond the impairment of voluntary control, SCI also results in loss of modulation of 86

involuntary reflex circuits. Clonus is a common manifestation of stretch reflex hyperexcitablity 87

(Decq 2003). Clonus manifests as involuntary 5-7 Hz joint oscillations (Wallace et al. 2005), 88

and commonly occurs at the ankle in individuals with motor-incomplete SCI, and other forms of 89

central nervous system pathology. Clonic soleus (SOL) activity due to inadequate modulation of 90

the SOL stretch reflex may impede walking progression (Yang et al. 1991), compromise 91

independent walking (Beres-Jones et al. 2003), and restrict quality of life in individuals with 92

motor-incomplete SCI. 93


4

Reflex pathophysiology contributes to ankle clonus after SCI (Katz and Rymer 1989). 94

Disrupted supraspinal input and maladaptive reorganization of spinal reflex pathways give rise to 95

impaired modulation of peripheral sensory feedback (Field-Fote 2004; Hultborn 2003) and 96

aberrant reflex responses (Gracies 2005); exaggerated soleus stretch reflexes emerge (Davies et 97

al. 1995; Dietz 1997) and contribute to clonus in plantar flexors. After SCI, impaired SOL 98

disynaptic reciprocal Ia inhibition (RI) manifests as SOL coactivation during TA contractions 99

(Boorman et al. 1996), SOL reciprocal facilitation after TA activation (a reversal of the normal 100

response) (Crone and Nielsen 1994; Crone et al. 2003; Morita et al. 2001), and antagonist 101

response to SOL or TA tendon tapping (Xia and Rymer 2005). In individuals with motor-102

incomplete SCI, decreased SOL PI contributes to soleus stretch reflex hyperexcitability (Dietz 103

2001; Faist et al. 1994). 104

Increased SOL motoneuron excitability, reduced post-activation depression of repeated 105

stretch activations, and antagonist coactivation may promote clonus. Elevated SOL H/M ratio, 106

found in individuals with clonus (Koelman et al. 1993), has been associated with increased 107

plantar flexor (PF) reflex threshold angle (the angle at clonus onset) (Manella and Field-Fote 108

2010) and number of clonic oscillations (Manella and Field-Fote 2009). SOL H-reflex low-109

frequency post-activation depression (LFD) (Lloyd and Wilson 1957; Pierrot-Deseilligny and 110

Burke 2005) is impaired in persons with chronic SCI (Grey et al. 2008; Schindler-Ivens and 111

Shields 2000), allowing the SOL to be easily activated with small amounts of sensory input. In 112

spastic hemiparesis, dorsiflexor force is inversely related to plantar flexor spasticity and EMG 113

SOL/TA cocontraction ratio (Levin and Hui-Chan 1994), suggesting a relationship between the 114

amount of voluntary TA control and SOL hyperexcitability. After motor-incomplete SCI, 115

SOL/TA coactivation during clonus has been reported (Manella and Field-Fote 2009). Therefore 116


5

a myriad of factors, including SOL reflex hyperexcitability, diminished SOL LFD, decreased 117

voluntary dorsiflexor activation, and SOL/TA coactivation appear to influence clonic activity. 118

Control of both voluntary movement and reflex activity are known to be modifiable with 119

operant conditioning training, wherein reward is contingent upon performance of a targeted 120

motor task. Operant conditioning training induces neuroplasticity in supraspinal and spinal reflex 121

pathways. Such training increases voluntary EMG responses (Brucker and Bulaeva 1996; 122

Seymour and Bassler 1977; Stein et al. 1990) and firing rates of spared motor units (Stein et al. 123

1990) after motor-incomplete SCI, and decreases stretch reflex excitability in individuals with 124

and without spasticity (Evatt et al. 1989; Segal 1997; Wolf and Segal 1996). Successful SOL H-125

reflex suppression in animals (Wolpaw and Herchenroder 1990; Wolpaw and Chen 2001; 126

Wolpaw and Chen 2006) and humans (Thompson et al. 2009, 2008, 2006) provides clear 127

evidence of activity-dependent plasticity in spinal circuitry induced by operant conditioning 128

training. 129

Based on the foregoing evidence that both impaired voluntary dorsiflexor control and 130

plantar flexor hyperexcitability contribute to functional deficits and spasticity in persons with 131

spastic hypertonia, we designed two operant conditioning-based interventions intended to 132

improve ankle motor control in individuals with motor-incomplete SCI. The interventions were: 133

1) TA EMG activation (TA↑) to enhance supraspinal drive to spinal circuits and 2) SOL H-134

reflex suppression during voluntary DF (SOL↓) to modulate reflex pathways at the spinal cord 135

level. TA↑ was intended to improve ability to voluntarily drive corticospinal activation of 136

spared TA motor neurons and activate inhibitory interneurons (Iles and Pisini 1992) comprising 137

SOL RI and PI pathways. We hypothesized TA↑ would be associated with increased DF active 138

ROM and DF strength, foot clearance in swing phase, and walking speed and distance. SOL↓ 139


6

was intended to suppress SOL H-reflex amplitude responses during DF. We hypothesized 140

SOL↓ would be associated with enhanced SOL H-reflex LFD, decreased SOL/TA clonus 141

coactivation, decreased clonus duration, decreased PF reflex threshold angle, and increased toe 142

tapping ability. We compared TA↑ and SOL↓ interventions and assessed 15 outcomes : 1) two 143

measures of ankle clonus (clonus duration, PF reflex threshold angle); 2) two measures of ankle 144

motor control (toe tap test, DF active ROM); 3) four measures of lower extremity strength (DF 145

strength, PF strength, [training and non-training leg]); 4) three walking-related measures (foot 146

clearance, speed, and distance); and 5) four neurophysiologic measures (SOL H-reflex RI, PI, 147

and LFD, and SOL/TA clonus coactivation ratio). Secondarily, we also explored relationships 148

among changes in outcome measures to determine if neurophysiological mechanisms are 149

associated with changes in training, clinical, or walking measures. 150

METHODS 151

Research participants 152

Of sixteen volunteers recruited from The Miami Project to Cure Paralysis research 153

subject registry; twelve signed written consent and were enrolled in the study (Figure 1). 154

Participants were 16 to 75 years old, with stable motor-incomplete SCI (> 1 year), lesion level of 155

T12 or above, and ability to walk 6 meters with devices or assistance as needed. All participants 156

exhibited an American Spinal Injury Association (ASIA) Impairment Scale D (AIS D) category 157

of SCI (ASIA 2002; Marino et al. 2003), and in at least one lower extremity had a palpable TA 158

contraction, a SOL H-reflex response, and ankle clonus elicited by manual plantar flexor stretch. 159

Participants were without other conditions impairing walking function, or cognitive deficits that 160

would impede adherence to the study protocol. Participants were stratified according to toe tap 161

score (toe taps in 10 seconds) into one of two categories indicating extent of voluntary ankle 162


7

control: low control (< 15 taps) and high control (≥ 15 taps) (Figure 1), with each training 163

session lasting approximately 2 hours including set-up time. Following stratification, 164

participants were randomly assigned to the TA↑ group (n = 6, mean age = 44.2, 6 men) or the 165

SOL↓ group (n = 6, mean age = 45.2, 4 men) (Table 1). We trained the ankle with most severe 166

clonus as assessed by clonus duration after manual plantar flexor stretch. Training was 167

conducted three times weekly for five weeks for a total of 15 sessions (3 baseline and 12 training 168

sessions) (Figure 1). Testing was performed the week before and the week after training. All 169

twelve participants completed the study (mean age = 44.6 years, 10 men). The study was 170

conducted in accordance with the standards of the Declaration of Helsinki and was approved by 171

the University of Miami Human Subjects Research Office. 172

<< INSERT FIGURE 1 AND TABLE 1 ABOUT HERE >> 173

Clinical Measures of Ankle Clonus, Motor Control, and Walking 174

Ankle clonus 175

The Ankle Clonus Drop Test (Manella and Field-Fote 2009) provides reliable ankle 176

clonus measures that correlate with the Spinal Cord Assessment Tool for Spastic reflexes clonus 177

score (Benz et al. 2005) and SOL H/M ratio (Manella and Field-Fote 2009). Briefly, with the 178

individual sitting, the foot was placed on a 10-cm high platform, the leg was lifted to a height of 179

10 cm above the platform and released; impact of the forefoot with the platform edge provided a 180

SOL stretch to elicit clonus (Figure 2). Kinematic ankle angles and SOL and TA EMG activity 181

were recorded for three trials. Clonus duration was measured from first PF deflection after 182

platform impact to last PF deflection of at least 1 degree greater than the rest angle. The PF 183

reflex threshold angle (Manella and Field-Fote 2010) was identified kinematically as first PF 184


8

deflection after platform impact. Mean clonus duration and PF reflex threshold angle were 185

calculated. 186

<< INSERT FIGURE 2 ABOUT HERE >> 187

Ankle motor control 188

Timed toe tapping (Knights and Moule 1967) was performed seated with the ball of the 189

foot positioned on a pressure sensitive footswitch embedded in a platform. Dorsiflexion 190

activated the footswitch that triggered data acquisition (Spike 2 v7, Cambridge Electronic 191

Design, Cambridge, England). With the heel maintaining platform contact, the participant 192

performed voluntary concentric and eccentric dorsiflexor contractions as quickly as possible for 193

10 seconds for four trials with 60-second rest periods between trials. Mean number of taps 194

achieved in the best 3 of 4 trials was calculated. 195

Voluntary DF active ROM was recorded using an 8-camera Peak Motus 8.5 motion 196

analysis system (Vicon, Centennial, CO, USA). Reflective markers were placed laterally on the 197

