<%BANNER%>

Using EMG spectral analysis as a quantitative measure of motor unit recruitment to evaluate stroke motor recovery

HIDE
 Title Page
 Acknowledgement
 Table of Contents
 List of Tables
 List of Figures
 Abstract
 Introduction and review of...
 Method
 Results
 Discussion
 Appendix: Pilot data
 References
 Biographical sketch
University of Florida Institutional Repository

PAGE 1

USING EMG SPECTRAL ANALYSIS AS A QUANTITATIVE MEASURE OF MOTOR UNIT RECRUITMENT TO EV ALUATE STROKE MOTOR RECOVERY By LINDSAY ANN MCMANUS A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2006

PAGE 2

Copyright 2006 by Lindsay Ann McManus

PAGE 3

ACKNOWLEDGMENTS I would like to thank my advisor and committee chairman, Dr. James Cauraugh, for all of his guidance and support for the past two years. He helped to create an environment here at the University of Florida that has utilized my strengths to their highest potential. From teaching the undergraduate motor learning course to collecting and analyzing data in the lab, Dr. Cauraugh has given me the chance to learn and discover what interests me and what paths I will follow in the future. I also would like to thank my other committee members, Dr. Chris Janelle and Dr. Mark Tillman, for their suggestions and guidance through my thesis writing experience. In addition, I have to say thank you to Steve Coombes who, first of all, made this thesis possible with his LabView programming skills. Otherwise, I would never have been able to explore something like spectral analysis. Steve has been a great person to work with in the lab, share ideas with, and learn from. I know he will be a successful researcher one day and I look forward to hearing about his innovative research in the future. Outside of the Motor Behavior Lab, James Milford and Lori Gibbs have been very valuable resources for me as a graduate student. I would like to thank James for all of his caring support from the day I arrived all the way through my thesis writing and graduation. Lori has also been extremely helpful in everything from helping me prepare classroom materials to gathering paperwork for various tasks. Lastly, I would like to take the chance to thank my family and friends for all of their love and support for the past two years. Some friends are far away, but have given me the encouragement to pursue my degree. Other iii

PAGE 4

friends I have met here at UF and I will value their friendship for years to come. My family has always told me to follow my dreams and encouraged me to keep going above and beyond what is expected. And with their love and support, I think that I have truly made the most of my time here. I will always be proud to be a Florida Gator! iv

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................iii LIST OF TABLES ............................................................................................................vii LIST OF FIGURES .........................................................................................................viii INTRODUCTION AND REVIEW OF LITERATURE.....................................................1 Normal Motor Unit Recruitment..................................................................................1 Recruitment Following Neurological Insult.................................................................3 Bilateral Training..........................................................................................................5 Coupled Bilateral Training and EMG-Triggered Stimulation Protocols......................7 Adding a Load to the Unimpaired Limb.......................................................................8 Spectral Analysis..........................................................................................................8 Purpose.......................................................................................................................12 Hypotheses..................................................................................................................12 METHOD..........................................................................................................................15 Subjects.......................................................................................................................15 Pretest-Posttest Instruments and Procedures..............................................................16 Motor Recovery Protocols: Training Procedures.......................................................17 EMG Spectral Analysis..............................................................................................18 Statistical Design and Analyses..................................................................................18 RESULTS..........................................................................................................................20 Power Band 1 (1-3 Hz) and Power Band 2 (5-12 Hz)................................................20 Power Band 3 (15-30 Hz)...........................................................................................20 Power Band 4 (30-55 Hz)...........................................................................................22 Mean Peak Frequency.................................................................................................24 Force...........................................................................................................................25 DISCUSSION....................................................................................................................28 APPENDIX. PILOT DATA.............................................................................................38 LIST OF REFERENCES...................................................................................................45 v

PAGE 6

BIOGRAPHICAL SKETCH.............................................................................................50 vi

PAGE 7

TABLE Table page 4-1 Averages collapsed across subjects and groups.......................................................36 vii

PAGE 8

LIST OF FIGURES Figure page 2-1 Subject performing the box and block test with his impaired right hand..............16 3-1 Group condition interaction in power band 3 (15-30 Hz)..................................21 3-2 Group limb interaction in power band 3 (15-30 Hz)..........................................21 3-3 Limb condition interaction in power band 4 (30-55 Hz)....................................23 3-4 Limb condition session interaction in power band 4 (30-55 Hz)....................23 3-5 Mean peak frequency limb main effect.................................................................24 3-6 Mean peak frequency group limb interaction.....................................................25 3-7 Condition main effect............................................................................................26 3-8 Limb main effect....................................................................................................26 3-9 Limb condition interaction.................................................................................27 4-1 Power spectrum. Frequency (Hz) on x-axis. Power (V2) on y-axis...................31 4-2 Power spectrum. Frequency (Hz) on x-axis. Power (V2) on y-axis...................31 4-3 Power spectrum. Frequency (Hz) on x-axis. Power (V2) on y-axis...................32 4-4 Power spectrum. Frequency (Hz) on x-axis. Power (V2) on y-axis...................32 4-5 Power spectrum. Frequency (Hz) on x-axis. Power (V 2 ) on y-axis....................33 4-6 Power spectrum. Frequency (Hz) on x-axis. Power (V2) on y-axis...................33 A-1 Average power (subject 1102) in pre and posttests in unimanual condition.........39 A-2 Average power (subject 1102) in pre and posttests in bimanual condition...........39 A-3 Average power (subject 2102) in pre and posttests in unimanual condition.........40 A-4 Average power (subject 2102) in pre and posttests in bimanual condition...........40 viii

PAGE 9

A-5 Mean peak frequency (subject 1102) in pre and posttests unimanual condition...41 A-6 Mean peak frequency (subject 1102) pre and posttests bimanual condition.........42 A-7 Mean peak frequency (subject 2102) pre and posttests unimanual condition.......42 A-8 Mean peak frequency (subject 2102) pre and posttests bimanual condition.........43 A-9 Mean force (subject 1102) in posttests unimanual vs. bimanual conditions.........44 A-10 Mean force (subject 2102) posttests unimanual vs. bimanual conditions.............44 ix

PAGE 10

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science USING EMG SPECTRAL ANALYSIS AS A QUANTITATIVE MEASURE OF MOTOR UNIT RECRUITMENT TO EVALUATE STROKE MOTOR RECOVERY By Lindsay Ann McManus August 2006 Chair: James Cauraugh Major Department: Applied Physiology and Kinesiology After stroke, motor function is often reestablished abnormally in the paretic limb, causing difficulty in daily tasks. The two primary purposes of this study were to apply spectral properties, namely mean peak frequency shifts and changes in average power over four frequency bins from surface EMG recordings, to describe motor unit recruitment in the muscles of the wrist/finger extensors, and to correlate functional improvements, namely increased force production, to improved motor unit recruitment patterns. Using data collected from 2002 through 2006 from two separate training protocols, twenty five chronic patients were selected for analysis. The first protocol consisted of a pretest, a 2 week (4 sessions) coupled bilateral training and active EMG-neuromuscular stimulation protocol, and a posttest. The second group (Weighted) underwent a pretest and posttest identical to previous group (Coupled), but participated in a coupled bilateral x

PAGE 11

training and active EMG-neuromuscular stimulation with an added weight on the affected hand. Separate analyses were conducted on the three dependent measures (average power, mean peak frequency, and mean force) using a four-way mixed ANOVA. Significant results were revealed in the low gamma power band, in addition to mean peak frequency and mean force (p < 0.05). In the low gamma band, the unaffected limb in the unimanual condition during the posttest exhibited significantly higher power than every other combination of condition, limb, and session. Moreover, the unaffected limb showed a significantly higher mean peak frequency than the affected limb. In addition, the unaffected limb in the coupled group showed higher mean peak frequency than the affected limb in the same group. The affected limb in the coupled group also displayed a lower mean peak frequency than the unaffected limb in the weighted group. Analysis of the force data revealed higher force when moving unimanually compared to moving bimanually. As expected, the affected limb produced a lower amount of force compared to the unaffected limb in both unimanual and bimanual conditions. Findings support previous research stating that the unaffected limb exhibits higher power than the affected limb, indicating that the unimpaired limb is able to harness larger motor neurons in its recruitment strategy, thus leading to larger force production. Each subject exhibited improvement in at least one category, number of blocks moved, increased mean peak frequency, a change in power in the 30-55 Hz bands, or amount of force produced. However, no single uniform characteristic of motor improvement was shown by each stroke subject. The current findings indicate that the training is affecting the stroke patients positively and merits further research xi

PAGE 12

INTRODUCTION AND REVIEW OF LITERATURE Normal Motor Unit Recruitment The principle of motor unit recruitment states that motor units, consisting of the motor neuron and the muscle fibers it innervates, are progressively activated with increasing strength of voluntary muscle contraction (Milner-Brown et al., 1973). The central nervous system increases the strength of muscle contraction by increasing the number of active motor units or increasing the firing rate of individual motor units. Both of these mechanisms occur concurrently, though the recruitment of more motor units takes precedence over an increase in firing rate until nearly all motor units are recruited. Once this level is reached, motor units are driven to fire in their secondary range to rates greater than 50 Hz (Petajan, 1991). Motor units are recruited according to the order of their size. The first motor units to fire are small in size, weak in force generation, and resistant to fatigue. Progressively larger motor units with greater force production capabilities are recruited with increasing strength of muscle contraction (Henneman et al., 1965). As large motor units are activated, it is important to note that the smaller units are still firing, although the large motor units assume primary responsibility for maximum force production. This sequential recruitment order is known as the Henneman size principle and it results in a smooth increase in muscle strength. Through their feline research paradigm, Henneman and colleagues discovered that axon diameter, conduction velocity, and cell size all increase with increased muscle contraction (Henneman et al., 1965). The three main 1

PAGE 13

2 types of motor units include Type I fibers, described as slow twitch, small, fatigue-resistant units which generate the least force and have the slowest contraction time, Type IIa units that are fast-twitch, larger, fatigue-resistant units able to generate greater forces and faster contraction times, and Type IIx units, known as the largest, fast-twitch, easily-fatigable units capable of the greatest force production and fastest contraction time. Motor unit recruitment sequences begin with Type I motor units and then sequentially advance to Type IIa and finally Type IIx units for high force production. The late recruited Type II fibers have larger diameters and thus generate higher motor unit action potentials (MUAPs) than the smaller Type I fibers. This size principle is used in electromyography (EMG) studies designed to analyze motor unit recruitment patterns by using the amplitude of the MUAP to define the size of a specific motor unit (Ertas et al., 1995). A confluence of MUAPs traveling along the sarcolemmas of the muscle fibers results in a muscle contraction. Myoelectric signals are formed on the basis of the interference pattern among these MUAPs and are detected using EMG. The conductance and resistance of the sarcolemma differ between fastand slow-twitch fibers (Luff & Atwood, 1972) and therefore, researchers postulate that different types of muscle fibers generate different MUAPs. Myoelectric signals provide information about the form of the MUAPs and are able to offer insight into the electrophysiology of the active muscle. Signals measured using intramuscular EMG in animals have shown that faster motor units generate higher frequencies in their power spectra (Wakeling & Rozitis, 2004). However, evidence from surface EMG in humans has remained elusive on this topic. Recent advances in EMG signal decomposition including wavelet techniques, Fourier

PAGE 14

3 Transform methods, and simulations that accurately mimic human muscle behavior have provided additional information not previously known, such as motor unit recruitment, muscle fatigue factors, and biomechanical adaptations (Karlsson et al., 2000). For example, when faster motor units are active, increases in mean and/or median myoelectric frequency are observed in the EMG signal, thus suggesting that there is an inherent property of the muscle fiber that can be distinguished within EMG signals (Farina et al., 2002). Recruitment Following Neurological Insult After a motor neuron injury such as a stroke, the relationship between motoneuron size and the number and size of the muscle fibers it innervates is usually lost or damaged. Reorganization of this relationship at motor endplates occurs over time through axon branching and neuromuscular activity (Nudo, 1999). This reorganization, known as neuroplasticity, occurs as the nervous system attempts to repair damaged motor neurons. Therefore, there must be persistent functional changes evident in the brain, and consequently in motor unit activity, to represent this new organization. Using spectral analysis to analyze MUAPs post-stroke and after bimanual coordination training will offer insight into abnormal motor unit recruitment strategies post-stroke and aid in the explanation of new strategies employed after a specific training protocol. After a cerebrovascular accident (CVA), such as stroke, motor unit recruitment is altered because of interrupted descending pathways as well as structural changes in the muscle (Mazevet et al., 2003). Brain damage secondary to stroke results in corticospinal and supraspinal motor pathway disruption and possibly leads to synaptic degeneration at the segmental level (McComas et al., 1973). This loss of neural signaling at the segmental level causes motor neuron loss and altered force control mechanisms. Muscle

PAGE 15

4 impairment following stroke can result from several sources including loss of motor units because of Wallerian degeneration (McComas et al., 1971). As a result of the Wallerian degeneration, all of the muscles fibers previously innervated by the axon are denervated and therefore, fewer motor units are available for muscle activation. Altered properties of the motor unit affect recruitment through altered recruitment order, discharge rate, and/or discharge pattern (Beaumont & Gardiner, 2002). This abnormal motor unit recruitment in turn, impacts the level of impairment post stroke. For instance, in neurologically healthy individuals, once the first recruited muscle unit reaches a firing frequency of 10 Hz, a second unit begins firing to increase muscle force. However, after stroke, this second unit is often missing or damaged and an increase in force can only be achieved by increasing the firing rate of the first motor unit(s). This abnormal recruitment and discharge pattern contributes to the inefficient and undesired movement patterns often observed in stroke patients. Adding to the underlying cause of such inefficient motor patterns is altered motor neuron pool excitability. Altered excitability affects the probability of motor unit activation, and in turn allows expression of unwanted or atypical coordination patterns (Morales et al., 1987). Low tonic rates are indicative of slow motor units while fast motor units are associated with higher, more phasic tonic rates (Denny-Brown, 1929; Granit et al., 1956). The close matching of motoneurons and muscle properties with respect to firing rates was first described by Kernell and colleagues in muscles of rats and cats (Granit et al., 1963) and later in single motor units of rats (Gardiner and Kernell, 1990) and supports the idea that motor unit recruitment and rate coding both contribute to voluntary force production (Kernell, 1992). That is, the number of motor units that are recruited and the rate of