5th metatarsal head, malleolus, heel, and knee. In sitting, hip and knee were flexed to 90° and 198

30°, respectively, with the lower leg supported allowing unrestricted ankle motion. Three trials 199

of maximal voluntary dorsiflexion and 5 seconds of relaxation were performed; mean maximum 200

dorsiflexion angle was calculated. 201

Bilateral lower extremity strength was assessed using ASIA lower extremity motor scores 202

(LEMS) (ASIA 2002; Marino et al. 2003). With the subject supine, hip flexors, knee extensors, 203

dorsiflexors (DF), great toe extensor, and plantar flexors (PF) were graded from 0 (no palpable 204

contraction) to 5 (maximal resistance against gravity); LEMS for DF, PF, training leg and non-205

training leg were recorded. 206

207


9

Walking function 208

Foot clearance and speed were analyzed using an 8-camera Peak Motus 8.5 motion 209

analysis system (Vicon, Centennial, CO, USA); data were sampled at 60 Hz, smoothed (zero-lag 210

Butterworth filter) and low-pass filtered (6 Hz). Twenty-one reflective markers were placed 211

bilaterally at lateral 5th metatarsal head, malleolus, heel, and knee, and greater trochanter, 212

anterior superior iliac spine, posterior wrist, lateral elbow and shoulder; and at C7, T10 and the 213

sacrum. Participants, wearing a safety harness, walked with assistance and/or devices as needed 214

at their fastest comfortable speed over a 6-meter walkway for three trials. Foot clearance, 215

defined as maximal vertical excursion of the 5th metatarsal marker, was quantified for three 216

swing phases for each trial. Walking speed was calculated for the central 4 meters of each trial. 217

Mean foot clearance and speed were computed. Distance traversed during a 2-minute period was 218

determined by participants walking, with assistance and/or devices as needed, at fastest 219

comfortable speed around an 80-ft (24.39 m) track. 220

Neurophysiologic Tests 221

SOL H-reflexes (peak-to-peak amplitude, mV) in response to RI, PI, and LFD 222

conditioning stimuli were recorded using surface electrodes (interelectode distance, 36 mm) 223

positioned as described by Rainoldi et al. (2004). SOL electrodes were placed at mid-line, 10 cm 224

proximal to the calcaneus. Instrumentation included a constant current stimulator (Digitimer 225

DS7A , Digitimer Ltd, Hertfordshire, England), AC amplifier (Grass P511, Grass Technologies, 226

West Warwick, RI, USA), and analog-to-digital converter (CED 1401, Cambridge Electronic 227

Design, Cambridge, England). Inter-stimulus intervals (ISIs) were controlled with an electrical 228

stimulator (Grass S88, Grass Technologies, West Warwick, RI, USA). A peripheral nerve 229

stimulator (MiniStim MS-IVA, Life-Tech Inc, Stafford, TX, USA) was used to locate the 230


10

posterior tibial nerve (PTN) in the popliteal fossa, common peroneal nerve (CPN) distal to the 231

fibular head, and femoral nerve (FN) in the femoral triangle. Cathodes were placed at the nerve 232

site where the largest muscle contraction was evoked with lowest stimulus intensity; anodes were 233

placed on the patella medially for PTN, laterally for CPN, and superiorly for FN stimulation. 234

Stimulus intensity (mA) was maintained between 10 – 30% Mmax (Crone et al. 1990). 235

Reciprocal Ia inhibition test 236

For testing RI of the SOL H-reflex (Crone et al. 1987), TA electrodes were placed on the 237

muscle belly, lateral to tibial crest and at mid-point between tibial tuberosity and inter-malleolar 238

line. Twenty interleaved test (PTN) and conditioned (CPN-PTN, 2 ms ISI) reflexes were evoked 239

at 0.10 Hz. TI testing was repeated using CPN-PTN conditioning stimuli at 3 ms ISI. The 240

conditioned reflex amplitude was normalized to the test amplitude and expressed as percent (%) 241

of control reflex; for each ISI interval the mean RI% was calculated and smallest mean RI% was 242

selected. 243

Presynaptic inhibition test 244

PI of the SOL H-reflex was assessed using heteronymous SOL H-reflex facilitation (Faist 245

et al. 1994). Electrodes were placed on quadriceps of the training leg at mid-line on the muscle 246

belly, 15 cm proximal to the patella. A 0.5 ms PTN test pulse was administered prior to a 0.1 ms 247

FN conditioning pulse. FN stimulus intensity was 4X motor threshold. The ISI at onset of SOL 248

H-reflex facilitation was identified and increased by 0.4 ms. Twenty interleaved test and 249

conditioned reflexes were evoked at 0.10 Hz. The difference between test and conditioned reflex 250

amplitudes was expressed as a percent of Mmax. Values ranged from 2 – 18%, consistent with 251

Faist et al. (1994). The largest PI% was selected for analysis. 252

253


11

Low-Frequency Depression test 254

SOL H-reflexes were evoked using paired-pulse PTN stimuli with ISI of 1 second. Ten 255

trials with 10-second rests between trials were recorded. The amplitude of the conditioned 256

(second) response was expressed as % of the control (first) reflex amplitude (LFD%); mean 257

LFD% was calculated (Field-Fote et al. 2006). 258

Coactivation during clonus 259

SOL and TA EMG activity were recorded during the ankle clonus test (described above). 260

The coactivation ratio of SOL and TA EMG bursts (± 2SD mean resting EMG) during a 2-261

second window of clonus was calculated using customized software (MATLAB, The 262

MathWorks, Inc., Natick, MA, USA) (Manella and Field-Fote 2009). Mean SOL/TA clonus 263

coactivation ratio for three trials was calculated. 264

Training Procedures 265

Surface recording electrodes were placed after standard skin preparation. For both 266

interventions, the participant was seated with approximately 100° hip flexion, 70° knee flexion 267

and 120° ankle plantar flexion; the training foot was positioned on a footswitch imbedded in a 268

platform. A shoe was worn if footswitch activation was achieved with it on; if not then a sock 269

and shower slipper were used. An auditory signal cued the participant to perform DF at 0.10 Hz 270

for 30 repetitions; 10 bouts of 30 repetitions were performed with 60-second rest intervals 271

between each bout. During baseline testing no visual or auditory feedback was provided. 272

During training EMG biofeedback and a visual tally of the number of repetitions s that met or 273

exceeded target during each bout were displayed (methods for determining target TA↑ and 274

SOL↓ EMG are described below in the respective sections). Immediate verbal reward of “good 275


12

job” or similar accolade was given and faded to summary feedback upon 80% success rate 276

(number of repetitions exceeding target/total repetitions) (Figure 1). 277

TA↑ 278

TA electrodes were placed as described previously. EMG signals were amplified (x2k) 279

(Grass amplifier, Grass Technologies, West Warwick, RI, USA), notch filtered (60 Hz), 280

bandpass filtered (10 Hz - 1000 Hz), sampled at 1k Hz, full-wave rectified, and smoothed (RMS, 281

8 ms moving average filter). The footswitch was activated by dorsiflexion, and triggered a TTL 282

input to an analog-to-digital converter (CED 1401, Cambridge Electronic Design, Cambridge, 283

England). For each participant, at session start, three 3-second TA maximal voluntary 284

contractions (MVCs) were performed with 60-second rest intervals. Mean MVC peak amplitude 285

was computed and %MVC amplitude target for the session was identified. TA EMG 286

biofeedback using customized software (Spike 2, v7, Cambridge Electronic Design, Cambridge, 287

England) was displayed on a split screen (Figure 3A). On the right, the %MVC amplitude target 288

was displayed (horizontal line), a vertical bar moved up or down reflecting increasing or 289

decreasing TA EMG activity. Instructions were to dorsiflex the foot and move the TA EMG bar 290

above the target line; when the target was exceeded the screen flashed green. On the left side of 291

the screen, reward, EMG, and footswitch activations were displayed; a gold icon appeared when 292

TA EMG activity exceeded target. For training session 1, the mean %MVC amplitude of the 293

prior baseline session was used as the training target. For training sessions 2 – 12, each %MVC 294

amplitude target was based on prior session performance; for session success rates of at least 295

50%, the target was increased 10% for the next session. 296

297

298


13

SOL↓ 299

Footswitch activation, EMG recording and signal amplification equipment and 300

parameters were as described above for TA↑. During training, SOL H-reflexes during brief 301

voluntary DF contractions were recorded as described by Crone et al. (1987), Crone and Nielson 302

(1989), Morin and Pierrot-Deseilligny (1977), and Pierrot-Deseilligny and Burke (2005). 303

Electrodes were placed as described earlier for PTN stimulation and SOL H-reflex recording and 304

the foot was positioned on the platform over the embedded footswitch. The footswitch was 305

connected to a stimulator (Grass S88, Grass Technologies, West Warwick, RI, USA) that 306

provided a 50-ms delay between footswitch activation and SOL H-reflex stimulation. This delay 307

reflected the time course of depression of SOL H-reflex by voluntary DF, which increases 308

greatly from 50 – 100 ms following movement onset in non-disabled individuals (Crone et al. 309

1987). For each participant, at session start, SOL H-reflex and M-wave recruitment curves at 310

rest and during voluntary DF were obtained by incrementally increasing the stimulus intensity by 311

2 mA; H-reflex and M-wave maximum amplitudes were recorded. PTN stimulus intensity was 312

maintained at 20% of M-wave maximum amplitude during DF throughout each session. 313

SOL H-reflex biofeedback using customized software (Signal, Cambridge Electronic 314

Design, Cambridge, England) was presented on a split screen (Figure 3B). On the right, the SOL 315

H-reflex amplitude (mV) target was displayed (horizontal line). A black square, representing the 316

peak-to peak amplitude of the SOL H-reflex response, appeared above or below the target line 317

for responses greater or smaller than the target, respectively. Instructions were to dorsiflex 318

(which triggered SOL H-reflex stimulation) and generate a black square (reflex amplitude) below 319

the target line; the screen flashed green for responses smaller than target amplitude. On the left 320

side of the screen, the SOL H-reflex signal was displayed (Figure 3B). For each training session, 321


14

the first bout was a baseline bout (no feedback) in which the median SOL H-reflex amplitude 322

target was calculated to determine the training target for the second bout; for subsequent training 323

bouts the median SOL H-reflex target was calculated from the previous training bout. 324

Participants were encouraged to employ any strategy to accomplish the task while remaining 325

seated with the foot on the platform and footswitch. 326


Data analysis 328

Non-parametric statistical tests were conducted using SPSS Statistics 18 software (SPSS, 329

Inc., Chicago, IL, USA) because the assumption of normal distribution required for use of 330

parametric test was not met due to small sample size. Mann-Whitney U or Chi-square statistics 331

were used to compare demographic and pre-intervention clinical characteristics for each 332

intervention group. Spearman correlation coefficients were used to determine pre-intervention 333

relationships between outcome measures. 334

Training dosage was compared between TA↑ and SOL↓ groups using Mann-Whitney U-335

test. For each group, the magnitude of change in the training measure was examined with one-336

tailed Wilcoxon signed ranks test. For the SOL↓ group, only SOL H-reflexes that were elicited 337

with stimulus intensities between 10 and 30% M-max were included in data analysis, to ensure 338

equivalence of stimuli across trials. Effect size was assessed with standardized response mean 339

(SRM) (Liang et al. 1990), which divides mean pre-post intervention change by standard 340

deviation of the change. SRM interpretation, based on Cohen’s criteria, was 0.20, 0.50 and 0.80 341

for small, moderate, and large effects, respectively (Liang et al. 1990). For TA↑, mean %MVC 342

amplitude was compared between three baseline sessions and the final three training sessions. 343

For SOL↓, stability of SOL Mmax across sessions was assessed using repeated measures 344


15

ANOVA; SOL Mmax values were obtained from the recruitment curve for each session and 345

were found to be stable (p = 0.24). Mean SOL H-reflex amplitude during dorsiflexion was 346

calculated for the last three training sessions (conditioned SOL HDF

) and was normalized to the 347

mean SOL HDF

for three baseline sessions and expressed as percent (%) of baseline (Thompson 348

et al. 2009). In four participants, the mean SOL HDF

value for baseline session 3 was identified 349

as an outlier (> 2 SD of mean of baseline sessions 1 and 2) and was removed from analysis. 350

Change was determined by subtracting the pretest from the posttest measure. One-tailed 351

Mann-Whitney U-tests were used to compare magnitude of change, between TA↑ and SOL↓ 352

groups for each outcome measure. Within each group, some outcome measures exhibited 353

change that appeared to be clinically relevant; therefore, for each group separately, we analyzed 354

change in outcome measures using one-tailed Wilcoxon signed ranks tests, calculated SRM 355

(Liang et al. 1990) for each outcome, and explored relationships among outcome changes with 356

one-tailed Spearman correlation coefficients. 357

RESULTS 358

Pre-intervention characteristics 359

The TA↑ and SOL↓ groups were similar in pre-intervention demographic, clinical, and 360

neurophysiologic measures. However, for clinical measures, the TA↑ group had greater DF 361

active ROM (96.9° vs. 112.4°, p = 0.04) and non-training leg strength (22.5 vs. 19.2 LEMS, p = 362

0.04) (Table 1). Stratification of participants by toe tap score resulted in equal distribution of 363

participants between the two groups based on level of ankle motor control; each group had four 364

individuals with low control (<15 taps) and two with high control (≥ 15 taps). 365

Change in training measures and effect size (standardized response mean) 366


16

Training dosage (repetitions) was similar between TA↑ and SOL↓ groups (3100 vs. 2740, 367

p = 0.21). For TA↑, TA%MVC amplitude increased between baseline (43% ± 17) and final 368

training sessions (91% ± 18, p = 0.01) (Table 2) and a large training effect was observed (SRM = 369

5.19). For SOL↓, the conditioned SOL HDF

% resulted in a decrease from baseline that 370

approached statistical significance (-16.17 % ± 26.22, p = 0.09) (Table 2) and a moderate 371

training effect was observed (SRM = -0.62). SOL HDF

-reflex responses during training sessions 372

2 and 11 for an example participant are illustrated in Figure 4. 373

<< INSERT FIGURE 4 AND TABLE 2 ABOUT HERE >> 374

Change in clinical measures and effect size (standardized response mean) 375

Change in PF reflex threshold angle was the only clinical measure for which there was a 376

between-groups difference that approached statistical significance (Table 2). This angle 377

decreased toward more dorsiflexion for TA↑ (-4.33°) and did not change significantly for SOL↓ 378

(+3.15°) (Table 2). Within-group analysis of data from the TA↑ group identified five clinical 379

measures that achieved statistical significance and had a large effect size. These outcomes were: 380

decreased PF reflex threshold angle (-4.33°) and DF active ROM angle (-4.32°), and increased 381

dorsiflexion LEMS (+0.8 points), walking foot clearance (+4.8 mm), and walking distance 382

(+12.09 m) (Table 2). In addition, an increase in LEMS for training leg (+2.2 points) and non-383

training leg (+0.8 points) approached statistical significance and demonstrated a large effect size 384

(Table 2). Within-group analysis of data from the SOL↓ group identified one clinical measure, 385

increased walking distance (+6.25 m), achieved statistical significance and had a large effect size 386

(Table 2). In addition, two clinical measures approached statistical significance: increased non-387

training leg LEMS (+1.8 points), which demonstrated a large effect size, and increased walking 388

speed (+0.02 m/s), which demonstrated a moderate effect size (Table 2). A power analysis of 389


17

the training and clinical outcomes that approached statistical significance revealed that our study 390

was considerably underpowered to detect change in the following measures: SOL HDF% (31%), 391

SOL and TA non-training LEMS (33%), TA non-training LEMS (20%), SOL walking speed 392

(14%), and between-groups difference for PF reflex threshold angle (34%). 393

Change in neurophysiologic measures and effect size (standardized response mean) 394

The only neurophysiologic measure in which we observed a significant change was 395

SOL/TA clonus coactivation ratio, which decreased (-0.21) for the SOL↓ group and exhibited a 396

moderate effect size (Table 2). 397

Relationship among training, clinical, and neurophysiologic outcomes 398

TA↑ 399

We had hypothesized TA↑, which was intended to activate SOL RI and PI pathways, 400

would be associated with increased DF active ROM and strength, walking foot clearance, speed, 401

and distance. We did not find a significant change in SOL RI, PI, or walking speed. However, 402

increased walking distance was moderately correlated with increased DF (r = 0.66, p = 0.08), and 403

training leg strength (r = 0.62, p = 0.10) (Figure 5). 404


SOL↓ 406

We had hypothesized SOL↓ would be associated with increased SOL H-reflex LFD and 407

decreased SOL/TA clonus coactivation that would be related to decreased clonus duration, 408

decreased PF reflex threshold angle, and enhanced toe tap count. We did not find significant 409

change in SOL LFD, clonus duration, PF reflex threshold angle, or toe tap count. However, 410

decreased SOL/TA clonus coactivation ratio (-0.21, p = 0.02) was moderately correlated with 411

increased walking speed and distance (both r = -0.60, p = 0.10) (Figure 6). 412


18


DISCUSSION 414

We investigated whether operant conditioning to either increase voluntary TA activation 415

or decrease SOL H-reflexes improved ankle clonus, motor control, or walking function in 416

individuals with AIS D SCI. We also explored relationships among change in training, clinical, 417

and neurophysiologic outcomes. Our hypotheses that TA↑ would increase DF active ROM and 418

strength, walking foot clearance, and walking distance were supported; however, improvement 419

in walking speed was not attained. In addition, an unexpected outcome of decreased PF reflex 420

threshold angle was observed. For SOL↓ our hypothesis of decreased SOL/TA coactivation was 421

supported; however, decreased clonus duration and PF reflex threshold angle, and increased toe 422

tap count were not observed. Unexpectedly, SOL↓ was associated with increased walking 423

distance. Only one measure, PF reflex threshold angle, differed substantially between the two 424

groups. Our findings provide evidence that both TA↑ and SOL↓ were associated with significant 425

improvement in distance walked during a 2-minute period. Each training protocol may have 426

modulated clinical and neurologic variables in unique ways. 427

Change in Training Measures 428

TA↑ was associated with an increase in TA contraction strength (%MVC amplitude) of 429

approximately 48% (Figure 5), a finding consistent with prior reports of increased voluntary 430

EMG responses after such training (Brucker and Bulaeva 1996; Seymour and Bassler 1977). 431

SOL↓ was associated with a decrease in SOL H-reflex excitability (conditioned SOL HDF

%) of 432

about 16% that approached statistical significance (Figure 6). This result concurs with 433

Thompson et al. (2008) who reported a 10 – 15% reduction after down-conditioning of SOL H-434

reflex in standing for three individuals with SCI. Our mean training dosage was approximately 435


19

3000 repetitions (300 X 12 training sessions) for TA↑ and SOL↓ and was less than the dosage of 436

5400 repetitions used in other studies (Birkenmeier et al. 2010; Thompson et al. 2009). Our 437

results are similar to Thompson et al. (2009) who used 225 repetitions per session and reported a 438

significant decrease in the conditioned SOL H-reflex within sessions 7 – 12 that continued to 439

decrease through session 24; change the investigators attributed to early phase task-dependent 440

adaptation. Thompson et al. (2009) also identified a significant change in the control SOL H-441

reflex within sessions 19 – 24 that was attributed to a long-term change in spinal cord plasticity. 442