PAGE 16

5 firing of the motoneurons both control force generation of the muscle during voluntary movement. Olson and colleagues successfully reported that a progressive recruitment of motor units occurs when increasing force is required of voluntary movements (Olson et al., 1968). However, impaired voluntary muscle strength results from neurological insult because of a loss of normal motor unit recruitment occurs because of Wallerian degeneration, altered recruitment and discharge patterns, and altered motor neuron pool excitability (Landau & Sahrmann, 2002). Moreover, drawing consistent conclusions from research involving impaired muscle control in hemiplegic subjects is difficult because of the inconsistencies across studies. For example, a limited number of subjects, a wide variation of lesion duration, and multiple causes of disability are prevalent in stroke motor recovery research. Therefore, the question as to whether the lack of muscular control because of disruption at the supraspinal level, intrinsic changes in the motor neuron pool, or changes in the properties of the muscle itself remains unanswered. However, the literature does agree on the principles that motor unit activation in hemiplegia subjects is non-uniform and motor unit firing rates tend to decrease overall relative to the ipsilesional limb (Rose & McGill, 1998). A consequence of this diminished activation and firing of motor units is a decreased ability to produce voluntary contractions. Bilateral Training Promising results have been reported in functional motor improvement post-stroke with bilateral training protocols. Mudie and Matyas (2000) demonstrated enhanced upper limb performance after a three week protocol of bilateral isokinematic therapy (BIT). Bilateral practice with both upper limbs successfully elicited improvements in unilateral performance of the affected limb alone. Motor improvements after bilateral

PAGE 17

6 training were greater than those in response to similar unilateral practice or practice sessions that involved the nonparetic limb assisting the paretic limb in various tasks. Significant increases in strength, range of motion, and functional motor ability were reported by Whitall and colleagues (2000) following a six week bilateral training protocol. In addition, Cauraugh and Kim (2002) demonstrated enhanced motor control after EMG-triggered neuromuscular stimulation coupled with bilateral training. These three studies provide support for a bilateral training technique because of the potential effectiveness of the various protocols in improving motor ability of individuals post stroke. The neural mechanisms and pathways targeted by bilateral training protocols remain uncertain; however, several possibilities have been proposed. One pathway involves an interaction in the affected hemisphere between surviving neurons of ipsilateral pathways and surviving neurons of the crossed corticospinal pathways, the involvement of indirect pathways from the affected hemisphere, and a facilitation of coupling or crosstalk from the unaffected hemisphere to ipsilateral pathways are all possible neural mechanisms responsible for motor improvement exhibited after bilateral training (Mudie & Matyas, 2000). The idea of the coupling principle between arms harnesses existing coupling abilities and encourages recovery of function through bilateral movements. For example, forcing the use of both upper extremeties, as opposed to the paretic arm only, has the potential to yield additional benefits by accessing intact pathways of the unaffected hemisphere and possible coupling mechanisms. Rose and Winstein (2004) reported clear temporal coupling in bimanual conditions and found a facilitation effect in 5 of 11 patients where the paretic limb exhibited a shorter movement

PAGE 18

7 time in the bimanual condition. Further, Cunningham and colleagues (2002) reported coupling in bimanual training citing smoother movement of the paretic limb when it was coupled with the nonparetic limb in elbow extension movements. Together these studies support neural coupling as a possible mechanism underlying bilateral training protocols. Coupled Bilateral Training and EMG-Triggered Stimulation Protocols In separate studies of EMG-triggered neuromuscular stimulation and bilateral coordination training, motor improvements in hemiparetic subjects were observed (Leonard, 1998; Popovic and Sinkjaer, 2001). These findings are based on accepted motor control theories, including sensorimotor integration and dynamical systems, which complement one another in that the body is viewed as a dynamic system executing movements through an interaction of sensory input and motor actions based on a set of imposed constraints. Both bilateral training and EMG-triggered neuromuscular stimulation essentially activate the same central and peripheral neural mechanisms associated with the pyramidal tract and provide a basis for combining the two protocols. The combination of the two interventions has the potential to accelerate the progress of motor recovery after stroke (Cauraugh & Kim, 2002). Specifically, Cauraugh and Kim reported an increase in performance of a functional task, decreased reaction times, and increased capability in the sustained force task for the coupled bilateral training and EMG-triggered stimulation group compared to the unilateral and control groups. The coupled intervention group exhibited motor improvement in all three testing categories, exhibiting decreased hemiparesis because of each subjects expanded motor control repertoire (Cauraugh & Kim, 2002; Cauraugh, 2004).

PAGE 19

8 Adding a Load to the Unimpaired Limb Interlimb coupling is a potentially useful phenomenon in the context of stroke and hemiplegia. It is beneficial to increase the symmetry between the impaired and unimpaired limb through bilateral training in an effort to allow the unimpaired cortex to facilitate the damaged cortex. Interlimb coupling increases with the frequency of the movement, in addition to the torque requirements of the task (Walter and Swinnen, 1990). Therefore, through inertial loading of the unimpaired limb, torque is manipulated and combined with the coupled bilateral training EMG-neuromuscular stimulation protocol for one group in this study. Spectral Analysis Raw EMG signals offer valuable information, but in a relatively useless form. Conclusions drawn from raw EMG signals cannot be quantitatively compared between subjects, and therefore most analyses are purely qualitative. Pioneering EMG work began in the 1950s and has undergone much refinement, especially in the past 20 years due to the advanced processing of computers and other instrumentation. The formal quantitative analysis of MUAP signals is not as cumbersome or time-consuming as it was in the past thanks to recent advances. Among these advances is the ability to use the Fast Fourier Transform (FFT) filter to break the EMG signal into its frequency components which are then graphed as a function of their probability of occurrence and used to determine muscle activation and fatigue. EMG signal analysis in the frequency domain involves parameters describing specific aspects of the frequency spectrum of the signal. The two most common parameters of the power density spectrum used to determine fatigue are mean and median frequency because they are easily applicable and provide useful information about the

PAGE 20

9 EMG signal. If the EMG spectrum has a normal distribution, these two frequencies will be equivalent, but deviation from normality will result in differing values. Spectral shifts in surface EMG recordings are attributed to the type of muscle fibers activated and are therefore used to characterize motor unit recruitment. Investigation into these shifts has been limited to changes in the mean and median frequency of the power spectrum calculated through FFT (Basmajian & DeLuca, 1985). The amplitude of the surface EMG is related to the recruitment and the discharge rates of the active motor units. Because of this relationship, EMG amplitude is used as an index of the level of activation provided by the spinal cord. Using FFT, the mean and median frequencies of the EMG signal are identified and analyzed. Shifts in the mean or median frequencies should reflect the recruitment of progressively larger and faster motor units based on Hennemans size principle. Stated differently, variation in muscle fiber conduction velocity has been proposed as an index of motor unit recruitment according to the size principle, and therefore should accurately illustrate motor unit recruitment (Farina et al., 2004). Data obtained from surface EMG techniques may not need to correlate directly with traditional needle EMG data to be of value in the study of nerve and muscle disorders. For instance, the decline in median frequency of the power spectrum with fatigue or increasing force is a measure that has no single correlate in needle EMG analysis. The combination of surface EMG and computerized signal processing has allowed for nerve and muscle studies that had previously only been analyzed using needle EMG technology. Toulouse and colleagues (1992) studied interference patterns in persons with Guillain-Barr syndrome, a disorder characterized by deterioration in the

PAGE 21

10 peripheral nervous system, and normal subjects. They found that surface EMG analysis was useful in quantifying observed changes over the recovery period. In addition, Priez et al. (2002) used surface EMG spectral analysis to discriminate between normal subjects and those with muscular dystrophy. Further, the technique was successfully applied to quantifying the severity of the disease. The information gathered by needle EMG is invaluable and superior to surface EMG because of the specificity and accuracy of the signals. However, researchers are encouraged to investigate the value of surface EMG analysis and its ability to provide insight into the study of motor unit recruitment and activity (Hogrel, 2005). The lowest frequency at which motor units are modulated is 2 Hz. This common drive is of central origin and has been found in muscles without muscle spindles (DeLuca, 1985). However, it has only been demonstrated during isometric contractions, where the force slowly increases or is constant (Farmer et al., 1993a). Low frequency common drive persists after stroke, and this further supports the hypothesis that the modulation is of central origin (Farmer et al., 1993b). The next phase of modulation in motor unit firing occurs at a frequency of 5-12 Hz. To date, there is no convincing evidence that this activity is directly controlled by the motor cortex, though peaks in this frequency range have been reported in coherence spectra of the magnetoencephalographic signal (MEG) from the motor cortex and EMG activity (Salenius et al., 1997). The next motor unit activity frequencies include beta (15-30 Hz) and low gamma (30-60 Hz) bands which are regulated by the motor cortex (Brown, 2000). Firing frequencies will vary based on the motor units, muscles, and persons. This rate modulation depends on the intensity of descending motor input which is reinforced

PAGE 22

11 by additional local and afferent inputs. Therefore, after stroke when conduction in the central nervous system motor pathways is impaired, motor unit firing patterns are altered (Gemperline et al. 1995). Employing spectral analysis of muscle force as an estimator of motor unit activity has shown that the activity spectra tends to shift to higher frequencies with increasing force in normal subjects. Moritani and Muro (1986) reported increases in surface EMG amplitude and mean power frequency with increasing force. They attributed these changes to progressive increases in the motor unit firing frequency of initially recruited motor units as well as newly recruited motor units. Thus, Moritani and Muro concluded that surface EMG spectral analysis can provide a sensitive measure of motor unit activity during force output. In addition to the normal motor unit activity recorded by surface EMG, Homberg et al. (1986) selected patients with different motor dysfunctions to illustrate how motor unit activity was reflected in the power spectra. They reported decreased and increased firing rates, as well as abnormal synchronization across the various motor dysfunctions. Surface EMG has been shown to be a valuable tool in evaluating motor unit activity. Although intracellular EMG technology is able to provide more detail and insight into single motor unit recruitment, surface EMG data offers a significant means to analyze muscle performance and fatigue. With the mathematical advances of spectral analysis, it is now possible to assess EMG signals in their frequency domains and investigate the effects of observed mean and median power shifts. Combining spectral analysis and surface EMG technology is an interesting approach to analyzing motor unit recruitment and, in turn, quantifying motor recovery after a nerve injury such as stroke.

PAGE 23

12 Purpose There were two primary purposes for this study: (1) to apply spectral properties, namely mean peak frequency shifts and changes in average power over four frequency bins from surface EMG recordings to describe motor unit recruitment in the muscles of the wrist/finger extensors, and (2) to correlate functional improvements, namely increased force production which has been previously reported (Cauraugh & Kim, 2002) in stroke patients after bilateral training, to improved motor unit recruitment patterns. Differences in the motor unit recruitment patterns of stroke subjects before and after participation in a coupled EMG-neuromuscular stimulation and bilateral coordination training protocol were analyzed. Recruitment patterns exhibited during a sustained contraction task were quantified by the mean peak frequency and average power of EMG recordings. Pre and post data were considered, in addition to unimanual (paretic limb moving alone) versus bimanual (paretic and nonparetic limb moving together) testing conditions. Hypotheses After participating in a motor recovery training protocol, force production of the paretic limb is expected to increase. Several mechanisms may be responsible for this increased force production. More specifically, increased force production occurs because of either faster motor neuron firing rates, as measured by mean peak frequency, or recruitment of larger, stronger motor neurons, quantified by average power. Overall, an increase in mean peak frequency is expected in the posttest, which would indicate that subjects are able to fire motor units at a faster rate than was possible before the coupled protocol training. The increase in mean peak frequency is predicted to be greater in the weighted group compared to the coupled group because of the increased torque demands.

PAGE 24

13 In the range of 1-3 Hz, stroke patients should exhibit similar average power preand post-training because this modulation has been reported to be preserved after stroke. If frequencies in this lower band are preserved after infarct, then it is likely that they are not controlled by the motor cortex and therefore possibly controlled at a sub-cortical level. There is no convincing evidence for motor cortex control of the next frequency band, 5-12 Hz, though peaks driven by the motor cortex have been reported. Hence, average power of the paretic and nonparetic limbs should mirror each other closely and there will be no significant differences between pre and posttests. In the next two bands, beta (15-30 Hz) and low gamma (30-60 Hz), the paretic limb will exhibit a significantly lower average power compared to the nonparetic limb. This modulation is controlled primarily by the motor cortex which has been damaged by the stroke and therefore, unable to elicit normal control over voluntary movements. The beta and low gamma bands are important in higher aspects of motor control. After motor unit injury such as stroke, there is an inability to shift to these higher frequencies and persons remain locked in a low frequency mode (Myers et al., 2004). Bilateral training protocols aim to facilitate crosstalk between the paretic and nonparetic limbs in an effort to increase the functional ability of the paretic limb. Therefore, after two weeks (4 days) of bilateral coordination training, motor unit recruitment in the paretic limb will improve in the beta and low gamma bands. More specifically, it is hypothesized that the average power in the beta and low gamma bands of the paretic limb will increase post-training in both the unilateral and bilateral testing conditions with a larger increase in the bimanual condition compared to the unimanual condition. A larger increase in average power in the beta and low gamma bands for the bimanual condition will be the result of a facilitation effect from the

PAGE 25

14 nonparetic hemisphere. A larger increase in power is also expected to be shown in the beta and low gamma bands by the weighted group. The heightened torque demands of the weighted group will harness greater interlimb coupling and result in higher average power in the beta and low gamma bands. Post-training, the paretic limb will be able to recruit larger motor units and fire at an increased frequency which has the potential to lead to greater force production.