In our participants with spinal cord injury, we observed a decrease in conditioned SOL H-reflex 443

that approached significance after 12 sessions; an additional six sessions (total of 5400 444

repetitions) may have improved the effectiveness of SOL↓ training. Further study is needed to 445

assess the most effective dosage-response parameters. 446

Effects of TA↑ and SOL↓ on Clinical Measures of Ankle Motor Control 447

Change in PF reflex threshold angle was the only outcome that differed substantially 448

between the two groups. TA↑ exhibited a decreased angle (i.e., towards more dorsiflexion, 449

indicating a more lengthened plantar flexor position at which stretch of the soleus evoked a 450

reflex contraction), signifying decreased responsiveness to muscle stretch and therefore 451

decreased spasticity (Ness and Field-Fote 2009). TA↑ was effective in enhancing TA activation, 452

as exhibited by increased DF active ROM and strength. In addition, increased DF strength was 453

associated with increased walking distance (Figure 5). Functional magnetic resonance imaging 454

studies suggest that 10° of voluntary ankle dorsiflexion is a useful measure of motor control of 455

walking that in depends, in part, upon increased TA activation (Dobkin et al. 2004). 456

In non-disabled individuals, PI of SOL motor neurons increases 60 – 80 ms after onset of 457

voluntary TA EMG activation, and the extent of inhibition is proportional to TA contraction 458


20

strength (Meunier and Morin 1989). Our findings for TA↑ of increased DF active ROM and 459

strength are congruent with a supraspinally-mediated enhancement of SOL PI. However, our 460

findings of no change in PI are in conflict with a mechanism of enhanced SOL PI. A possible 461

explanation is that SOL PI was tested at rest, which may not reflect modulation of the PI 462

pathway during active dorsiflexion. 463

TA↑ was associated with an increase in both training leg and non-training leg strength 464

that approached statistical significance (Figure 5), while SOL↓ was associated with an increase 465

in non-training leg strength that approached statistical significance (Figure 6). The finding of 466

increased strength in the non-training leg for both groups was unexpected. For TA↑ increased 467

DF strength and training leg strength were correlated with increased walking distance (Figure 5). 468

We surmise that TA↑ improved the ability to isolate and activate dorsiflexors in the weaker 469

training leg, which likely contributed to the ability to walk farther distances. Participants were 470

not restricted from walking outside of training sessions and reported the ability to walk farther 471

distances during their daily activities, which likely provided an opportunity for re-conditioning 472

of the non-training leg. Our finding of increased non-training leg strength is also congruent with 473

studies of contralateral strength training, wherein increased motor neuron output, increased 474

strength, and decreased H-reflex gain in the untrained homonymous muscle have been reported 475

(Dragert and Zehr 2012, Carroll et al 2006, Hortobagyi 2005). Wolf et al 1995 reported a 476

contralateral effect after down-training of biceps brachii stretch reflexes in human subjects 477

wherein contralateral responses were reduced to 80% of baseline. The authors suggested that 478

both intra-spinal mechanisms and bilateral descending systems may contribute to this 479

contralateral training effect. 480

481


21

Effects of TA↑ and SOL↓ on Walking Function 482

Both TA↑ and SOL↓ were associated with increased walking distances (Table 2) that 483

exceeded the minimally important difference of 4 meters reported for the 2-minute walk test in 484

individuals with motor-incomplete SCI (Field-Fote and Roach 2011). The cumulative effects of 485

TA↑ (decreased PF reflex threshold angle and increased DF active ROM, DF and leg strength, 486

foot clearance and walking distance) (Figure 5) are consistent with a mechanism of increased 487

corticospinal activation. Upon initiation of voluntary DF, the corticospinal tract activates TA 488

motor neurons in parallel with activation of interneurons mediating PI and RI of SOL motor 489

neurons (Iles and Pisini 1992). Our findings are consistent with studies of locomotor training 490

effects in chronic motor-incomplete SCI that report increased voluntary activation of TA and 491

slope of TMS-elicited TA motor evoked potential recruitment curves; findings that provide 492

evidence of a training-induced increase in corticospinal excitability (Thomas and Gorassini 493

2005). 494

For SOL↓, decreased SOL/TA coactivation was associated with increased walking 495

distance and increased walking speed that approached statistical significance (Figure 6). 496

However, the increase in speed did not exceed the minimally important difference of 0.05 m/s 497

reported for individuals with SCI (Musselman 2007). The SOL↓ effects are consistent with 498

improved SOL reflex modulation during walking. Both excitatory and inhibitory Ib interneuron 499

pathways to ankle extensors are easily activated by the corticospinal tract during voluntary 500

movement (Jankowska 1992), and appear to regulate stance-to-swing phase transition during 501

locomotion (Pearson et al. 1992). The cumulative effects of SOL↓ may be attributable to a Ib 502

neuroplastic effect of SOL↓ that may have contributed to improved walking function (Figure 6). 503

504


22

Effects of TA↑ and SOL↓ on Neurophysiologic Measures 505

The only substantial neurophysiologic effect, exhibited by SOL↓, was decreased SOL/TA 506

clonus coactivation from 52% to 31% that was associated with the ability to walk faster and 507

farther (Figure 6). These findings may suggest that reduced SOL/TA coactivation during clonus 508

was associated with improved SOL↓ walking outcomes. 509

Prior studies have described adverse antagonist coactivation and impaired SOL RI and PI 510

after SCI (Boorman et al. 1996; Crone and Nielsen, 1994; Crone et al. 2003; Dietz 2001; Faist et 511

al. 1994; Morita et al. 2001; Xia and Rymer 2005). We observed decreased SOL/TA 512

coactivation after SOL↓ training; however, no significant changes in SOL RI or PI were 513

identified. In individuals with PF hyperexcitability, Crone et al. (2003) reported reciprocal 514

facilitation of plantar flexors after CPN stimulation (as opposed to reciprocal inhibition) and 515

proposed that activation of disynaptic Ib pathways may contribute to adverse antagonist 516

cocontraction. Our finding of a predominant pattern of SOL/TA coactivation during clonus is 517

consistent with Xia and Rymer (2005) who reported altered agonist and antagonist activation 518

after SCI as a consequence of reciprocal facilitation of TA during SOL tendon taps, as well as 519

response latencies that are consistent with oligosynaptic short-latency pathways. Citing animal 520

studies, these investigators proposed Ib afferent excitation as a possible source of reciprocal 521

facilitation after SCI (Xia and Rymer 2005). 522

Recently it has been postulated that increased facilitation or suppressed inhibition of the 523

antagonist muscle following perturbation of the agonist involves Ib force-sensitive pathways 524

(Lewis et al. 2010). Yanagawa et al. (1991) observed that weak TA contractions increased Ib 525

inhibition of SOL H-reflexes and concluded that Ib inhibition, which arrived prior to spinal-level 526

TA-induced inhibitory effects on SOL Ia reflex, was corticospinally activated. We surmise that 527


23

decreased coactivation of SOL and TA during clonus following SOL↓ (which employed weak 528

TA contractions) may be attributable to enhanced Ib inhibitory mechanisms. 529

Conclusion and Functional Implications 530

Two operant conditioning training programs, one to increase voluntary dorsiflexor motor 531

control and the other to decrease plantar flexor stretch reflex excitability, were both associated 532

with improved walking function in individuals with chronic motor-incomplete spinal cord injury. 533

TA↑ decreased plantar flexor spasticity, increased ankle motor control and was associated 534

with increased walking foot clearance and walking distance. Step initiation is mediated by 535

supraspinal input to spinal interneurons that activate stepping (Gerasimenko et al. 2002) and the 536

corticospinal tract is more closely linked with TA and leg flexor spinal motor circuits (Capaday 537

et al. 1999; Schubert et al. 1997). Intensive, repetitive TA EMG activation during TA↑ may 538

have unmasked dormant corticospinal pathways that preferentially increased recruitment of TA 539

and leg flexor motor neurons, step initiation and walking function. 540

SOL↓ was associated with decreased SOL/TA coactivation during clonus and increased 541

walking distance. Antagonist coactivation modulation has been attributed to Ib input to spinal 542

interneurons (Feldman 1993, Morita et al. 2006, Yanagawa et al. 1991). It is possible that 543

intensive training to inhibit SOL H-reflexes during weak voluntary TA contractions enhanced 544

corticospinal activation of SOL Ib interneurons, and the combined effects of decreased SOL/TA 545

clonus coactivation and increased SOL stretch reflex inhibition improved ankle motor control 546

and walking function. 547

Our study was considerably underpowered to detect change in SOL HDF%, TA training 548

LEMS, TA and SOL non-training LEMS, and SOL walking speed. However, change in these 549

measures did approach statistical significance, and since the minimal clinically important 550


24

difference (MCID) for most of these measures is not known, it is possible that the changes 551

observed may be clinically relevant. The MCID for walking speed is 0.05m/s, the mean change 552

of 0.4m/s in the TA↑ approached this value. There were subjects in both groups that met the 553

criteria for clinically meaningful improvement in walking speed. We included these outcomes in 554

our discussion and interpretation of results to elucidate clinically relevant relationships among 555

change in training and clinical measures. 556

LIMITATIONS 557

Our two-group pre-test post-test randomized design allowed us to compare two different 558

interventions. Our study design lacked a no-treatment control group, which limits interpretation 559

of whether either training approach was more beneficial than no intervention (Portney and 560

Watkins 2000). However, given the chronicity of SCI in our participants, little change in motor 561

function and reflex activity would be expected in the absence of an intervention. We did not 562

monitor or control for background SOL EMG during SOL↓ training, therefore the effects of SOL 563