PAGE 26

METHOD Subjects Chronic stroke subjects (N=15; 10 Left CVA, 5 Right CVA) were selected from two coupled EMG-neuromuscular stimulation and bilateral training protocols from 2002-2006. The number of participants was increased from 15 to 25 (N=25; 18 Left CVA, 7 Right CVA) because of an abundance of data, which also increases the power of the results. Volunteer participants had a stroke at least 1 year ago and displayed mild to moderate upper extremity chronic hemiparesis. Although participants were selected from two different studies, each subject participated in the same pre and posttests and participated in a coupled bilateral training and active neuromuscular stimulation protocol. All subjects met six admission criteria: (1) diagnosis of at least 1 CVA and no more than 2 CVAs on the same side of the brain; (2) an upper limit cutoff point of 80% motor recovery, as assessed by rectified EMG activation patterns and sustained force contractions during direct comparison of the impaired and unimpaired limbs; (3) a lower limit cutoff point of 10 of voluntary wrist or finger extension against gravity from a 90 flexed position; (4) absence of other neurological deficits, including a pacemaker; (5) no use of drugs for spasticity; and (6) no enrollment in another motor recovery rehabilitation protocol. Before testing began, subjects read and signed an informed consent form approved by the Institutional Review Board. 15

PAGE 27

16 Pretest-Posttest Instruments and Procedures Motor functions of the upper extremity were evaluated with three categories of measurement. The first category was a functional manual dexterity test termed the Box and Block timed manipulation test. For 1 minute, subjects attempted to reach and grasp a 2.54-cm cube, transport it over a short barrier, release the block, and return to the original side for another block. Figure 1-1 shows a subject performing the task (Cauraugh & Kim, 2002). Figure 2-1. Subject performing the Box and Block test with his impaired right hand. The other two categories of measurements were laboratory-based chronometric and force generation tasks: (1) simple reaction time for speed of information processing and rapid muscle onset and (2) sustained muscle contractions and force modulation (stability control). Simple reaction time tests and sustained muscle contractions were performed by the paretic limb alone, the nonparetic limb alone, and both limbs together in a randomized order for each subject. The hands and arms were inserted into separate devices so that isometric wrist/finger extension movements were executed against 11.4-kg load cells. For both arms, force and EMG signals were recorded online. EMG activity of

PAGE 28

17 the wrist/finger extensor muscles was recorded with surface electrodes (silversilver chloride electrodes with an epoxy-mounted preamplifier). Only the surface EMG data collected during the sustained contraction task were analyzed using spectral analysis. This was in an effort to target the motor unit recruitment strategies post-stroke, which are more clearly represented during a sustained contraction task. Motor Recovery Protocols: Training Procedures A treatment session started with stretching, and surface electrodes were attached to the extensor communis digitorum and extensor carpi ulnaris muscles of the impaired limb. The electrodes were connected to an Automove (AM 800) EMG Facilitation Stimulator microprocessor, and when a target threshold level of EMG activity was voluntarily achieved, the unit immediately provided a surface neuromuscular electric stimulation (i.e., 1-second ramp up, 5 seconds of biphasic stimulation at 50 Hz, pulse width of 200 s, 1-second ramp down, and mA range of 16 to 29) that assisted the muscles to execute a full range of motion. The initial threshold was set at 50 V, and as participants successfully achieved the target threshold level, the microprocessor unit automatically increased the target level slightly higher. If the threshold level was not met, then the unit decreased the threshold level closer to the amount of voluntary activity that the individual could produce. Trials were separated by 25 seconds of rest. During each day of training, subjects completed three sets of 30 successful EMG-triggered neuromuscular stimulation trials (approximately 1 hour and 30 minutes) according to the motor recovery protocol group assignments (i.e., bilateral or unilateral movements). The 6 hours of training (4 days) were completed during 2 weeks. This procedure was followed by one group (Coupled, N=16), while a second group (Weighted,

PAGE 29

18 N=9) underwent the same routine with an additional weight on the unaffected limb. The weight was approximately 100% of the unaffected hands mass. To accomplish this, subjects wore a nylon glove with pockets sewn to the top and bottom where small metal weights were placed. EMG Spectral Analysis Maximum sustained contractions were used to determine motor unit recruitment patterns. EMG signals were recorded at a sampling rate of 1000 Hz from the wrist extensor muscles, including the extensor carpi radialis longus, the extensor carpi ulnaris, and the extensor carpi radialis brevis. Signals were then lowpass filtered with a 55 Hz cutoff to avoid electrical outlet noise at 60 Hz. After cutting the first and last second off of each trial to avoid any inconsistencies during the sustained contraction test, the power spectrum of the EMG signal during 6 seconds of the sustained contraction phase was calculated. The proportion of power in each of four frequency bins 1 to 3 Hz, 5 to 12 Hz, 15 to 30 Hz, and 30 to 55 Hz was calculated. These frequency bins were used because of their established relation with population changes in motor unit activity (Myers et al., 2004), although only signal frequencies below 55 Hz were analyzed to avoid interference due to electrical outlet noise at 60 Hz. A window of 1,024 data points with a 50% overlap was used in isometric actions to calculate mean power frequency (MPF) and mean force (MF). Using the same FFT, average power for each of the four frequency bands was calculated. Statistical Design and Analyses Separate analyses were conducted on the three dependent measures. The average power for each frequency band was analyzed in a four-way mixed design ANOVA. The between-subjects factors were training group (2: coupled, weighted) and limb (2: paretic,

PAGE 30

19 nonparetic). The within-subjects factors were session (2: pretest, posttest) and impaired limb testing condition (2: bimanual, unimanual). The same mixed-design was used to analyze both mean power frequency and mean force.

PAGE 31

RESULTS The findings below include all of the experimental treatment data for 25 subjects. All dependent variables were analyzed in a Group Condition Limb Session (2 2 2 2) mixed ANVOA with repeated measures on the last two factors. Subjects were divided into two training groups, weighted and coupled, and performed wrist extension movements in bimanual and unimanual conditions with both affected and unaffected limbs. This process was followed in the pretest as well as the posttest. All statistical tests were conducted with set at 0.05. When appropriate, mean comparisons were computed with Tukeys HSD follow-up procedure. Power Band 1 (1-3 Hz) and Power Band 2 (5-12 Hz) No significant differences between groups or sessions in either power band were exhibited. There were also no significant differences found between the affected and unaffected limb in either the unimanual or bimanual condition in these lower two power bands. Power Band 3 (15-30 Hz) No significant main effects were identified for Condition, Limb, or Session. The analysis indicated a trend for a Group Condition interaction (F (1, 23) = 3.804, p < 0.07) exhibited in Figure 3-1. Subjects in the coupled protocol group exhibited higher power in the unimanual condition compared to subjects in the weighted group in both the unimanual and bimanual conditions. In addition, the analyses also revealed a trend for a Group Limb interaction (F (1, 23) = 3.721, p < 0.07) as seen in Figure 3-2. 20

PAGE 32

21 Power Band 3Group Condition Interaction0 0.0002 0.0004 0.0006 0.0008 0.0010.0012 0.0014 0.0016 0.0018 0.002Coupled Weighted Bimaual Unimanual Average Power V 2 Figure 3-1. Group condition interaction in power band 3 (15-30 Hz). Power Band 3Group Limb Interaction0 0.0005 0.0010.0015 0.0020.0025 Coupled Weighted A ffected Unaffected Average Power V 2 Figure 3-2. Group limb interaction in power band 3 (15-30 Hz).

PAGE 33

22 The affected limb in the coupled group exhibited a higher power that was statistically different from the unaffected limb in the coupled group as well as both the unaffected and affected limbs in the weighted group. The power of the affected limb in the weighted group was significantly lower than the affected limb in the coupled group in addition to the unaffected limbs in both groups. Power Band 4 (30-55 Hz) Analyses indicated a trend for a Limb Condition interaction (F (1, 23) = 3.616, p < 0.07) displayed in Figure 3-3. The unaffected limb in the bimanual condition showed a higher power than the affected limb in both unimanual and bimanual conditions. Further, the mixed ANOVA revealed a significant three-way interaction among Limb Condition Session (F (1, 23) = 4.417, p < 0.05) shown in Figure 3-4. The unaffected limb in the unimanual condition during the posttest exhibited significantly higher power than every other combination of condition, limb, and session; that is, the unaffected limb in the bimanual condition in the pre and posttest and the affected limb in either the unimanual or bimanual condition in the pre and posttest.

PAGE 34

23 Power Band 4Limb Condition Interaction0 0.0010.0020.0030.0040.0050.0060.007Bimanual Unimanual A ffected Unaffected Average Power V 2 Figure 3-3. Limb condition interaction in power band 4 (30-55 Hz). Power Band 4Limb Condition Session Interaction0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 Pretest Posttes t Bimanual Affected Bimanual Unaffected Unimanual Affected Unimanual Unaffected Average Power V 2 Figure 3-4. Limb condition session interaction in power band 4 (30-55 Hz).

PAGE 35

24 Mean Peak Frequency Analyses from the mixed ANOVA showed a significant main effect for Limb (F (1, 23) = 39.987, p < 0.001) shown in Figure 3-5. The unaffected limb showed a significantly higher mean peak frequency than the affected limb. Further, a significant Group Limb interaction (F (1, 23) = 5.165, p < 0.04) was found (see Figure 3-6). For the coupled group, the unaffected limb performed significantly higher in MPF than the affected limb in the same group. Moreover, the affected limb in the coupled group displayed a significantly lower MPF than the unaffected limb in the weighted group. Mean Peak FrequencyLimb Main Effect36.53737.53838.53939.54040.541AffectedUnaffectedHertz Figure 3-5. Mean peak frequency limb main effect.

PAGE 36

25 Mean Peak FrequencyGroup Limb Interaction35 36 37 38 39 40 41 Coupled WeightedHertz\ A ffected Unaffected Figure 3-6. Mean peak frequency group limb interaction. Force Analyses of the force data revealed a significant main effect for Condition (F (1, 23) = 8.749 p < 0.01). Subjects were capable of producing more force unimanually than bimanually. Further, the mixed ANOVA exhibited a significant main effect for Limb (F (1, 23) = 66.268, p < 0.0001). Comparing the two means indicated that the unaffected limb produced more force than the affected limb. Lastly, a significant Limb Condition interaction (F (1, 23) = 8.708, p < 0.01) was found. Follow-up test revealed that the affected limb in the bimanual condition produced a significantly lower amount of force when compared to the unaffected limb in both unimanual and bimanual conditions. The unaffected limb produced significantly less force bimanually than unimanually. Significant main effects and interaction are displayed in Figures 3-7, 3-8, and 3-9 respectively.

PAGE 37

26 ForceCondition Main Effect66.26.46.66.877.27.4BimanualUnimanualKilograms Figure 3-7. Condition main effect. ForceLimb Main Effect012345678910AffectedUnaffectedKilograms Figure 3-8. Limb main effect.

PAGE 38

27 ForceLimb Condition Interaction0 2 4 6 8 10 12 Bimanual UnimanualKilograms A ffected Unaffected Figure 3-9. Limb condition interaction.

PAGE 39

DISCUSSION Wrist and finger extension control is one of the most difficult movements to regain after stroke. Given that this control is the basis for many simple, everyday movements such as picking up a cup or turning a doorknob, it is extremely important to understand its rehabilitation and motor recovery. If researchers can identify the physiological basis for regaining motor function in the wrist and fingers, then potentially, stroke therapy could target these specific aspects of motor improvement, making therapy more effective and efficient. In accordance with the original hypotheses and previous research findings, there were no significant differences between the unaffected and affected limbs in power bands 1 and 2 (1-3 Hz and 5-12 Hz), nor was a training effect found. These results support the idea that the motor cortex does not control movements at low frequencies below 12 Hz and that the firing frequencies in the 1-3 Hz and 5-12 Hz bands remain unchanged post stroke (Myers et al., 2004). However, increased power was expected after training in the 15-30 Hz and 30-55 Hz power bands as a result of motor cortex rehabilitation. Although the effect for session was not significant, indicating that there were no differences between the pre and posttests, other results were evident. The coupled protocol group exhibited a higher power in the beta band (15-30 Hz) compared to the weighted group when moving unimanually. More specifically, the affected limb in the coupled protocol group showed a higher power than the unaffected limb in the same group as well as the affected and 28

PAGE 40

29 unaffected limbs of the weighted group. These findings do not indicate that the coupled training protocol is definitively more effective than the weighted protocol, but the specific subjects in the coupled group responded to that treatment better. Additional sources of variability include age, movement ability before training, and time since stroke; each may influence the efficacy of the training program. Further, the left hemisphere has been linked to bilateral movement control, so damage to this side may affect training performances (Wyke, 1971). In fact, 18 of 25 subjects in the present study suffered a left CVA, which may explain the absence of the statistically pretest-posttest effect. For the low gamma band (30-55 Hz), findings revealed a high correspondence with the current hypotheses and previous publications. The unaffected limb exhibited higher power than the affected limb, indicating that the unimpaired limb is able to harness larger motor neurons in its recruitment strategy. The recruitment of larger motor neurons theoretically leads to a larger force production, which is supported by this research. The most force is produced by the unimpaired limb when moving unimanually in the posttest; therefore inferring that the nonparetic hand tends to decrease its force production when coupled with the paretic hand. Research has shown that the nonparetic limb matches the movement time and trajectory of the paretic limb when moving together (Rose & Winstein, 2004). In addition, the current results support previous research (Van Dieen et al., 2003) in that the nonparetic limb attempted to match the paretic limb in force production. Although the affected limb did not exhibit a significant change after the coupled bimanual training protocol, in Figure 3-4, a trend for increased power in the low gamma band exists for the affected limb in the bimanual and unimanual conditions,

PAGE 41

30 though a steeper slope is exhibited in the bimanual condition. According to the hypotheses, the affected limb was expected to display an increase in power with a larger increase in the bimanual condition. Though Tukeys HSD post-hoc analyses did not reveal a significant increase, a trend for increased power in the low gamma band in the affected limb is evident. This trend for increased in power in the low gamma band translates into the ability to recruit larger motor neurons and therefore the potential to produce larger amounts of force. In an effort to explain the results revealed in the four power bands from 1-55 Hz, the entire continuous power spectrum was examined (personal communication, J. Garland, March 31, 2006). Instead of averaging power for each subject across a specific power band, the continuous power spectrum may reveal phenomenon not evident in averaging. After examining each subjects power spectrum, results were inconclusive. There were no trends for an increase or decrease in power in either limb. Subjects showed increases in both limbs (Figures 4-1 and 4-2), decreases in both limbs (Figures 4-3 and 4-4), and increases in one limb coupled with a decrease in the opposite limb (Figure 4-5 and 4-6). Although each subjects power spectrum displays a lower power in the affected limb compared to the unaffected limb in the pretest, each subject responded differently to the training which is clearly supported by the various power spectrum results displayed in Figures 4-1, 4-2, 4-3, 4-4, 4-5, and 4-6.