EMG activity on modulation of the SOL H-reflex are not known, which may have limited our 564

ability to fully assess the extent of inhibition. Additionally, the neurophysiologic tests for SOL 565

RI, PI, and LFD were performed at rest. Performing these tests during dorsiflexion may have 566

better elucidated modulation of these pathways during movement. Our conclusions are restricted 567

to outcomes observed in a small sample of individuals with AIS D SCI. Further research is 568

warranted to corroborate effects of TA↑ and SOL↓ in AIS D SCI and generalizability to 569

individuals with other similar upper motor neuron pathology and functional limitations. 570

ACKNOWLEDGEMENTS 571

The authors thank Simon Gray for software design, Ian Hentall, PhD, for footswitch 572

design, Meggie Safford, DPT, and Rachel Cowen, PhD, for data processing assistance, the 573


25

support of the Miami Project to Cure Paralysis, the contributions of Christine K. Thomas, PhD, 574

and T. George Hornby, PT, PhD to manuscript development, and the time and effort volunteered 575

by our research participants with spinal cord injury. 576

. 577

578

579

580


26

REFERENCES 581

American Spinal Injury Association. Reference Manual of the International Standards for 582

Neurological Classification of Spinal Cord Injury Chicago, IL: American Spinal Injury 583

Association, 2002. 584

Bakker M, Overeem S, Snijders AH, Borm G, van Elswijk G, Toni I, Bloem BR. Motor 585

imagery of foot dorsiflexion and gait: effects on corticospinal excitability. Clin Neurophysiol 586

119: 11: 2519-2527, 2008. 587

Beekhuizen KS, Field-Fote EC. Massed practice versus massed practice with stimulation: 588

effects on upper extremity function and cortical plasticity in individuals with incomplete 589

cervical spinal cord injury. Neurorehabil Neural Repair 19: 1: 33-45, 2005. 590

Benz EN, Hornby TG, Bode RK, Scheidt RA, Schmit BD. A physiologically based clinical 591

measure for spastic reflexes in spinal cord injury. Arch Phys Med Rehabil 86; 52-59, 2005. 592

Beres-Jones JA, Johnson TD, Harkema SJ. Clonus after human spinal cord injury cannot be 593

attributed solely to recurrent muscle-tendon stretch. Exp Brain Res 149: 2: 222-236, 2003. 594

Birkenmeier RL, Prager EM, Lang CE. Translating animal doses of task-specific training to 595

people with chronic stroke in 1-hour therapy sessions: a proof-of-concept study. Neurorehabil 596

Neural Repair 24: 7: 620-635, 2010. 597

Boorman GI, Lee RG, Becker WJ, Windhorst UR. Impaired "natural reciprocal inhibition" in 598

patients with spasticity due to complete spinal cord injury. Electromyogr Clin Neurophysiol 599

101: 84-92, 1996. 600

Brucker BS, Bulaeva NV. Biofeedback effect on electromyography responses in patients with 601

spinal cord injury. Arch Phys Med Rehabil 77: 2: 133-137, 1996. 602


27

Carroll TJ, Herbert RD, Munn J, Lee ML, Gandevia SC. Contralateral effects of unilateral 603

strength training: evidence and possible mechanisms. J Appl Physiol 101: 1514-1522, 2006. 604

Crone C, Hultborn H, Jespersen B, Nielsen J. Reciprocal Ia inhibition between ankle flexors 605

and extensors in man. J Physiol 389: 163-185, 1987. 606

Crone C, Hultborn H, Mazieres L, Morin C, Nielsen J, Pierrot-Deseilligny E. Sensitivity of 607

monosynaptic test reflexes to facilitation and inhibition as a function of the test reflex size: a 608

study in man and the cat. Exp Brain Res 81: 1: 35-45, 1990. 609

Crone C, Johnsen LL, Biering-Sorensen F, Nielsen JB. Appearance of reciprocal facilitation 610

of ankle extensors from ankle flexors in patients with stroke or spinal cord injury. Brain 126: 611

Pt 2: 495-507, 2003. 612

Crone C, Nielsen J. Spinal mechanisms in man contributing to reciprocal inhibition during 613

voluntary dorsiflexion of the foot. J Physiol 416:255-272, 1989. 614

Crone C, Nielsen J. Central control of disynaptic reciprocal inhibition in humans. Acta Physiol 615

Scand 152: 4: 351-363, 1994. 616

Davies S, Illis LS, Raisman G. Regeneration in the central nervous system and related factors. 617

Summary of the Bermuda Paraplegia Conference, April 1994 (International Spinal Research 618

Trust). Paraplegia 33: 1: 10-17, 1995. 619

Decq P. Pathophysiology of spasticity. Neurochirurgie 49: 2-3 Pt 2: 163-184, 2003. 620

Dietz V. Spinal cord lesion: effects of and perspectives for treatment. Neural Plast 8: 1-2: 83-90, 621

2001. 622

Dietz V. Locomotor recovery after spinal cord injury. Trends Neurosci 20: 8: 346-347, 1997. 623

Dragert K, Zehr EP. High-intensity unilateral dorsiflexor resistance training results in bilateral 624

neuromuscular plasticity after stroke. Exp Brain Res Nov 30 2012 [Epub ahead of print]. 625


28

Evatt ML, Wolf SL, Segal RL. Modification of human spinal stretch reflexes: preliminary 626

studies. Neurosci Lett 105: 3: 350-355, 1989. 627

Faist M, Mazevet D, Dietz V, Pierrot-Deseilligny E. A quantitative assessment of presynaptic 628

inhibition of Ia afferents in spastics. Differences in hemiplegics and paraplegics. Brain 117 ( 629

Pt 6): Pt 6: 1449-1455, 1994. 630

Feldman AG. The coactivation command for antagonist muscles involving Ib interneurons in 631

mammalian motor control systems: an electrophysiologically testable model. Neurosci Lett 632

155: 2: 167-170, 1993. 633

Field-Fote EC. Electrical stimulation modifies spinal and cortical neural circuitry. Exerc Sport 634

Sci Rev 32: 4: 155-160, 2004. 635

Field-Fote EC, Brown KM, Lindley SD. Influence of posture and stimulus parameters on post-636

activation depression of the soleus H-reflex in individuals with chronic spinal cord injury. 637

Neurosci Lett 410: 1: 37-41, 2006. 638

Field-Fote EC, Roach KE. Influence of a locomotor training approach on walking speed and 639

distance in people with chronic spinal cord injury: a randomized clinical trial. Phys Ther 91: 640

1: 48-60, 2011. 641

Gerasimenko YP, Makarovskii AN, Nikitin OA. Control of locomotor activity in humans and 642

animals in the absence of supraspinal influences. Neurosci Behav Physiol 32: 4: 417-423, 643

2002. 644

Gracies JM. Pathophysiology of spastic paresis. II: Emergence of muscle overactivity. Muscle 645

Nerve 31: 5: 552-571, 2005. 646


29

Grey MJ, Klinge K, Crone C, Lorentzen J, Biering-Sorensen F, Ravnborg M, Nielsen JB. 647

Post-activation depression of soleus stretch reflexes in healthy and spastic humans. Exp Brain 648

Res 185: 2: 189-197, 2008. 649

Hortobagyi T. Cross education and the human central nervous system. IEEE Eng Med Biol Mag 650

24: 22-28, 2005. 651

Hultborn H. Changes in neuronal properties and spinal reflexes during development of 652

spasticity following spinal cord lesions and stroke: studies in animal models and patients. J 653

Rehabil Med (41 Suppl): 41 Suppl: 46-55, 2003. 654

Iles JF, Pisini JV. Cortical modulation of transmission in spinal reflex pathways of man. J 655

Physiol 455: 425-446, 1992. 656

Jankowska E. Interneuronal relay in spinal pathways from proprioceptors. Prog Neurobiol 38: 657

335-378, 1992. 658

Katz RT, Rymer WZ. Spastic hypertonia: mechanisms and measurement. Arch Phys Med 659

Rehabil 70: 2: 144-155, 1989. 660

Knights RM, Moule AD. Normative and reliability data on finger and foot tapping in children. 661

Percept Mot Skills 25: 3: 717-720, 1967. 662

Koelman JH, Bour LJ, Hilgevoord AA, van Bruggen GJ, Ongerboer de Visser BW. Soleus 663

H-reflex tests and clinical signs of the upper motor neuron syndrome. J Neurol Neurosurg 664

Psychiatry 56: 7: 776-781, 1993. 665

Levin MF, Hui-Chan C. Ankle spasticity is inversely correlated with antagonist voluntary 666

contraction in hemiparetic subjects. Electromyogr Clin Neurophysiol 34: 7: 415-425, 1994. 667


30

Lewis GN, MacKinnon CD, Trumbower R, Perreault EJ. Co-contraction modifies the stretch 668

reflex elicited in muscles shortened by a joint perturbation. Exp Brain Res 207: 1-2: 39-48, 669

2010. 670

Liang MH, Fossel AH , Larson MG. Comparisons of five health status instruments for 671

orthopedic evaluation. Med Care 28: 7: 632-642, 1990. 672

Lloyd DP , Wilson VJ. Reflex depression in rhythmically active monosynaptic reflex pathways. 673

J Gen Physiol 40: 3: 409-426, 1957. 674

Manella KJ , Field-Fote EC. Ankle Clonus Drop Test and angle of muscle reaction for 675

assessing plantar flexor spasticity after spinal cord injury. 2010 Combined Sections Meeting, 676