PAGE 42

31 PretestSubject 50030 0.020.040.060.080.10.120.141 5 9 1317 212529333741454953 A ffected Limb Unaffected Limb Figure 4-1. Power spectrum. Frequency (Hz) on x-axis. Power (V2) on y-axis. PosttestSubject 50030 0.02 0.04 0.06 0.08 0.1 0.12 0.14 1 5 9 13 17 212529333741454953 A ffected Limb Unaffected Limb Figure 4-2. Power spectrum. Frequency (Hz) on x-axis. Power (V2) on y-axis.

PAGE 43

32 PretestSubject 50050 0.002 0.004 0.006 0.008 0.010.012 0.014 1 5 9 13 17212529333741454953 Unaffected Limb A ffected Limb Figure 4-3. Power spectrum. Frequency (Hz) on x-axis. Power (V2) on y-axis. PosttestSubject 5005 0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 1 5 9 13 17 212529333741454953 Unaffected Limb A ffected Limb Figure 4-4. Power spectrum. Frequency (Hz) on x-axis. Power (V2) on y-axis.

PAGE 44

33 PretestSubject 50070 0.010.020.030.040.050.061 5 9 1317 212529333741454953 Unaffected Limb A ffected Limb Figure 4-5. Power spectrum. Frequency (Hz) on x-axis. Power (V 2 ) on y-axis. PosttestSubject 50070 0.01 0.02 0.03 0.04 0.05 0.06 1 5 9 13 17 212529333741454953 Unaffected Limb A ffected Limb Figure 4-6. Power spectrum. Frequency (Hz) on x-axis. Power (V 2 ) on y-axis.

PAGE 45

34 Efficiency is the conservation of effort to accomplish a given task goal involving an interaction with the environment and more specifically, a ratio of the effective work output and the energy input into any system (Hatfield & Hillman, 2001). Kraemer (1994) expounded on this idea of efficiency using a muscle endurance training example, stating that the level of motor unit activation needed to maintain a given submaximal force decreases as the skill is acquired. Additionally, Kraemer argued that improved endurance performance may result in neural activity changes and lead to more efficient performance with lower energy expenditure. In the field of motor unit recruitment, DeVries and Housh (1994) used EMG measurements to show a gradual decrease in motor unit recruitment as a result of resistance training. Even the simplest central nervous system involves many neuromuscular events overlapping in time. Kelso, Tuller, and Harris (1983) suggested a need for an organizing principle to direct the proper timing and activation of several sets of muscles for a coordinated action to occur. The precise facilitation and inhibition that must take place in a coordinated movement is not simply several actions at one joint, but a synergy of muscles and joints constrained to act as a single unit. In other words, a complex action with many degrees of freedom is simplified to a more efficient model by the central nervous system. According to an extension of the general adaptation syndrome (GAS) proposed by Hans Selye (1976), when an organism is initially confronted with a specific task, it generally responds in an inefficient manner by firing both relevant and irrelevant neural pathways. As skill level increases, the neural networks are able to suppress irrelevant pathways and coordinate a skillful movement. This learning process is termed pruning, described as clipping away

PAGE 46

35 ineffective pathways and clearly activating effective neural pathways which lead to successful movements (Bell & Fox, 1996). In the present pilot data, the average power decreases in the bimanual condition which indicates that the system is not recruiting larger motor units to create more force. Further, the system does not appear to be firing these motor units at a faster rate because the MPF does not increase in the bimanual condition. Therefore, one interpretation is that in the bimanual condition, the system is able to recruit more motor neurons of the same size and fire at the same frequency as in the unimanual condition to create greater force. In this method, smaller motor units are utilized to their fullest potential in an effort to preserve energy because they are fatigue-resistant. Based on this analysis, it can be concluded that after bimanual coordination training coupled with neuromuscular stimulation, stroke subjects perform more efficiently by producing a greater amount of force with less effort in a bimanual performance condition. Overall, following a bimanual coordination protocol coupled with neuromuscular stimulation, stroke subjects are able to produce greater force in both unimanual and bimanual conditions. In addition, the system is more efficient when performing bimanually as is evidenced by the decreased average power in the bimanual condition. Although stroke subjects appear to respond uniquely to the training administered, each group showed progress in the functional task, exhibiting a practical improvement. The coupled training protocol group showed improved performance in the Box and Block Test by increasing the average number of blocks moved by 8. Whereas, the weighted protocol group improved its average number of blocks moved by 3. In addition, though not statistically significant, there was an average increase when collapsing across subjects

PAGE 47

36 in power bands 3 and 4, as well as MPF and force. These averages are shown in Table 4-1. Table 4-1. Averages collapsed across subjects and groups. Dependent Measure Average Pretest Average Posttest Average Increase Power Band 3 (V 2 ) 0.00135715 0.00190377 +0.00054662 Power Band 4 (V 2 ) 0.00350579 0.00456317 +0.00105738 Mean Peak Frequency (Hz) 37.93329836 38.02371529 +0.09041693 Force (kg) 4.76181822 4.78578040 +0.02396218 Each subject exhibited improvement in at least one category, number of blocks moved, increased mean peak frequency, a change in power in the 30-55 Hz bands, or amount of force produced. However, no single uniform characteristic of motor improvement was shown by each stroke subject. Just as each stroke is unique and variable, so is each patients rehabilitation and adaptation strategy. The short training period and less intense training schedule, may not yield physiological changes in the central nervous system or the muscles of the wrist and finger extensors as could be expected with a more strenuous schedule. Physiological adaptations poststroke because of a training regimen may only be evident through highly intensive therapy for a longer period of time which was not administered in the present studies (Oullette et al., 2004). However, the improvements shown by each subject in the present protocol support the efficacy of both the weighted and coupled therapy. Whether range of motion, flexibility, strength, or a combination of these and other factors, is improved is unclear; however, the current findings indicate that the training is affecting the stroke subjects positively and this merits further research. In addition to discovering the type of training best suited for individual stroke survivors, it is also important to utilize spectral analysis as a

PAGE 48

37 noninvasive method of investigating motor unit recruitment. Although it is imperative to develop uniform subject groups when employing spectral analysis to eliminate strong variations which hide significant changes in the motor unit recruitment pattern of individuals, the analysis technique lends itself well to human trials because it uses surface EMG recordings. The use of spectral analysis to evaluate stroke motor recovery is important because is investigates recovery not a functional level, but at a physiological level. Employing spectral analysis in the investigation of motor unit recruitment changes in the stroke population is an important research avenue deserving of future attention.

PAGE 49

APPENDIX PILOT DATA Two subjects from a bimanual coordination training protocol were used as pilot data. The data are presented for subjects 1102 and 2102 in both the unimanual and bimanual conditions for preand post-tests. In the unimanual and bimanual conditions, both subjects show an increase in average power in the beta and low gamma bands, increased mean peak frequency, and increased force production in the paretic limb. Increased average power is indicative of a larger reliance on Type II muscle fibers, and thus an ability to recruit larger motor units (Bilodeau et al., 1995). Therefore, it can be concluded that after participation in the bimanual training protocol, stroke subjects improve their ability to capitalize on Hennemans size principle to produce larger force values. That is, they were able to recruit larger motor units that are capable of greater force output. This is a critical finding which demonstrates the effectiveness of the training protocol. The increased average power is displayed in the beta and low gamma frequency bands which are controlled by the motor cortex. Therefore, a logical conclusion is that bimanual coordination exercises coupled with neuromuscular stimulation activates and strengthens the motor cortex. Figures A-1, A-2, A-3, A-4 display the increased average power in the beta and low gamma bands for subjects 1102 and 2102 pre and posttraining in the unimanual and bimanual conditions. 38

PAGE 50

39 1102 Unimanual Condition0 0.0005 0.0010.0015 0.0020.0025 1-3 Hz 5-12 Hz15-30 Hz30-55 Hz Frequency BandsAverage Power V 2 Pretest Posttest Figure A-1. Average power (subject 1102) in pre and posttests in unimanual condition. 1102 Bimanual Condition0 0.0002 0.0004 0.0006 0.0008 0.0010.0012 0.0014 1-3 Hz 5-12 Hz15-30 Hz30-55 Hz Frequency BandsAverage Power V 2 Pretest Posttest Figure A-2. Average power (subject 1102) in pre and posttests in bimanual condition.

PAGE 51

40 2102 Unimanual Condition0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 1-3 Hz 5-12 Hz15-30 Hz30-55 Hz Frequency BandsAverage Power V 2 Pretest Posttest Figure A-3. Average power (subject 2102) in pre and posttests in unimanual condition. 2102 Bimanual Condition0 0.001 0.002 0.003 0.004 0.005 0.006 1-3 Hz 5-12 Hz15-30 Hz30-55 Hz Frequency BandsAverage Power V 2 Pretest Posttest Figure A-4. Average power (subject 2102) in pre and posttests in bimanual condition.

PAGE 52

41 Mean peak frequency (MPF) is indicative of the firing rate of individual motor units (Bilodeau et al., 1995). Posttests reveal increased MPF for both subjects (see Figures A-5, A-6, A-7, A-8). An increase in MPF also indicates more of a reliance on Type II muscle fibers, just as increased average power shows. Hence, the same conclusions can be deduced. After training, stroke subjects are not only using larger motor units but also firing at higher frequencies which are the underlying mechanisms of the increased force production. 1102 Unimanual Condition39.5 39.6 39.7 39.8 39.9 40 40.1 40.2 40.3 40.4 PretestPosttest Mean Peak Frequency (Hz) Figure A-5. Mean peak frequency (MPF) for subject 1102 in pre and posttests in unimanual condition.

PAGE 53

42 1102 Bimanual Condition38.7 38.8 38.9 39 39.1 39.2 39.3 39.4 39.5 39.6 PretestPosttest Mean Peak Frequency (Hz) Figure A-6. Mean peak frequency (MPF) for subject 1102 pre and posttests in bimanual condition. 2102 Unimanual Condition33.8 34 34.2 34.4 34.6 34.8 35 35.2 35.4 PretestPosttest Mean Peak Frequency (Hz) Figure A-7. Mean peak frequency (MPF) for subject 2102 pre and posttests in unimanual condition.

PAGE 54

43 2102 Bimanual Condition35.5 36 36.5 37 37.5 38 38.5 39 39.5 40 PretestPosttest Mean Peak Frequency (Hz) Figure A-8. Mean peak frequency (MPF) for subject 2102 pre and posttests in bimanual condition. Research has demonstrated that in both the unimanual and bimanual conditions, stroke subjects exhibit increased average power in the beta and low gamma bands, increased mean peak frequency, and increased force production, all as expected. When comparing the post-tests of the unimanual and bimanual conditions, a decrease in power coupled with an increase in force production is observed (Figures A-9 and A-10). MPF in these conditions stays essentially equal and therefore will not be considered a factor for increasing the force in the bimanual condition. The justification for an increase in force accompanied by a decrease in average power of the beta and low gamma bands can be explained through a model of efficiency.

PAGE 55

44 1102 Posttest0 0.5 1 1.5 2 2.5 3 3.5 4 UnimanualBimanual Mean Force (kg) Figure A-9. Mean force (subject 1102) in posttests in the unimanual vs. bimanual conditions. 2102 Posttest0 0.5 1 1.5 2 2.5 3 3.5 4 UnimanualBimanual Mean Force (kg) Figure A-10. Mean force (subject 2102) posttests in the unimanual vs. bimanual conditions.

PAGE 56

LIST OF REFERENCES Basmajian, J.V. and De Luca, C.V. Muscles alive. Baltimore, MD: Williams & Wilkins, 1985. Beaumont E, Gardiner P. Effects of daily spontaneous running on the electrophysiological properties of hindlimb motoneurones in rats. J Physiol. 2002;540:129-38. Bell, M. A., & Fox, N .A. Crawling experience is related to changes in cortical organization during infancy: Evidence from EEG coherence. Dev Psychobiol. 1996;29,551. Bilodeau M, Cincera M, Gervais S, Arsenault AB, Gravel D, Lepage Y, McKinley P. Changes in the electromyographic spectrum power distribution caused by a progressive increase in the force level. Eur J Appl Physiol Occup Physiol. 1995;71(2-3):113-23. Brown P. Cortical drives to human muscle: the Piper and related rhythms. Prog Neurobiol. 2000;60(1):97-108. Cauraugh JH. Coupled rehabilitation protocols and neural plasticity: Upper extremity improvements in chronic hemiparesis. Restor Neurol and Neurosci. 2002;22:337-347. Cauraugh JH, Kim SB. Two coupled motor recovery protocols are better than one: Electromyogram-triggered neuromuscular stimulation and bilateral movements. Stroke. 2002;33:1589. Cauraugh JH, Summers JJ. Neural plasticity and bilateral movements: A rehabilitation approach for chronic stroke. Prog Neurobiol. 2005;75(5):309-20. Cunningham CL, Stoykov ME, Walter CB. Bilateral facilitation of motor control in chronic hemiplegia. Acta Psychol. 2002;110: 321. Cunningham CL, Stoykov MEP, Walter CB. Bilateral facilitation of motor control in chronic hemiplegia. Acta Psychol. 2002;110:321-337. DeLuca CJ. Control properties of motor units. J Exp Biol. 1985;115:125-36. Denny-Brown D. On the nature of postural reflexes. Proc. R. Soc. Lond. B Biol. Sci. 1929;104:252-301. 45

PAGE 57

46 DeVries HA, Housh TJ. Physiology of exercise 5th Edition. Madison, Wisconsin: Brown and Benchmark, 1994. Ertas M, Stalberg E, Falck B: Can the size principle be detected in conventional EMG recordings? Muscle Nerve. 1995;18(4):435-439 Farina D, Macaluso A, Ferguson RA, and De Vito G. Effect of power, pedal rate, and force on average muscle fiber conduction velocity during cycling. J Appl Physiol. 2004;97(6):2035-2041. Farina D, Fosci M, Merletti, R. Motor unit recruitment strategies investigated by surface EMG variables. J Appl Physiology. 2002;92,235. Farmer SF, Bremner FD, Halliday DM, Rosenberg JR and Stephens JA. The frequency content of common presynaptic inputs to motoneurones studied during voluntary isometric contractions in man. J Physiol. 1993a; 470 :127. Farmer SF, Bremner FD, Halliday DM, Rosenberg JR and Stephens JA. Changes in motor-unit synchronisation following central nervous lesions in man. J Physiol. 1993b; 463 :83. Gardiner PF & Kernell D. The "fastness" of rat motoneurones: time-course of afterhyperpolarization in relation to axonal conduction velocity and muscle unit contractile speed. Pflugers Arch. 1990;415,762-766. Gemperline JJ, Allen S, Walk D, Rymer WZ. Characteristics of motor unit discharge in subjects with hemiparesis. Muscle Nerve 1995;18:1101. Granit R., Henatsch MD, and Steg G. Tonic and phasic ventral horn cells differentiated by post-tetanic potentiation in cat extensors. Acta Physiol Scand. 1956;37:114. Granit R, Kernell D, and Shortess KS. Quantitative aspects of repetitive firing of mammalian motoneurons, caused by injected currents. J Physiol. 1963;168:911-931. Hatfield BD, Hillman CH. The psychophysiology of sport: A mechanistic understanding of the psychology of superior performance. In R. Singer, H. Hausenblas, & C. Janelle (Eds.), Handbook of sport psychology. New York: Wiley & Sons, 2001;362-386. Henneman E, Somjen G, Carpenter DO: Functional significance of cell size in spinal motoneurons. J Neurophysiol. 1965;28:560-80. Hogrel JY. Clinical applications of surface electromyography in neuromuscular disorders. Neurophysiol Clin. 2005;35(2-3):59-71.

PAGE 58

47 Homberg V, Reiners K, Hefter H, and Freund HJ. The muscle activity spectrum: spectral analysis of muscle force as an estimator of overall motor unit activity. Electroencephalogr Clin Neurophysiol. 1986;63:209-222. Kamen G, DeLuca CJ. Firing rate interactions among human Orbicularis Oris motor units. Int J Neurosci. 1992; 64 :167. Karlsson S, Yu J, and Akay M. Timefrequency analysis of myoelectric signals during dynamic contractions: a comparative study. IEEE Trans Biomed Eng. 2000;47: 228. Kelso JAS, Tuller BH, Harris KS. A dynamic pattern perspective on the control and coordination of movement. In P. MacNeilage (Ed.), The Production of speech. New York: Springer-Verlag. 1983;137-173. Kernell, D. Organized variability in the neuromuscular system: a survey of task-related adaptations. Arch ItalBiol.1992;130:19. Kraemer, W. J. (1994). General adaptations to resistance and endurance training programs. In T. Baechle (Eds.), Essentials of strength training and conditioning (pp. 127-150). Champaign: Human Kinetics. Landau WM, Sahrmann SA. Preservation of directly stimulated muscle strength in hemiplegia due to stroke. Arch Neurol. 2002;59(9):1453-7. Leonard CT. The Neuroscience of human movement. St Louis, Mo: Mosby; 1998. Luff AR. and Atwood HL. Membrane properties and contraction of single muscle fibres in the mouse. Am J Physiol. 1972;222,1435. Mazevet D, Meunier S, Pradat-Diehl P, Marchand-Pauvert V, and Pierrot-Deseilligny E. Changes in propriospinally mediated excitation of upper limb motoneurons in stroke patients. Brain. 2003;126(4):988-1000. McComas AJ, Sica RE, Upton AR, and Aguilera N. Functional changes in motoneurones of hemiparetic patients. J Neurol Neurosurg Psychiatry. 1973;36:183-193. McComas AJ, Fawcett PRW, Campbell MJ, Sica REP. Electrophysiological estimation of the number of motor units within a human muscle. J Neurol Neurosurg Psychiatry. 1971; 34:121-131. Milner-Brown HS. Stein, RB, and Yemm R. Changes in firing rate of human motor units during linearly changing voluntary contractions. J Physiol. 1973;230:371. Morales FR, Engelhardt JK, Soja PJ, Pereda AE, and Chase MH. Motoneuron properties during motor inhibition produced by microinjection of carbachol into the pontine reticular formation of the decerebrate cat. J Neurophysiol. 1987;57:1118-1129.

PAGE 59

48 Moritani T, Muro M, and Nagata A. Intramuscular and surface electromyogram changes during muscle fatigue. J Appl Physiol. 1986;60:1179-1185. Mudie MH, Matyas TA. Can simultaneous bilateral movement involve the undamaged hemisphere in reconstruction of neural networks damaged by stroke? Disabil Rehabil. 2000;22:23 Myers LJ, Erim Z, and Lowery MM. Time and frequency domain methods for quantifying common modulation of motor unit firing patterns. J Neuroengineering Rehabil. 2004;1:2. Nudo RJ. Recovery after damage to motor cortical areas. Curr Opin Neurobiol. 1999;9:740. Olson CB, Carpenter DO, and Henneman E. Orderly recruitment of muscle action potentials. Arch Neurol. 1968;19:591-597. Ouellette MM, LeBrasseur NK, Bean JF, Phillips E, Stein J, Frontera WR, Fielding RA.Petajan JH. Stroke. 2004;35(6):1404-9. Petajan, JH. AAEM minimonograph #3: motor unit recruitment. Muscle Nerve. 1991;14(6):489-502. Priez A, Duchene J, Goubel F. Duchenne muscular dystrophy quantification: a multivariate analysis of surface EMG. Med Biol Eng Comput. 1992;30:283. Popovic D, Sinkjaer T. Control of Movement in the Physically Disabled. London, UK: Springer-Verlag; 2001 Rose J, McGill KC. The motor unit in spastic cerebral palsy. Develop Med Child Neurol, 1998;40:270-277. Rose DK, Winstein CJ. Bimanual training after stroke: are two hands better than one? Top Stroke Rehabil. 2004;11;20-30. Salenius S, Portin K, Kajola M, Salmelin R and Hari R. Cortical control of human motoneuron firing during isometric contraction. J Neurophysiol. 1997; 77: 34013405. Selye H. Stress in health and disease. Reading, MA: Butterworth. 1976. Toulouse P, Carrault G, Le Rumeur E, Coatrieux JL: Surface electromyogram automatic analysis and Guillain-Barr syndrome follow up. Electromyogr Clin Neurophysiol. 1992;32(1-2):51-62. Van Dieen JH, Futoshi O, De Haan A. Reduced neural drive in bilateral exertions: A performance-limiting factor. Med Sci Sports Exerc. 2003;35(1):111-118.

PAGE 60

49 Wakeling JM and Rozitis AI. Spectral properties of myoelectric signals from different motor units in the leg extensor muscles. J Exp Biol. 2004; 207:2519-2528. Walter CB, Swinnen JP. Kinetic attraction during bimanual coordination. J Mot Behav. 1990;22(4):451-73. Whitall J, Waller S, Silver K, Macko R. Repetitive bilateral arm training with rhythmic auditory cueing improves motor function in chronic hemiparetic stroke. Stroke. 2000;31:2390. William M. Landau, MD; Shirley A. Sahrmann, PhD. Preservation of directly stimulated muscle strength in hemiplegia due to stroke. Arch Neurol. 2002;59:1453-1457. Wyke, M. The effects of brain lesions on the performance of bilateral arm movements. Neuropsychologia. 1971; 9: 33.

PAGE 61

BIOGRAPHICAL SKETCH Lindsay McManus was born in North Carolina in 1982. She graduated with honors from North Carolina State University with a Bachelor of Science degree in biomedical engineering and biological engineering. During a summer internship in Cincinnati, Lindsay joined a research team studying stroke and spinal cord injury recovery, which eventually led her to the University of Florida in 2004. Lindsay has worked with Dr. James Cauraugh in the Motor Behavior Laboratory investigating chronic stroke recovery through bilateral training and EMG-neuromuscular stimulation. While at the University of Florida, Lindsay also worked in the Biomechanics Laboratory with Dr. John Chow and Dr. Mark Tillman. In addition, she was also an instructor for the undergraduate motor learning/control course for four semesters. After graduating with a Master of Science degree in August 2006, Lindsay plans to enter the field of medical devices and implants. 50


Permanent Link: http://ufdc.ufl.edu/UFE0015780/00001

Material Information

Title: Using EMG spectral analysis as a quantitative measure of motor unit recruitment to evaluate stroke motor recovery
Physical Description: xi, 50 p. ; ill.
Language: English
Creator: McManus, Lindsay A. ( Dissertant )
Cauraugh, James H. ( Thesis advisor )
Janelle, Chris ( Reviewer )
Tillman, Mark ( Reviewer )
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2006
Copyright Date: 2006

Subjects

Subjects / Keywords: Applied Physiology and Kinesiology thesis, M.S
Dissertations, Academic -- UF -- Applied Physiology and Kinesiology

Notes

Abstract: After stroke, motor function is often reestablished abnormally in the paretic limb, causing difficulty in daily tasks. The two primary purposes of this study were to apply spectral properties, namely mean peak frequency shifts and changes in average power over four frequency bins from surface EMG recordings, to describe motor unit recruitment in the muscles of the wrist/finger extensors, and to correlate functional improvements, namely increased force production, to improved motor unit recruitment patterns. Using data collected from 2002 through 2006 from two separate training protocols, fifteen chronic patients were selected for analysis. The first protocol consisted of a pretest, a 2 week (4 sessions) coupled bilateral training and active EMG-neuromuscular stimulation protocol, and a posttest. The second group (Weighted) underwent a pretest and posttest identical to previous group (Coupled), but participated in a coupled bilateral training and active EMG-neuromuscular stimulation with an added weight on the affected hand. Separate analyses were conducted on the three dependent measures (average power, mean peak frequency, and mean force) using a four-way mixed ANOVA. Significant results were revealed in the low gamma power band, in addition to mean peak frequency and mean force (p < 0.05). In the low gamma band, the unaffected limb in the unimanual condition during the posttest exhibited significantly higher power than every other combination of condition, limb, and session. The unaffected limb also shows a significantly higher mean peak frequency than the affected limb. The unaffected limb in the coupled group showed higher mean peak frequency than the affected limb in the same group. The affected limb in the coupled group also displays a lower mean peak frequency than the unaffected limb in the weighted group. After analyzing force data, subjects showed more force production capability when moving unimanually compared to moving bimanually. And as expected, the affected limb produced a lower amount of force compared to the unaffected limb in both unimanual and bimanual conditions. Findings support previous research in that the unaffected limb exhibits higher power than the affected limb, indicating that the unimpaired limb is able to harness larger motor neurons in its recruitment strategy which leads to larger force production. Each subject exhibited improvement in at least one category, number of blocks moved, increased mean peak frequency, a change in power in the 30-55 Hz bands, or amount of force produced. However, no single uniform characteristic of motor improvement was shown by each stroke subject. The current findings indicate that the training is affecting the stroke patients positively and merits further future research.
Subject: analysis, bilateral, EMG, motor, neuromuscular, recruitment, spectral, stimulation, stroke, training, unit
General Note: Title from title page of source document.
General Note: Document formatted into pages; contains 61 pages.
General Note: Includes vita.
Thesis: Thesis (M.S.)--University of Florida, 2006.
Bibliography: Includes bibliographical references.
Original Version: Text (Electronic thesis) in PDF format.

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0015780:00001

Permanent Link: http://ufdc.ufl.edu/UFE0015780/00001

Material Information

Title: Using EMG spectral analysis as a quantitative measure of motor unit recruitment to evaluate stroke motor recovery
Physical Description: xi, 50 p. ; ill.
Language: English
Creator: McManus, Lindsay A. ( Dissertant )
Cauraugh, James H. ( Thesis advisor )
Janelle, Chris ( Reviewer )
Tillman, Mark ( Reviewer )
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2006
Copyright Date: 2006

Subjects

Subjects / Keywords: Applied Physiology and Kinesiology thesis, M.S
Dissertations, Academic -- UF -- Applied Physiology and Kinesiology

Notes

Abstract: After stroke, motor function is often reestablished abnormally in the paretic limb, causing difficulty in daily tasks. The two primary purposes of this study were to apply spectral properties, namely mean peak frequency shifts and changes in average power over four frequency bins from surface EMG recordings, to describe motor unit recruitment in the muscles of the wrist/finger extensors, and to correlate functional improvements, namely increased force production, to improved motor unit recruitment patterns. Using data collected from 2002 through 2006 from two separate training protocols, fifteen chronic patients were selected for analysis. The first protocol consisted of a pretest, a 2 week (4 sessions) coupled bilateral training and active EMG-neuromuscular stimulation protocol, and a posttest. The second group (Weighted) underwent a pretest and posttest identical to previous group (Coupled), but participated in a coupled bilateral training and active EMG-neuromuscular stimulation with an added weight on the affected hand. Separate analyses were conducted on the three dependent measures (average power, mean peak frequency, and mean force) using a four-way mixed ANOVA. Significant results were revealed in the low gamma power band, in addition to mean peak frequency and mean force (p < 0.05). In the low gamma band, the unaffected limb in the unimanual condition during the posttest exhibited significantly higher power than every other combination of condition, limb, and session. The unaffected limb also shows a significantly higher mean peak frequency than the affected limb. The unaffected limb in the coupled group showed higher mean peak frequency than the affected limb in the same group. The affected limb in the coupled group also displays a lower mean peak frequency than the unaffected limb in the weighted group. After analyzing force data, subjects showed more force production capability when moving unimanually compared to moving bimanually. And as expected, the affected limb produced a lower amount of force compared to the unaffected limb in both unimanual and bimanual conditions. Findings support previous research in that the unaffected limb exhibits higher power than the affected limb, indicating that the unimpaired limb is able to harness larger motor neurons in its recruitment strategy which leads to larger force production. Each subject exhibited improvement in at least one category, number of blocks moved, increased mean peak frequency, a change in power in the 30-55 Hz bands, or amount of force produced. However, no single uniform characteristic of motor improvement was shown by each stroke subject. The current findings indicate that the training is affecting the stroke patients positively and merits further future research.
Subject: analysis, bilateral, EMG, motor, neuromuscular, recruitment, spectral, stimulation, stroke, training, unit
General Note: Title from title page of source document.
General Note: Document formatted into pages; contains 61 pages.
General Note: Includes vita.
Thesis: Thesis (M.S.)--University of Florida, 2006.
Bibliography: Includes bibliographical references.
Original Version: Text (Electronic thesis) in PDF format.