San Diego, CA: American Physical Therapy Association, 2010. Online. 677

http://www.neuropt.org/. 678

Manella KJ , Field-Fote EC. Biomechanical measures of ankle clonus correlate with clinical 679

and electrophysiologic reflex measures. Program No. 861.16. 2009 Neuroscience Meeting 680

Planner. Chicago, IL: Society for Neuroscience, 2009. Online. http://www.sfn.org/ 681

Marino RJ, Barros T, Biering-Sorensen F, Burns SP, Donovan WH, Graves DE, Haak M, 682

Hudson LM, Priebe MM , ASIA Neurological Standards Committee 2002. International 683

standards for neurological classification of spinal cord injury. J Spinal Cord Med 26 Suppl 1: 684

S50-6, 2003. 685

Meunier S, Morin C. Changes in presynaptic inhibition of Ia fibres to soleus motoneurones 686

during voluntary dorsiflexion of the foot. Exp Brain Res 76: 3: 510-518, 1989. 687

Morin C, Pierrot-Deseilligny E. Role of Ia afferents in the soleus motoneurones inhibition 688

during a tibialis anterior voluntary contraction in man. Exp Brain Res 27:509-522, 1977. 689


31

Morita H, Crone C, Christenhuis D, Petersen NT , Nielsen JB. Modulation of presynaptic 690

inhibition and disynaptic reciprocal Ia inhibition during voluntary movement in spasticity. 691

Brain 124: Pt 4: 826-837, 2001. 692

Morita H, Shindo M, Momoi H, Yanagawa S, Ikeda S , Yanagisawa N. Lack of modulation 693

of Ib inhibition during antagonist contraction in spasticity. Neurology 67: 1: 52-56, 2006. 694

Musselman K. Clinical significance testing in rehabilitation research: what, why, and how? 695

Phys Ther Rev 12: 287-296, 2007. 696

Ness LL , Field-Fote EC. Effect of whole-body vibration on quadriceps spasticity in individuals 697

with spastic hypertonia due to spinal cord injury. Restor Neurol Neurosci 27: 6: 621-631, 698

2009. 699

Pearson KG, Ramirez JM, Jiang W. Entrainment of the locomotor rhythm by group Ib 700

afferents from ankle extensor muscles in spinal cats. Exp Brain Res 90: 557-566, 1992. 701

Pierrot-Deseilligny E, Burke D. Monosynaptic Ia excitation and post-activation depression. In: 702

The Circuitry of the Human Spinal Cord. New York: Cambridge University Press, p. 63-106, 703

2005. 704

Portney LG, Watkins MP. Experimental Design. In: Foundations of Clinical Research: 705

Applications to Practice. New Jersey: Prentice Hall Health, p. 180, 2000. 706

Rainoldi A, Melchiorri G , Caruso I. A method for positioning electrodes during surface EMG 707

recordings in lower limb muscles. J Neurosci Methods 134: 1: 37-43, 2004. 708

Schindler-Ivens S, Shields RK. Low frequency depression of H-reflexes in humans with acute 709

and chronic spinal-cord injury. Exp Brain Res 133: 2: 233-241, 2000. 710

Schubert M, Curt A, Jensen L, Dietz V. Corticospinal input in human gait: modulation of 711

magnetically evoked motor responses. Exp Brain Res 115: 2: 234-246, 1997. 712


32

Segal RL. Plasticity in the central nervous system: operant conditioning of the spinal stretch 713

reflex. Top Stoke Rehabil 3: 76-87, 1997. 714

Seymour RJ , Bassler CR. Electromyographic biofeedback in the treatment of incomplete 715

paraplegia. Phys Ther 57: 10: 1148-1150, 1977. 716

Stein RB, Brucker BS , Ayyar DR. Motor units in incomplete spinal cord injury: electrical 717

activity, contractile properties and the effects of biofeedback. J Neurol Neurosurg Psychiatry 718

53: 10: 880-885, 1990. 719

Thomas SL, Gorassini MA. Increases in corticospinal tract function by treadmill training after 720

incomplete spinal cord injury. J Neurophysiol 94: 4: 2844-2855, 2005. 721

Thompson AK, Chen XY, Wolpaw JR. Acquisition of a simple motor skill: task-dependent 722

adaptation plus long-term change in the human soleus H-reflex. J Neurosci 29: 18: 5784-723

5792, 2009. 724

Thompson AK, Estabrooks KL, Chong SL, Stein RB. Spinal reflexes and ankle flexor and 725

extensor muscles after chronic central nervous system lesions and functional electrical 726

stimulation. Neurorehabil Neural Repair 23; 133-142, 2009. 727

Thompson AK, Stein RB, Chen XY , Wolpaw JR. Modulation in spinal circuits and 728

corticospinal connections following nerve stimulation and operant conditioning. Conf Proc 729

IEEE Eng Med Biol Soc 1: 2138-2141, 2006. 730

Thompson A, Wolpaw JR, Lichtman SW, Defrancesco E, Pomerantz F. Operant down-731

conditioning of soleus H-reflex in incomplete spinal cord injury. Program No. 73.4. 2008 732

Neuroscience Meeting Planner. Washington, DC: Society for Neuroscience, 2008. Online. 733

Wallace DM, Ross BH , Thomas CK. Motor unit behavior during clonus. J Appl Physiol 99: 6: 734

2166-2172, 2005. 735


33

Wolf SL, Segal RL. Reducing human biceps brachii spinal stretch reflex magnitude. J 736

Neurophysiol 75: 4: 1637-1646, 1996. 737

Wolf SL, Segal RL, Heter ND, Catlin PA. Contralateral and long latency effects of human 738

biceps brachii stretch reflex conditioning. Exp Brain Res 107: 96-102, 1995. 739

Wolpaw JR, Chen XY. The cerebellum in maintenance of a motor skill: a hierarchy of brain 740

and spinal cord plasticity underlies H-reflex conditioning. Learn Mem 13: 2: 208-215, 2006. 741

Wolpaw JR, Chen XY. Operant conditioning of rat H-reflex: effects on mean latency and 742

duration. Exp Brain Res 136: 2: 274-279, 2001. 743

Wolpaw JR, Herchenroder PA. Operant conditioning of H-reflex in freely moving monkeys. J 744

Neurosci Methods 31: 2: 145-152, 1990. 745

Xia R, Rymer WZ. Reflex reciprocal facilitation of antagonist muscles in spinal cord injury. 746

Spinal Cord 43: 1: 14-21, 2005. 747

Yanagawa S, Shindo M, Nakagawa S. Increase in Ib inhibition by antagonistic voluntary 748

contraction in man. J Physiol 440: 311-323, 1991. 749

Yang JF, Fung J, Edamura M, Blunt R, Stein RB , Barbeau H. H-reflex modulation during 750

walking in spastic paretic subjects. Can J Neurol Sci 18: 4: 443-452, 1991. 751

Zijdewind I, Thomas CK. Motor unit firing during and after voluntary contractions of human 752

thenar muscles weakened by spinal cord injury. J Neurophysiol 89: 4: 2065-2071, 2003. 753

754

755


34

FIGURE LEGEND 756 757 Figure 1. Methodological design of study. RI, reciprocal inhibition; PI, presynaptic inhibition; 758 LFD, low frequency depression; SOL/TA, soleus/tibialis anterior; PF, plantar flexor; DF, 759 dorsiflexion; ROM, range of motion; LE, lower extremity; MVC, maximum voluntary 760 contraction. 761 762 Figure 2. Schematic of ankle clonus drop test set-up and kinematic, kinetic, and EMG output. 763 PF, plantar flexor; SOL, soleus; MG, medial gastrocnemius; TA, tibialis anterior. 764 765 Figure 3. Feedback displays and change in performance measures. A. TA↑ training display: left 766 screen, TA EMG feedback and reward; right screen, TA%MVC amplitude target and response. B. 767 SOL↓ training display: left screen SOL H-reflex stimulus, M-wave, and HDF-reflex; right screen, 768 SOL HDF-reflex amplitude target and response. C. TA↑ graph of change in mean TA %MVC 769 amplitude, (*) p = 0.02. D. SOL↓ graph of change in conditioned SOL H

DF reflex (% baseline), 770

(+) p = 0.09. Week 1= Baseline, Week 5 = last week of training. (Wilcoxon signed ranks one-771 tailed tests) 772 773 Figure 4. Example of effect of SOL↓ training on SOL HDF –reflex amplitude responses in one 774 participant. A. Training session 2. B. Training session 11. Black line, SOL HDF –reflex 775 responses; Grey line, M-wave responses evoked at 10-30% Mmax stimulus intensity. Baseline 776 bout, 30 repetitions without feedback; Training bouts, 30 repetitions with feedback with 1-minute 777 rests between bouts. 778 779 Figure 5. Effects of TA (↑) training. Change in outcome measures and change score 780 correlations. Solid single headed arrow, p ≤ 0.05, change achieved significance; Dashed single 781 headed arrow, 0.05 < p ≤ 0.10, change approached statistical significance; Dashed double 782 headed arrow, moderate correlation, 0.5 ≤ r < 0.75, 0.05 < p ≤ 0.10; Spearman correlation 783 coefficients, one-tailed test. 784 785 Figure 6. Effects of SOL (↓) training. Change in outcome measures and change score 786 correlations. Solid single headed arrow, p ≤ 0.05, change achieved statistical significance; 787 Dashed single headed arrow, 0.05 < p ≤ 0.10, change approached statistical significance; Dashed 788 double headed arrow, moderate correlation, 0.5 ≤ r < 0.75, 0.05 < p ≤ 0.10; Spearman correlation 789 coefficients, one-tailed test. 790