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0015780:00001


This item has the following downloads:


Table of Contents
    Title Page
        Page i
        Page ii
    Acknowledgement
        Page iii
        Page iv
    Table of Contents
        Page v
        Page vi
    List of Tables
        Page vii
    List of Figures
        Page viii
        Page ix
    Abstract
        Page x
        Page xi
    Introduction and review of literature
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
    Method
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
    Results
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
    Discussion
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
    Appendix: Pilot data
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
    References
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
    Biographical sketch
        Page 50
Full Text












USING EMG SPECTRAL ANALYSIS AS A QUANTITATIVE MEASURE OF
MOTOR UNIT RECRUITMENT TO EVALUATE STROKE MOTOR RECOVERY














By

LINDSAY ANN MCMANUS


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA


2006

































Copyright 2006

by

Lindsay Ann McManus















ACKNOWLEDGMENTS

I would like to thank my advisor and committee chairman, Dr. James Cauraugh, for

all of his guidance and support for the past two years. He helped to create an

environment here at the University of Florida that has utilized my strengths to their

highest potential. From teaching the undergraduate motor learning course to collecting

and analyzing data in the lab, Dr. Cauraugh has given me the chance to learn and

discover what interests me and what paths I will follow in the future. I also would like to

thank my other committee members, Dr. Chris Janelle and Dr. Mark Tillman, for their

suggestions and guidance through my thesis writing experience. In addition, I have to

say thank you to Steve Coombes who, first of all, made this thesis possible with his

LabView programming skills. Otherwise, I would never have been able to explore

something like spectral analysis. Steve has been a great person to work with in the lab,

share ideas with, and learn from. I know he will be a successful researcher one day and I

look forward to hearing about his innovative research in the future. Outside of the Motor

Behavior Lab, James Milford and Lori Gibbs have been very valuable resources for me

as a graduate student. I would like to thank James for all of his caring support from the

day I arrived all the way through my thesis writing and graduation. Lori has also been

extremely helpful in everything from helping me prepare classroom materials to

gathering paperwork for various tasks. Lastly, I would like to take the chance to thank

my family and friends for all of their love and support for the past two years. Some

friends are far away, but have given me the encouragement to pursue my degree. Other









friends I have met here at UF and I will value their friendship for years to come. My

family has always told me to follow my dreams and encouraged me to keep going above

and beyond what is expected. And with their love and support, I think that I have truly

made the most of my time here. I will always be proud to be a Florida Gator!
















TABLE OF CONTENTS



A C K N O W L E D G M E N T S ................................................................................................. iii

LIST OF TABLES ................................ ........ .............................. vii

L IS T O F F IG U R E S ..................................................... .. ............................................. v iii

INTRODUCTION AND REVIEW OF LITERATURE .................................................1...

N orm al M otor U nit R ecruitm ent ............................................................................ 1...
Recruitment Following Neurological Insult ................ ...................................3...
B ilateral T raining ....................................... . ...... ............ ................. ............ .. .5
Coupled Bilateral Training and EMG-Triggered Stimulation Protocols...................7...
A dding a Load to the U nim paired Lim b.................................................. ...............8...
Spectral A naly sis .............. ................................................ ............................ . 8
P u rp o se ...................................................................................................... ........ .. 12
H ypotheses .................................................................................. ....................... 12

M E T H O D ..........................................................................................................................1 5

Su bjects................................................ ........ .................... .....................15
Pretest-Posttest Instruments and Procedures .........................................................16
Motor Recovery Protocols: Training Procedures ..................................................17
E M G Spectral A analysis .. .................................................................... ............... 18
Statistical D esign and A nalyses............................................................. ............... 18

R E S U L T S .......................................................................................................... ........ .. 2 0

Power Band 1 (1-3 Hz) and Power Band 2 (5-12 Hz)...........................................20
P ow er B and 3 (15-30 H z) ................................................................... ................ 20
P ow er B and 4 (30-55 H z) ................................................................... ................ 22
M ean P eak F requ en cy .................................................................................................2 4
F o rc e ......................................................................................................... ....... .. 2 5

D IS C U S S IO N .................................................................................................................. .. 2 8

A PPEN D IX PILO T D A TA .................................................................... ................ 38

L IST O F R E FE R E N C E S ... ........................................................................ ................ 45



v









BIO GR APH ICAL SK ETCH .................................................................... ................ 50
















TABLE

Table page

4-1 Averages collapsed across subjects and groups ................................. ................ 36















LIST OF FIGURES


Figure page

2-1 Subject performing the box and block test with his impaired right hand.............. 16

3-1 Group x condition interaction in power band 3 (15-30 Hz). ................................21

3-2 Group x limb interaction in power band 3 (15-30 Hz) .....................................21

3-3 Limb x condition interaction in power band 4 (30-55 Hz)................................23

3-4 Limb x condition x session interaction in power band 4 (30-55 Hz)..................23

3-5 M ean peak frequency lim b m ain effect ............................................ ................ 24

3-6 M ean peak frequency group x limb interaction................................ ................ 25

3-7 C condition m ain effect ................................................................. ................ 26

3-8 Limb main effect ......................... .......... ........................26

3-9 L im b x condition interaction .................................... ..................... ................ 27

4-1 Power spectrum. Frequency (Hz) on x-axis. Power (utV2) on y-axis................31

4-2 Power spectrum. Frequency (Hz) on x-axis. Power (utV2) on y-axis................31

4-3 Power spectrum. Frequency (Hz) on x-axis. Power (utV2) on y-axis................32

4-4 Power spectrum. Frequency (Hz) on x-axis. Power (utV2) on y-axis................32

4-5 Power spectrum. Frequency (Hz) on x-axis. Power ([tV2) on y-axis.................33

4-6 Power spectrum. Frequency (Hz) on x-axis. Power (utV2) on y-axis................33

A-i Average power (subject 1102) in pre and posttests in unimanual condition ........39

A-2 Average power (subject 1102) in pre and posttests in bimanual condition ..........39

A-3 Average power (subject 2102) in pre and posttests in unimanual condition ........40

A-4 Average power (subject 2102) in pre and posttests in bimanual condition ..........40









A-5 Mean peak frequency (subject 1102) in pre and posttests unimanual condition...41

A-6 Mean peak frequency (subject 1102) pre and posttests bimanual condition .........42

A-7 Mean peak frequency (subject 2102) pre and posttests unimanual condition .......42

A-8 Mean peak frequency (subject 2102) pre and posttests bimanual condition .........43

A-9 Mean force (subject 1102) in posttests unimanual vs. bimanual conditions .........44

A-10 Mean force (subject 2102) posttests unimanual vs. bimanual conditions. ............44















Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

USING EMG SPECTRAL ANALYSIS AS A QUANTITATIVE MEASURE OF
MOTOR UNIT RECRUITMENT TO EVALUATE STROKE MOTOR RECOVERY


By

Lindsay Ann McManus

August 2006

Chair: James Cauraugh
Major Department: Applied Physiology and Kinesiology

After stroke, motor function is often reestablished abnormally in the paretic limb,

causing difficulty in daily tasks. The two primary purposes of this study were to apply

spectral properties, namely mean peak frequency shifts and changes in average power

over four frequency bins from surface EMG recordings, to describe motor unit

recruitment in the muscles of the wrist/finger extensors, and to correlate functional

improvements, namely increased force production, to improved motor unit recruitment

patterns.

Using data collected from 2002 through 2006 from two separate training protocols,

twenty five chronic patients were selected for analysis. The first protocol consisted of a

pretest, a 2 week (4 sessions) coupled bilateral training and active EMG-neuromuscular

stimulation protocol, and a posttest. The second group (Weighted) underwent a pretest

and posttest identical to previous group (Coupled), but participated in a coupled bilateral









training and active EMG-neuromuscular stimulation with an added weight on the affected

hand.

Separate analyses were conducted on the three dependent measures (average

power, mean peak frequency, and mean force) using a four-way mixed ANOVA.

Significant results were revealed in the low gamma power band, in addition to mean peak

frequency and mean force (p < 0.05). In the low gamma band, the unaffected limb in the

unimanual condition during the posttest exhibited significantly higher power than every

other combination of condition, limb, and session. Moreover, the unaffected limb

showed a significantly higher mean peak frequency than the affected limb. In addition,

the unaffected limb in the coupled group showed higher mean peak frequency than the

affected limb in the same group. The affected limb in the coupled group also displayed a

lower mean peak frequency than the unaffected limb in the weighted group. Analysis of

the force data revealed higher force when moving unimanually compared to moving

bimanually. As expected, the affected limb produced a lower amount of force compared

to the unaffected limb in both unimanual and bimanual conditions.

Findings support previous research stating that the unaffected limb exhibits higher

power than the affected limb, indicating that the unimpaired limb is able to harness larger

motor neurons in its recruitment strategy, thus leading to larger force production. Each

subject exhibited improvement in at least one category, number of blocks moved,

increased mean peak frequency, a change in power in the 30-55 Hz bands, or amount of

force produced. However, no single uniform characteristic of motor improvement was

shown by each stroke subject. The current findings indicate that the training is affecting

the stroke patients positively and merits further research















INTRODUCTION AND REVIEW OF LITERATURE

Normal Motor Unit Recruitment

The principle of motor unit recruitment states that motor units, consisting of the

motor neuron and the muscle fibers it innervates, are progressively activated with

increasing strength of voluntary muscle contraction (Milner-Brown et al., 1973). The

central nervous system increases the strength of muscle contraction by increasing the

number of active motor units or increasing the firing rate of individual motor units. Both

of these mechanisms occur concurrently, though the recruitment of more motor units

takes precedence over an increase in firing rate until nearly all motor units are recruited.

Once this level is reached, motor units are driven to fire in their secondary range to rates

greater than 50 Hz (Petajan, 1991).

Motor units are recruited according to the order of their size. The first motor units

to fire are small in size, weak in force generation, and resistant to fatigue. Progressively

larger motor units with greater force production capabilities are recruited with increasing

strength of muscle contraction (Henneman et al., 1965). As large motor units are

activated, it is important to note that the smaller units are still firing, although the large

motor units assume primary responsibility for maximum force production. This

sequential recruitment order is known as the Henneman size principle and it results in a

smooth increase in muscle strength. Through their feline research paradigm, Henneman

and colleagues discovered that axon diameter, conduction velocity, and cell size all

increase with increased muscle contraction (Henneman et al., 1965). The three main









types of motor units include Type I fibers, described as slow twitch, small, fatigue-

resistant units which generate the least force and have the slowest contraction time, Type

IIa units that are fast-twitch, larger, fatigue-resistant units able to generate greater forces

and faster contraction times, and Type IIx units, known as the largest, fast-twitch, easily-

fatigable units capable of the greatest force production and fastest contraction time.

Motor unit recruitment sequences begin with Type I motor units and then sequentially

advance to Type IIa and finally Type IIx units for high force production. The late

recruited Type II fibers have larger diameters and thus generate higher motor unit action

potentials (MUAPs) than the smaller Type I fibers. This size principle is used in

electromyography (EMG) studies designed to analyze motor unit recruitment patterns by

using the amplitude of the MUAP to define the size of a specific motor unit (Ertas et al.,

1995).

A confluence of MUAPs traveling along the sarcolemmas of the muscle fibers

results in a muscle contraction. Myoelectric signals are formed on the basis of the

interference pattern among these MUAPs and are detected using EMG. The conductance

and resistance of the sarcolemma differ between fast- and slow-twitch fibers (Luff &

Atwood, 1972) and therefore, researchers postulate that different types of muscle fibers

generate different MUAPs. Myoelectric signals provide information about the form of

the MUAPs and are able to offer insight into the electrophysiology of the active muscle.

Signals measured using intramuscular EMG in animals have shown that faster motor

units generate higher frequencies in their power spectra (Wakeling & Rozitis, 2004).

However, evidence from surface EMG in humans has remained elusive on this topic.

Recent advances in EMG signal decomposition including wavelet techniques, Fourier









Transform methods, and simulations that accurately mimic human muscle behavior have

provided additional information not previously known, such as motor unit recruitment,

muscle fatigue factors, and biomechanical adaptations (Karlsson et al., 2000). For

example, when faster motor units are active, increases in mean and/or median

myoelectric frequency are observed in the EMG signal, thus suggesting that there is an

inherent property of the muscle fiber that can be distinguished within EMG signals

(Farina et al., 2002).

Recruitment Following Neurological Insult

After a motor neuron injury such as a stroke, the relationship between motoneuron

size and the number and size of the muscle fibers it innervates is usually lost or damaged.

Reorganization of this relationship at motor endplates occurs over time through axon

branching and neuromuscular activity (Nudo, 1999). This reorganization, known as

neuroplasticity, occurs as the nervous system attempts to repair damaged motor neurons.

Therefore, there must be persistent functional changes evident in the brain, and

consequently in motor unit activity, to represent this new organization. Using spectral

analysis to analyze MUAPs post-stroke and after bimanual coordination training will

offer insight into abnormal motor unit recruitment strategies post-stroke and aid in the

explanation of new strategies employed after a specific training protocol.