791


35

Appendix 1. List of Abbreviations 792 793 AIS – ASIA Impairment Scale 794

ASIA – American Spinal Injury Association 795

CPN – common peroneal nerve 796

DF - dorsiflexion 797

DFAROM – dorsiflexion active range of motion 798

EMG - electromyography 799

FN – femoral nerve 800

H/M – H-reflex/M-wave ratio 801

HDF

– H-reflex during dorsiflexion 802

HmaxDF

– H-reflex maximum amplitude during dorsiflexion 803

Hmaxrest

- H-reflex maximum amplitude at rest 804

ISI – inter-stimulus interval 805

LEMS – lower extremity motor score 806

LFD – low frequency depression 807

MmaxDF

– M-wave maximum amplitude during dorsiflexion 808

Mmaxrest

– M-wave maximum amplitude at rest 809

MVC – maximum voluntary contraction 810

PF – plantar flexion 811

PFRTA – plantar flexion reflex threshold angle 812

PI – presynaptic inhibition 813

PTN – posterior tibial nerve 814

RI – reciprocal inhibition 815


36

ROM – range of motion 816

SCI – spinal cord injury 817

SOL - soleus 818

SRM – standardized response mean 819

TA – tibialis anterior 820

TTL – Transistor-Transistor Logic 821

822 823 824 825 826

827 828


37

Appendix 2A. Pre-intervention relationships between neurophysiologic measures and clinical and walking measures. 829 Measure Clonus

duration Plantar flexor Reflex Threshold Angle

Toe Tap Test

Dorsiflexion active ROM

Dorsiflexion strength

Plantar flexion strength

Training leg strength

Non-training leg strength

Foot Clear-ance

Distance Speed

RI (%)

0.035 0.457

-0.280 0.189

-0.028 0.465

0.077 0.406

-0.085 0.396

0.486 0.054

0.169 0.299

0.139 0.333

-0.042 0.448

0.217 0.249

0.105 0.373

PI (%)

-0.287 0.183

0.014 0.483

0.392 0.104

-0.294 0.177

0.074 0.410

0.247 0.219

0.261 0.206

0.275 0.193

-0.294 0.177

0.448 0.072

0.573 0.026

LFD (%)

0.350 0.133

-0.343 0.138

-0.018 0.478

-0.049 0.440

-0.052 0.436

-0.020 0.475

-0.141 0.331

-0.018 0.478

0.140 0.332

0.161 0.309

0.077 0.406

SOL/TA Ratio

0.112 0.365

0.133 0.340

0.042 0.448

0.524 0.040

-0.333 0.145

0.056 0.432

-0.533 0.037

-0.543 0.034

-0.364 0.123

-0.140 0.332

-0.091 0.389

830 Appendix 2B. Pre-intervention relationships between gait and clinical measures. 831 Measure Clonus

duration Plantar flexor Reflex Threshold Angle

Toe Tap Test

Dorsiflexion active ROM

Dorsiflexion strength

Plantar flexion strength

Training leg strength

Non-training leg strength

Foot Clearance

-0.259 0.208

-0.580 0.024

-0.409 0.093

-0.154 0.317

-0.026 0.468

-0.187 0.280

0.014 0.483

0.079 0.404

Speed

-0.538 0.035

-0.091 0.389

0.166 0.303

-0.594 0.021

0.100 0.379

0.614 0.017

0.476 0.059

0.718 0.004

Distance

-0.294 0.177

0.021 0.474

0.374 0.116

-0.524 0.040

0.344 0.137

0.702 0.005

0.646 0.012

0.711 0.005

Spearman correlation coefficients (top number) and p-value (bottom number) within each cell; RI = reciprocal inhibition; 832 PI = presynaptic inhibition; LFD = low frequency depression; SOL/TA = Soleus/Tibialis Anterior; ROM = range of motion. 833


38

Appendix 3A. Relationship between changes in neurophysiologic and clinical, gait and performance measures for TA↑ group 834 Measure ↓

Clonus duration

↓ Plantar flexor reflex threshold angle

↑ Toe Tap Test

↓ Dorsi flexion active ROM

↑ Dorsi flexion strength

↑ Plantar flexion strength

↑ Training leg strength

↑ Non-training leg strength

↑ Foot Clear-ance

↑ Distance

↑ Speed

↑ %MVC Ampli-tude

↑ RI (%)

-0.086 0.436

0.657 0.078

0.609 0.100

0.429 0.198

-0.393 0.221

-0.525 0.143

-0.883 0.010

-0.278 0.297

0.714 0.055

-0.429 0.198

-0.257 0.311

-0.486 0.164

↑ PI (%)

-0.086 0.436

-0.829 0.021

-0.551 0.129

-0.771 0.036

-0.393 0.221

-0.216 0.341

0.265 0.306

0.802 0.027

-0.371 0.234

-0.486 0.164

0.429 -0.198

0.657 0.078

↓LFD (%)

-0.600 0.104

-0.086 0.436

0.145 0.392

-0.714 0.055

-0.655 0.079

-0.463 0.178

-0.618 0.096

0.309 0.276

-0.200 0.352

-0.714 0.055

0.600 0.104

-0.086 0.436

↓ SOL/TA Ratio

-0.029 0.479

0.257 0.311

0.348 0.250

0.143 0.394

0.655 0.079

0.309 0.276

0.265 0.306

-0.772 0.036

-0.314 0.272

0.771 0.036

-0.200 0.352

-0.771 0.036

835 Appendix 3B. Relationship between changes in gait and clinical and performance measures for TA↑ group 836 Measure ↓

Clonus duration

↓ Plantar flexor Reflex Threshold Angle

↑ Toe Tap Test

↓ Dorsiflexion active ROM angle

↑ Dorsiflexion strength




↑ % MVC Amplitude

↑ Foot Clearance

0.257 0.311

0.314 0.272

0.058 0.457

0.714 0.055

-0.131 0.402

-0.432 0.196

-0.441 0.190

-0.093 0.431

0.029 0.479

↑ Speed

0.143 0.394

-0.486 0.164

0.290 0.289

-0.543 0.133

-0.655 0.079

-0.926 0.004

-0.265 0.306

0.463 0.178

-0.200 0.352

↑ Distance

0.314 0.272

0.143 0.394

0.174 0.371

0.275 0.311

0.655 0.079

0.617 0.096

0.618 0.096

-0.463 0.178

-0.314 0.272

↑%MVC Amplitude

0.143 0.394

-0.486 0.164

-0.754 0.042

-0.086 0.436

-0.131 0.402

-0.185 0.363

0.353 0.246

0.679 0.069

1.000

Arrows indicate direction of change; Spearman correlation coefficients (top number) and p-value (bottom number) within each cell; 837 RI = reciprocal inhibition; PI = presynaptic inhibition; LFD = low frequency depression; SOL/TA = Soleus/Tibialis Anterior; ROM = 838 range of motion. Yellow = measures related to hypotheses; Blue: direction of correlation differed from results unclear relationship; 839 Grey = moderate or strong correlations not related to original hypotheses 840


39

841 842 Appendix 4A. Relationship between changes in neurophysiologic and clinical, gait and performance measures for SOL↓ group 843

Measure ↓ Clonus duration

↑ Plantar flexor reflex threshold angle

↑ Toe Tap Test

↓ Dorsi flexion active ROM angle

↑ Dorsi flexion strength




↑ Foot Clear-ance

↑ Distance

↑ Speed

↓ H

DF %

of baseline

↑ RI (%)

-0.314 0.272

-0.714 0.055

0.290 0.289

-0.086 0.436

-0.169 0.354

0.372 0.234

0.232 0.329

0.464 0.177

-0.754 0.042

-0.200 0.352

-0.143 0.394

-0.029 0.479

↓ PI (%)

0.086 0.436

-0.314 0.272

0.754 0.042

-0.143 0.394

0.304 0.279

-0.270 0.302

0.609 0.100

0.174 0.371

-0.116 0.413

0.029 0.479

-0.143 0.394

-0.143 0.394

↑ LFD (%)

0.667 0.074

0.203 0.350

-0.132 0.401

0.551 0.129

0.017 0.487

-0.051 0.461

0.206 0.348

0.162 0.380

-0.309 0.276

-0.232 0.329

-0.232 0.329

0.551 0.129

↓ SOL/TA Ratio

0.371 0.234

-0.143 0.394

-0.029 0.478

0.714 0.055

-0.845 0.017

0.778 0.034

-0.580 0.114

0.464 0.177

-0.377 0.231

-0.600 0.104

-0.600 0.104

0.371 0.234

844 Appendix 4B. Relationship between changes in gait and clinical and performance measures for SOL↓ group 845 Measure ↓

Clonus duration

↑ Plantar flexor Reflex Threshold Angle

↑ Toe Tap Test

↓ Dorsiflexion active ROM angle

↑ Dorsiflexion strength




↓ H

DF %

of baseline

↑ Foot Clearance

0.116 0.413

0.638 0.087

-0.059 0.456

-0.058 0.457

-0.017 0.487

-0.189 0.360

-0.471 0.173

-0.426 0.200

-0.928 0.004

↑ Distance

-0.771 0.036

-0.257 0.311

-0.319 0.269

-0.886 0.009

0.778 0.034

-0.372 0.234

0.638 0.087

-0.116 0.413

0.257 0.311

↑ Speed

-0.771 0.036

-0.429 0.198

0.029 0.478

-0.543 0.133

0.372 0.234

-0.788 0.034

0.290 0.289

-0.812 0.025

0.086 0.436

↓ HDF

% of

baseline

-0.029 0.479

-0.543 0.133

-0.058 0.457

-0.200 0.352

-0.034 0.475

0.068 0.449

0.377 0.231

0.232 0.329

1.000

Arrows indicate direction of change; Spearman correlation coefficients (top number) and p-value (bottom number) within each cell; 846 RI, reciprocal inhibition; PI, presynaptic inhibition; LFD, low frequency depression; SOL/TA, Soleus/Tibialis Anterior; ROM, range 847 of motion; HDF, SOL H-reflex during dorsiflexion 848