After a cerebrovascular accident (CVA), such as stroke, motor unit recruitment is

altered because of interrupted descending pathways as well as structural changes in the

muscle (Mazevet et al., 2003). Brain damage secondary to stroke results in corticospinal

and supraspinal motor pathway disruption and possibly leads to synaptic degeneration at

the segmental level (McComas et al., 1973). This loss of neural signaling at the

segmental level causes motor neuron loss and altered force control mechanisms. Muscle









impairment following stroke can result from several sources including loss of motor units

because of Wallerian degeneration (McComas et al., 1971). As a result of the Wallerian

degeneration, all of the muscles fibers previously innervated by the axon are denervated

and therefore, fewer motor units are available for muscle activation. Altered properties

of the motor unit affect recruitment through altered recruitment order, discharge rate,

and/or discharge pattern (Beaumont & Gardiner, 2002). This abnormal motor unit

recruitment in turn, impacts the level of impairment post stroke. For instance, in

neurologically healthy individuals, once the first recruited muscle unit reaches a firing

frequency of 10 Hz, a second unit begins firing to increase muscle force. However, after

stroke, this second unit is often missing or damaged and an increase in force can only be

achieved by increasing the firing rate of the first motor unit(s). This abnormal

recruitment and discharge pattern contributes to the inefficient and undesired movement

patterns often observed in stroke patients. Adding to the underlying cause of such

inefficient motor patterns is altered motor neuron pool excitability. Altered excitability

affects the probability of motor unit activation, and in turn allows expression of unwanted

or atypical coordination patterns (Morales et al., 1987).

Low tonic rates are indicative of slow motor units while fast motor units are

associated with higher, more phasic tonic rates (Denny-Brown, 1929; Granit et al., 1956).

The close matching of motoneurons and muscle properties with respect to firing rates was

first described by Kernell and colleagues in muscles of rats and cats (Granit et al., 1963)

and later in single motor units of rats (Gardiner and Kernell, 1990) and supports the idea

that motor unit recruitment and rate coding both contribute to voluntary force production

(Kernell, 1992). That is, the number of motor units that are recruited and the rate of









firing of the motoneurons both control force generation of the muscle during voluntary

movement. Olson and colleagues successfully reported that a progressive recruitment of

motor units occurs when increasing force is required of voluntary movements (Olson et

al., 1968). However, impaired voluntary muscle strength results from neurological insult

because of a loss of normal motor unit recruitment occurs because of Wallerian

degeneration, altered recruitment and discharge patterns, and altered motor neuron pool

excitability (Landau & Sahrmann, 2002).

Moreover, drawing consistent conclusions from research involving impaired

muscle control in hemiplegic subjects is difficult because of the inconsistencies across

studies. For example, a limited number of subjects, a wide variation of lesion duration,

and multiple causes of disability are prevalent in stroke motor recovery research.

Therefore, the question as to whether the lack of muscular control because of disruption

at the supraspinal level, intrinsic changes in the motor neuron pool, or changes in the

properties of the muscle itself remains unanswered. However, the literature does agree

on the principles that motor unit activation in hemiplegia subjects is non-uniform and

motor unit firing rates tend to decrease overall relative to the ipsilesional limb (Rose &

McGill, 1998). A consequence of this diminished activation and firing of motor units is a

decreased ability to produce voluntary contractions.

Bilateral Training

Promising results have been reported in functional motor improvement post-stroke

with bilateral training protocols. Mudie and Matyas (2000) demonstrated enhanced

upper limb performance after a three week protocol of bilateral isokinematic therapy

(BIT). Bilateral practice with both upper limbs successfully elicited improvements in

unilateral performance of the affected limb alone. Motor improvements after bilateral









training were greater than those in response to similar unilateral practice or practice

sessions that involved the nonparetic limb assisting the paretic limb in various tasks.

Significant increases in strength, range of motion, and functional motor ability were

reported by Whitall and colleagues (2000) following a six week bilateral training

protocol. In addition, Cauraugh and Kim (2002) demonstrated enhanced motor control

after EMG-triggered neuromuscular stimulation coupled with bilateral training. These

three studies provide support for a bilateral training technique because of the potential

effectiveness of the various protocols in improving motor ability of individuals post

stroke.

The neural mechanisms and pathways targeted by bilateral training protocols

remain uncertain; however, several possibilities have been proposed. One pathway

involves an interaction in the affected hemisphere between surviving neurons of

ipsilateral pathways and surviving neurons of the crossed corticospinal pathways, the

involvement of indirect pathways from the affected hemisphere, and a facilitation of

coupling or crosstalk from the unaffected hemisphere to ipsilateral pathways are all

possible neural mechanisms responsible for motor improvement exhibited after bilateral

training (Mudie & Matyas, 2000). The idea of the coupling principle between arms

harnesses existing coupling abilities and encourages recovery of function through

bilateral movements. For example, forcing the use of both upper extremeties, as opposed

to the paretic arm only, has the potential to yield additional benefits by accessing intact

pathways of the unaffected hemisphere and possible coupling mechanisms. Rose and

Winstein (2004) reported clear temporal coupling in bimanual conditions and found a

facilitation effect in 5 of 11 patients where the paretic limb exhibited a shorter movement









time in the bimanual condition. Further, Cunningham and colleagues (2002) reported

coupling in bimanual training citing smoother movement of the paretic limb when it was

coupled with the nonparetic limb in elbow extension movements. Together these studies

support neural coupling as a possible mechanism underlying bilateral training protocols.

Coupled Bilateral Training and EMG-Triggered Stimulation Protocols

In separate studies of EMG-triggered neuromuscular stimulation and bilateral

coordination training, motor improvements in hemiparetic subjects were observed

(Leonard, 1998; Popovic and Sinkjaer, 2001). These findings are based on accepted

motor control theories, including sensorimotor integration and dynamical systems, which

complement one another in that the body is viewed as a dynamic system executing

movements through an interaction of sensory input and motor actions based on a set of

imposed constraints. Both bilateral training and EMG-triggered neuromuscular

stimulation essentially activate the same central and peripheral neural mechanisms

associated with the pyramidal tract and provide a basis for combining the two protocols.

The combination of the two interventions has the potential to accelerate the progress of

motor recovery after stroke (Cauraugh & Kim, 2002).

Specifically, Cauraugh and Kim reported an increase in performance of a functional

task, decreased reaction times, and increased capability in the sustained force task for the

coupled bilateral training and EMG-triggered stimulation group compared to the

unilateral and control groups. The coupled intervention group exhibited motor

improvement in all three testing categories, exhibiting decreased hemiparesis because of

each subject's expanded motor control repertoire (Cauraugh & Kim, 2002; Cauraugh,

2004).









Adding a Load to the Unimpaired Limb

Interlimb coupling is a potentially useful phenomenon in the context of stroke and

hemiplegia. It is beneficial to increase the symmetry between the impaired and

unimpaired limb through bilateral training in an effort to allow the unimpaired cortex to

facilitate the damaged cortex. Interlimb coupling increases with the frequency of the

movement, in addition to the torque requirements of the task (Walter and Swinnen,

1990). Therefore, through inertial loading of the unimpaired limb, torque is manipulated

and combined with the coupled bilateral training EMG-neuromuscular stimulation

protocol for one group in this study.

Spectral Analysis

Raw EMG signals offer valuable information, but in a relatively useless form.

Conclusions drawn from raw EMG signals cannot be quantitatively compared between

subjects, and therefore most analyses are purely qualitative. Pioneering EMG work

began in the 1950's and has undergone much refinement, especially in the past 20 years

due to the advanced processing of computers and other instrumentation. The formal

quantitative analysis of MUAP signals is not as cumbersome or time-consuming as it was

in the past thanks to recent advances. Among these advances is the ability to use the Fast

Fourier Transform (FFT) filter to break the EMG signal into its frequency components

which are then graphed as a function of their probability of occurrence and used to

determine muscle activation and fatigue.

EMG signal analysis in the frequency domain involves parameters describing

specific aspects of the frequency spectrum of the signal. The two most common

parameters of the power density spectrum used to determine fatigue are mean and median

frequency because they are easily applicable and provide useful information about the









EMG signal. If the EMG spectrum has a normal distribution, these two frequencies will

be equivalent, but deviation from normality will result in differing values. Spectral shifts

in surface EMG recordings are attributed to the type of muscle fibers activated and are

therefore used to characterize motor unit recruitment. Investigation into these shifts has

been limited to changes in the mean and median frequency of the power spectrum

calculated through FFT (Basmajian & DeLuca, 1985).

The amplitude of the surface EMG is related to the recruitment and the discharge

rates of the active motor units. Because of this relationship, EMG amplitude is used as an

index of the level of activation provided by the spinal cord. Using FFT, the mean and

median frequencies of the EMG signal are identified and analyzed. Shifts in the mean or

median frequencies should reflect the recruitment of progressively larger and faster motor

units based on Henneman's size principle. Stated differently, variation in muscle fiber

conduction velocity has been proposed as an index of motor unit recruitment according to

the size principle, and therefore should accurately illustrate motor unit recruitment

(Farina et al., 2004).

Data obtained from surface EMG techniques may not need to correlate directly

with traditional needle EMG data to be of value in the study of nerve and muscle

disorders. For instance, the decline in median frequency of the power spectrum with

fatigue or increasing force is a measure that has no single correlate in needle EMG

analysis. The combination of surface EMG and computerized signal processing has

allowed for nerve and muscle studies that had previously only been analyzed using

needle EMG technology. Toulouse and colleagues (1992) studied interference patterns in

persons with Guillain-Barre syndrome, a disorder characterized by deterioration in the









peripheral nervous system, and normal subjects. They found that surface EMG analysis

was useful in quantifying observed changes over the recovery period. In addition, Priez

et al. (2002) used surface EMG spectral analysis to discriminate between normal subjects

and those with muscular dystrophy. Further, the technique was successfully applied to

quantifying the severity of the disease. The information gathered by needle EMG is

invaluable and superior to surface EMG because of the specificity and accuracy of the

signals. However, researchers are encouraged to investigate the value of surface EMG

analysis and its ability to provide insight into the study of motor unit recruitment and

activity (Hogrel, 2005).

The lowest frequency at which motor units are modulated is 2 Hz. This "common

drive" is of central origin and has been found in muscles without muscle spindles

(DeLuca, 1985). However, it has only been demonstrated during isometric contractions,

where the force slowly increases or is constant (Farmer et al., 1993a). Low frequency

common drive persists after stroke, and this further supports the hypothesis that the

modulation is of central origin (Farmer et al., 1993b). The next phase of modulation in

motor unit firing occurs at a frequency of 5-12 Hz. To date, there is no convincing

evidence that this activity is directly controlled by the motor cortex, though peaks in this

frequency range have been reported in coherence spectra of the

magnetoencephalographic signal (MEG) from the motor cortex and EMG activity

(Salenius et al., 1997). The next motor unit activity frequencies include beta (15-30 Hz)

and low gamma (30-60 Hz) bands which are regulated by the motor cortex (Brown,

2000). Firing frequencies will vary based on the motor units, muscles, and persons. This

rate modulation depends on the intensity of descending motor input which is reinforced









by additional local and afferent inputs. Therefore, after stroke when conduction in the

central nervous system motor pathways is impaired, motor unit firing patterns are altered

(Gemperline et al. 1995).

Employing spectral analysis of muscle force as an estimator of motor unit activity

has shown that the activity spectra tends to shift to higher frequencies with increasing

force in normal subjects. Moritani and Muro (1986) reported increases in surface EMG

amplitude and mean power frequency with increasing force. They attributed these

changes to progressive increases in the motor unit firing frequency of initially recruited

motor units as well as newly recruited motor units. Thus, Moritani and Muro concluded

that surface EMG spectral analysis can provide a sensitive measure of motor unit activity

during force output. In addition to the normal motor unit activity recorded by surface

EMG, Homberg et al. (1986) selected patients with different motor dysfunctions to

illustrate how motor unit activity was reflected in the power spectra. They reported

decreased and increased firing rates, as well as abnormal synchronization across the

various motor dysfunctions.

Surface EMG has been shown to be a valuable tool in evaluating motor unit

activity. Although intracellular EMG technology is able to provide more detail and

insight into single motor unit recruitment, surface EMG data offers a significant means to

analyze muscle performance and fatigue. With the mathematical advances of spectral

analysis, it is now possible to assess EMG signals in their frequency domains and

investigate the effects of observed mean and median power shifts. Combining spectral

analysis and surface EMG technology is an interesting approach to analyzing motor unit

recruitment and, in turn, quantifying motor recovery after a nerve injury such as stroke.









Purpose

There were two primary purposes for this study: (1) to apply spectral properties,

namely mean peak frequency shifts and changes in average power over four frequency

bins from surface EMG recordings to describe motor unit recruitment in the muscles of

the wrist/finger extensors, and (2) to correlate functional improvements, namely

increased force production which has been previously reported (Cauraugh & Kim, 2002)

in stroke patients after bilateral training, to improved motor unit recruitment patterns.

Differences in the motor unit recruitment patterns of stroke subjects before and after

participation in a coupled EMG-neuromuscular stimulation and bilateral coordination

training protocol were analyzed. Recruitment patterns exhibited during a sustained

contraction task were quantified by the mean peak frequency and average power of EMG

recordings. Pre and post data were considered, in addition to unimanual pareticc limb

moving alone) versus bimanual pareticc and nonparetic limb moving together) testing

conditions.

Hypotheses

After participating in a motor recovery training protocol, force production of the

paretic limb is expected to increase. Several mechanisms may be responsible for this

increased force production. More specifically, increased force production occurs because

of either faster motor neuron firing rates, as measured by mean peak frequency, or

recruitment of larger, stronger motor neurons, quantified by average power. Overall, an

increase in mean peak frequency is expected in the posttest, which would indicate that

subjects are able to fire motor units at a faster rate than was possible before the coupled

protocol training. The increase in mean peak frequency is predicted to be greater in the

weighted group compared to the coupled group because of the increased torque demands.









In the range of 1-3 Hz, stroke patients should exhibit similar average power pre- and

post-training because this modulation has been reported to be preserved after stroke. If

frequencies in this lower band are preserved after infarct, then it is likely that they are not

controlled by the motor cortex and therefore possibly controlled at a sub-cortical level.