Participants (n=12)

Pre-training Tests

SOL H-Reflex RI PI LFD

Clonus Drop Test Clonus duratiom PF reflex threshold angle SOL/TA Coactivation

Ankle Motor Control 10-s Toe Tap Test DF Active ROM

Walking 6-m walk Foot Clearance Speed 2-min Distance

LE Motor Scores Dorsiflexors Plantar flexors Training leg Non-training leg

Randomization Toe Tap Score

< 15

TA↑ (n=6)

≥ 15

SOL↓ (n=6)

15 Sessions 30 repetitions/bout 1 minute rest between bouts 10 bouts 300 total repetitions/session 2-hr, 3X/week

12 Training Sessions Dosage: 3600 repetitions

3 Baseline Sessions Dosage: 900 repetitions

Post-Training Tests

Training Parameter Increase % TA MVC Amplitude

Training Parameter Decrease SOL H-reflex amplitude during active dorsiflexion

EMG

TA (mV)

MG (mV)

SOL (mV) 1.00

0.00

-1.00

0.25

0.00

0.25

0.10

0.00

0.10

║

║

║

║

║

Time (sec)

Time (sec)

Time (sec)

115

95

0.00

75

30.00

-50.00

Ankle Angles (degrees) ↓ Dorsiflexion

Ground Reaction Forces (newtons) ↑ Dorsiflexion

PF reflex threshold angle

Time (sec)

Time (sec)

EEMG

0

20

40

60

80

100

1 5

TA %

MV

C A

mpl

itude

Week

1

2

3

4

5

6

mean

A. TA ↑

*

%MVC target

100%

0%

%MVC Amplitude response

mV

Stimulus

H-reflex M-wave

mV

HDF-reflex target (mV)

HDF-reflex response

mV

+

C. TA↑

D. SOL↓ B. SOL↓

40

60

80

100

120

1 5

Week

123456mean

Con

ditio

ned

SOL

HD

F ref

lex

(% b

asel

ine)

EMG

Reward

%MVC Amp (RMS)

Footswitch

0

0.5

1

1.5

2

2.5

3

3.5

1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 103

109

115

121

127

133

139

145

151

157

163

169

175

181

187

193

199

205

211

217

223

229

235

241

247

253

259

265

271

277

283

289

Am

plitu

de (m

V)

Repetition Baseline Bout

Training Bouts

0

0.5

1

1.5

2

2.5

3

3.5

1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96 101

106

111

116

121

126

131

136

141

146

151

156

161

166

171

176

181

186

191

196

201

206

211

216

221

226

231

236

241

246

251

Am

plitu

de (m

V)

Repetition Training

Bouts Baseline

Bout

A. Training Session 2

B. Training Session 11

1

↓ PF reflex threshold angle

↓ Clonus

↓ DF active ROM angle

↑ ROM

↑ Distance ↑ Foot Clearance

↑ Walking Function

↑ TA % MVC Amplitude

TA ↑

↑ Target Stimulus

↑ Non-training Leg

↑ Training Leg ↑ DF

↑ Strength (LEMS)

↓ Conditioned SOL HDF reflex

↓ Target Stimulus

↓ SOL/TA Coactivation

↓ Clonus

↑ Non-training leg

↑ Strength (LEMS)

↑ Speed ↑ Distance

↑ Walking Function

SOL ↓

2

Table 1. Analysis of differences in participant baseline characteristics and pre-intervention outcome measures (mean ± SD) between TA↑ and SOL↓ groups TA↑ SOL↓ p-value Characteristics: Age (yrs) 44.2 ± 12.0 45.2 ± 12.8 1.00 Post-injury (yrs) 10.8 ± 10.0 10.8 ± 08.8 1.00 Gender 6 men 4 men, 2 women 0.12 (Х2) Injury Level C7 (median) C5 (median) 0.39 (X2) AIS D (all) D (all) N/A Clinical Measures: Clonus Duration (sec) 11.3 ± 14.5 12.4 ±10.2 0.75 PFRTA (degrees) 83.7 ± 5.5 85.2 ± 7.2 0.75 Toe Tap Test Score 13.5 ± 10.9 11.5 ± 6.3 0.75 DF AROM (degrees) 96.9 ± 13.7 112.4 ± 11.8 0.05 * DF LEMS 3.5 ± 1.1 3.8 ± 0.8 0.55 PF LEMS 2.5 ± 0.8 2.5 ± 0.6 0.78 Training leg LEMS 17.3 ± 3.4 16.2 ± 3.2 0.42 Non-training leg LEMS 22.5 ± 2.0 19.2 ± 2.8 0.04 * Walking Measures: Foot Clearance (mm) 5.0 ± 2.0 5.1 ± 1.6 0.87 Speed (m/s) 0.29 ± 0.26 0.14 ± 0.08 0.15 § Distance (m) 44.9 ± 49.3 21.3 ± 17.2 0.52 § Neurophysiologic Measures: RI SOL H-reflex (%) 98.9 ± 4.5 99.6 ± 1.1 0.87 PI SOL H-reflex (%) 7.8 ± 5.8 4.2 ± 1.6 0.20 LFD SOL H-reflex (%) 91.9 ± 16.5 95.0 ± 10.0 0.75 SOL/TA Ratio 0.49 ± 0.23 0.52 ± 0.39 0.63 Mann-Whitney U-test for all analyses, except Chi-square (X2) test for nominal data; C, cervical; AIS, American Spinal Injury Association Impairment Scale (ASIA, 2002); RI, reciprocal inhibition as a % of control reflex; PI, presynaptic inhibition as a % of Mmax; LFD, low frequency depression as a % of control reflex; H/M, H-reflex/M-wave; SOL/TA, soleus/tibialis anterior; PFRTA, plantar flexor reflex threshold angle; DF, dorsiflexion; AROM, active range of motion; LEMS, lower extremity motor score; PF, plantar flexion; (*) significant difference (p ≤ 0.05); (§) large but not significant difference.

Table 2. Analysis of change in training and outcome measures within each group (Wilcoxon Signed Rank one-tailed test) and between groups (Mann-Whitney U-test) TA↑ Group ___ SOL↓ Group _ Between-Groups Mean ± SD p-value SRM Mean ± SD p-value SRM p-value Training Measures: TA %MVC Amplitude 48.19 ± 9.29 0.01 * 5.19 Conditioned SOL H

DF (% baseline) -16.17 ± 26.22 0.09 + -0.62

Clinical Measures: Clonus duration (sec) -7.38 ± 12.63 0.22 -0.58 -2.39 ± 4.69 0.11 -0.51 0.87 PF reflex threshold angle (degrees) -4.33 ± 2.23 0.02 * -1.86 3.15 ± 11.53 0.22 0.27 0.06 + Toe Tap Test (score) 2.3 ± 5.3 0.16 0.43 1.3 ± 2.0 0.22 0.43 0.62 DF active ROM angle (degrees) -4.32 ± 4.53 0.05 * -0.95 -4.20 ± 9.24 0.22 -0.45 0.87 Dorsiflexion strength (LEMS) 0.8 ± 0.4 0.03 * 2.08 0.5 ± 0.8 1.00 0.63 0.25 Plantar flexion strength (LEMS) 0.3 ± 0.8 0.19 0.41 0.5 ± 0.8 1.00 0.63 0.93 Training leg strength (LEMS) 2.2 ± 1.9 0.06 + 1.16 1.3 ± 2.0 0.13 0.65 0.46 Non-training leg strength (LEMS) 0.8 ± 0.8 0.06 + 1.11 1.8 ± 2.1 0.06 + 0.86 0.33 Walking Measures: Foot Clearance (mm) 4.8 ± 5.3 0.05 * 0.91 1.6 ± 6.0 0.42 0.27 0.26 Speed (m/s) 0.04 ± 0.07 0.11 0.54 0.02 ± 0.04 0.08 + 0.69 1.00 Distance (m) 12.09 ± 9.03 0.02 * 1.34 6.25 ± 2.96 0.02 * 2.11 0.17 Neurophysiologic Measures: RI SOL H-reflex (%) 0.1 ± 0.5 0.50 -0.02 0.5 ± 4.8 0.42 -0.12 0.87 PI SOL H-reflex (%) 1.0 ± 5.4 0.42 -0.18 -1.3 ± 3.6 0.34 0.36 0.52 LFD SOL H-reflex (%) -3.1 ± 14.6 0.42 0.24 6.4 ± 10.2 0.16 -0.63 0.34 SOL/TA Clonus Coactivation Ratio -0.13 ± 0.36 0.28 -0.36 -0.21 ± 0.29 0.02 * -0.72 0.75 TA, tibialis anterior; SOL, soleus; SRM, standardized response mean; MVC, maximum voluntary contraction; H

DF, H-reflex during dorsiflexion; Mmax

DF,

maximum M-wave during dorsiflexion; RI, reciprocal inhibition as a % of control reflex; PI, presynaptic inhibition as a % of Mmax; LFD, low frequency depression as a % of control reflex; PF, plantar flexor; DF, dorsiflexion; ROM, range of motion; LEMS, lower extremity motor score;(*) = p ≤ 0.05, achieved statistical significance; (+) = p > 0.05 but ≤ 0.10, approached statistical significance; SRM interpretation of effect size = 0.20, 0.50, and 0.80 for small, moderate, and large effects, respectively.

Operant Conditioning to Increase Ankle Control or Decrease ... · 48 Ankle clonus is common after...

Documents

Transcript of Operant Conditioning to Increase Ankle Control or Decrease ... · 48 Ankle clonus is common after...