There is no convincing evidence for motor cortex control of the next frequency band, 5-

12 Hz, though peaks driven by the motor cortex have been reported. Hence, average

power of the paretic and nonparetic limbs should mirror each other closely and there will

be no significant differences between pre and posttests. In the next two bands, beta (15-

30 Hz) and low gamma (30-60 Hz), the paretic limb will exhibit a significantly lower

average power compared to the nonparetic limb. This modulation is controlled primarily

by the motor cortex which has been damaged by the stroke and therefore, unable to elicit

normal control over voluntary movements. The beta and low gamma bands are important

in higher aspects of motor control. After motor unit injury such as stroke, there is an

inability to shift to these higher frequencies and persons remain "locked" in a low

frequency mode (Myers et al., 2004). Bilateral training protocols aim to facilitate

crosstalk between the paretic and nonparetic limbs in an effort to increase the functional

ability of the paretic limb. Therefore, after two weeks (4 days) of bilateral coordination

training, motor unit recruitment in the paretic limb will improve in the beta and low

gamma bands. More specifically, it is hypothesized that the average power in the beta

and low gamma bands of the paretic limb will increase post-training in both the unilateral

and bilateral testing conditions with a larger increase in the bimanual condition compared

to the unimanual condition. A larger increase in average power in the beta and low

gamma bands for the bimanual condition will be the result of a facilitation effect from the






14


nonparetic hemisphere. A larger increase in power is also expected to be shown in the

beta and low gamma bands by the weighted group. The heightened torque demands of

the weighted group will harness greater interlimb coupling and result in higher average

power in the beta and low gamma bands. Post-training, the paretic limb will be able to

recruit larger motor units and fire at an increased frequency which has the potential to

lead to greater force production.















METHOD

Subjects

Chronic stroke subjects (N=15; 10 Left CVA, 5 Right CVA) were selected from

two coupled EMG-neuromuscular stimulation and bilateral training protocols from 2002-

2006. The number of participants was increased from 15 to 25 (N=25; 18 Left CVA, 7

Right CVA) because of an abundance of data, which also increases the power of the

results. Volunteer participants had a stroke at least 1 year ago and displayed mild to

moderate upper extremity chronic hemiparesis. Although participants were selected from

two different studies, each subject participated in the same pre and posttests and

participated in a coupled bilateral training and active neuromuscular stimulation protocol.

All subjects met six admission criteria: (1) diagnosis of at least 1 CVA and no more

than 2 CVAs on the same side of the brain; (2) an upper limit cutoff point of 80% motor

recovery, as assessed by rectified EMG activation patterns and sustained force

contractions during direct comparison of the impaired and unimpaired limbs; (3) a lower

limit cutoff point of 100 of voluntary wrist or finger extension against gravity from a 900

flexed position; (4) absence of other neurological deficits, including a pacemaker; (5) no

use of drugs for spasticity; and (6) no enrollment in another motor recovery rehabilitation

protocol. Before testing began, subjects read and signed an informed consent form

approved by the Institutional Review Board.









Pretest-Posttest Instruments and Procedures

Motor functions of the upper extremity were evaluated with three categories of

measurement. The first category was a functional manual dexterity test termed the Box

and Block timed manipulation test. For 1 minute, subjects attempted to reach and grasp a

2.54-cm cube, transport it over a short barrier, release the block, and return to the original

side for another block. Figure 1-1 shows a subject performing the task (Cauraugh & Kim,

2002).

















Figure 2-1. Subject performing the Box and Block test with his impaired right hand.

The other two categories of measurements were laboratory-based chronometric and

force generation tasks: (1) simple reaction time for speed of information processing and

rapid muscle onset and (2) sustained muscle contractions and force modulation (stability

control). Simple reaction time tests and sustained muscle contractions were performed by

the paretic limb alone, the nonparetic limb alone, and both limbs together in a

randomized order for each subject. The hands and arms were inserted into separate

devices so that isometric wrist/finger extension movements were executed against 11.4-kg

load cells. For both arms, force and EMG signals were recorded online. EMG activity of









the wrist/finger extensor muscles was recorded with surface electrodes (silver-silver

chloride electrodes with an epoxy-mounted preamplifier).

Only the surface EMG data collected during the sustained contraction task were

analyzed using spectral analysis. This was in an effort to target the motor unit

recruitment strategies post-stroke, which are more clearly represented during a sustained

contraction task.

Motor Recovery Protocols: Training Procedures

A treatment session started with stretching, and surface electrodes were attached to

the extensor communis digitorum and extensor carpi ulnaris muscles of the impaired

limb. The electrodes were connected to an Automove (AM 800) EMG Facilitation

Stimulator microprocessor, and when a target threshold level of EMG activity was

voluntarily achieved, the unit immediately provided a surface neuromuscular electric

stimulation (i.e., 1-second ramp up, 5 seconds of biphasic stimulation at 50 Hz, pulse

width of 200 is, 1-second ramp down, and mA range of 16 to 29) that assisted the

muscles to execute a full range of motion. The initial threshold was set at 50 iLV, and as

participants successfully achieved the target threshold level, the microprocessor unit

automatically increased the target level slightly higher. If the threshold level was not met,

then the unit decreased the threshold level closer to the amount of voluntary activity that

the individual could produce. Trials were separated by 25 seconds of rest.

During each day of training, subjects completed three sets of 30 successful EMG-

triggered neuromuscular stimulation trials (approximately 1 hour and 30 minutes)

according to the motor recovery protocol group assignments (i.e., bilateral or unilateral

movements). The 6 hours of training (4 days) were completed during 2 weeks. This

procedure was followed by one group (Coupled, N=16), while a second group (Weighted,









N=9) underwent the same routine with an additional weight on the unaffected limb. The

weight was approximately 100% of the unaffected hand's mass. To accomplish this,

subjects wore a nylon glove with pockets sewn to the top and bottom where small metal

weights were placed.

EMG Spectral Analysis

Maximum sustained contractions were used to determine motor unit recruitment

patterns. EMG signals were recorded at a sampling rate of 1000 Hz from the wrist

extensor muscles, including the extensor carpi radialis longus, the extensor carpi ulnaris,

and the extensor carpi radialis brevis. Signals were then lowpass filtered with a 55 Hz

cutoff to avoid electrical outlet noise at 60 Hz. After cutting the first and last second off

of each trial to avoid any inconsistencies during the sustained contraction test, the power

spectrum of the EMG signal during 6 seconds of the sustained contraction phase was

calculated. The proportion of power in each of four frequency bins 1 to 3 Hz, 5 to 12 Hz,

15 to 30 Hz, and 30 to 55 Hz was calculated. These frequency bins were used because of

their established relation with population changes in motor unit activity (Myers et al.,

2004), although only signal frequencies below 55 Hz were analyzed to avoid interference

due to electrical outlet noise at 60 Hz.

A window of 1,024 data points with a 50% overlap was used in isometric actions to

calculate mean power frequency (MPF) and mean force (MF). Using the same FFT,

average power for each of the four frequency bands was calculated.

Statistical Design and Analyses

Separate analyses were conducted on the three dependent measures. The average

power for each frequency band was analyzed in a four-way mixed design ANOVA. The

between-subjects factors were training group (2: coupled, weighted) and limb (2: paretic,






19


nonparetic). The within-subjects factors were session (2: pretest, posttest) and impaired

limb testing condition (2: bimanual, unimanual). The same mixed-design was used to

analyze both mean power frequency and mean force.















RESULTS

The findings below include all of the experimental treatment data for 25 subjects.

All dependent variables were analyzed in a Group x Condition x Limb x Session (2 x 2 x

2 x 2) mixed ANVOA with repeated measures on the last two factors. Subjects were

divided into two training groups, weighted and coupled, and performed wrist extension

movements in bimanual and unimanual conditions with both affected and unaffected

limbs. This process was followed in the pretest as well as the posttest. All statistical

tests were conducted with a set at 0.05. When appropriate, mean comparisons were

computed with Tukey's HSD follow-up procedure.

Power Band 1 (1-3 Hz) and Power Band 2 (5-12 Hz)

No significant differences between groups or sessions in either power band were

exhibited. There were also no significant differences found between the affected and

unaffected limb in either the unimanual or bimanual condition in these lower two power

bands.

Power Band 3 (15-30 Hz)

No significant main effects were identified for Condition, Limb, or Session. The

analysis indicated a trend for a Group x Condition interaction (F (1, 23) = 3.804, p <

0.07) exhibited in Figure 3-1. Subjects in the coupled protocol group exhibited higher

power in the unimanual condition compared to subjects in the weighted group in both the

unimanual and bimanual conditions. In addition, the analyses also revealed a trend for a

Group x Limb interaction (F (1, 23) = 3.721, p < 0.07) as seen in Figure 3-2.












Power Band 3
Group x Condition Interaction


--- Bimaual
--- Unimanual


Coupled Weighted

Figure 3-1. Group x condition interaction in power band 3 (15-30 Hz).


Power Band 3
Groupx Limb Interaction


-- Affected
--w- Unaffected


Coupled Weighted


Figure 3-2. Group x limb interaction in power band 3 (15-30 Hz).


0.002

0.0018

0.0016

0.0014

0.0012

0.001

0.0008

0.0006

0.0004

0.0002

0


0.0025



0.002

M

-0.0015

o
0.
0.001



0.0005



0









The affected limb in the coupled group exhibited a higher power that was

statistically different from the unaffected limb in the coupled group as well as both the

unaffected and affected limbs in the weighted group. The power of the affected limb in

the weighted group was significantly lower than the affected limb in the coupled group in

addition to the unaffected limbs in both groups.

Power Band 4 (30-55 Hz)

Analyses indicated a trend for a Limb x Condition interaction (F (1, 23) = 3.616, p

< 0.07) displayed in Figure 3-3. The unaffected limb in the bimanual condition showed a

higher power than the affected limb in both unimanual and bimanual conditions. Further,

the mixed ANOVA revealed a significant three-way interaction among Limb x Condition

x Session (F (1, 23) = 4.417, p < 0.05) shown in Figure 3-4. The unaffected limb in the

unimanual condition during the posttest exhibited significantly higher power than every

other combination of condition, limb, and session; that is, the unaffected limb in the

bimanual condition in the pre and posttest and the affected limb in either the unimanual

or bimanual condition in the pre and posttest.











Power Band 4
Limb x Condition Interaction


--- Affected
-->- Unaffected


Figure 3-3. Limb x


0.008

Sn00'7


J.

0.

0.

0.


006-

005-

004-


0.003

0.002

0.001

0


condition interaction in power band 4 (30-55 Hz).


Power Band 4
Limb x Condition x Session Interaction









F --- Bimanual Affected
--- Bimanual Unaffected
.... ......-.. Unimanual Unaffected
Unimanual Unaffected


Pretest


Posttest


Figure 3-4. Limb x condition x session interaction in power band 4 (30-55 Hz).


a



U-

p


0.007


0.006

0.005


0.004

0.003


0.002

0.001


0


Bimanual


Unimanual
















Mean Peak Frequency



Analyses from the mixed ANOVA showed a significant main effect for Limb (F (1,



23) = 39.987,p < 0.001) shown in Figure 3-5. The unaffected limb showed a



significantly higher mean peak frequency than the affected limb. Further, a significant



Group x Limb interaction (F (1, 23) = 5.165, p < 0.04) was found (see Figure 3-6). For



the coupled group, the unaffected limb performed significantly higher in MPF than the



affected limb in the same group. Moreover, the affected limb in the coupled group



displayed a significantly lower MPF than the unaffected limb in the weighted group.





Mean Peak Frequency

Limb Main Effect




41


40.5


40


39.5


39


I 385 F


38


37.5


37


36 5


Affected



Figure 3-5. Mean peak frequency limb main effect.


Unaffected


.............
............
.............
............
.............
............
.............
............
.............
............
.............
............
.............
............
.............
............
.............
............
.............
............
.............
...........











Mean Peak Frequency
Group x Limb Interaction

41

40
4 0 ------------------------------------------

39

38 -*- Affected
O 3 8 ---------^ ^----------
3 --u- Unaffected

37


36


35 -
Coupled Weighted

Figure 3-6. Mean peak frequency group x limb interaction.

Force

Analyses of the force data revealed a significant main effect for Condition (F (1,

23) = 8.749 p < 0.01). Subjects were capable of producing more force unimanually than

bimanually. Further, the mixed ANOVA exhibited a significant main effect for Limb (F

(1, 23) = 66.268,p < 0.0001). Comparing the two means indicated that the unaffected

limb produced more force than the affected limb. Lastly, a significant Limb x Condition

interaction (F (1, 23) = 8.708,p < 0.01) was found. Follow-up test revealed that the

affected limb in the bimanual condition produced a significantly lower amount of force

when compared to the unaffected limb in both unimanual and bimanual conditions. The

unaffected limb produced significantly less force bimanually than unimanually.

Significant main effects and interaction are displayed in Figures 3-7, 3-8, and 3-9

respectively.







































7.2





7





E 6.8


o

6.6





6.4





6.2








Bimanual





Figure 3-7. Condition main effect.


.............
.............
.............
.............
.............
.............
.............
.............
.............
.............
.............
.............
.............
.............
.............
.............
.............
.............
.............
.............
.............
.............
.............
.............
.............
.............
.............
.............
.............
.............
. . . .
. . . .


Affected


Figure 3-8. Limb main effect.


Force


Condition Main Effect


Unimanual


Force


Limb Main Effect


Unaffected












Force
Limb x Condition Interaction


- a

U- - -


- Affected
--u- Unaffected


Unimanual


Figure 3-9. Limb x condition interaction.


Bimanual















DISCUSSION

Wrist and finger extension control is one of the most difficult movements to regain

after stroke. Given that this control is the basis for many simple, everyday movements

such as picking up a cup or turning a doorknob, it is extremely important to understand

its rehabilitation and motor recovery. If researchers can identify the physiological basis

for regaining motor function in the wrist and fingers, then potentially, stroke therapy

could target these specific aspects of motor improvement, making therapy more effective

and efficient.

In accordance with the original hypotheses and previous research findings, there

were no significant differences between the unaffected and affected limbs in power bands

1 and 2 (1-3 Hz and 5-12 Hz), nor was a training effect found. These results support the

idea that the motor cortex does not control movements at low frequencies below 12 Hz

and that the firing frequencies in the 1-3 Hz and 5-12 Hz bands remain unchanged post

stroke (Myers et al., 2004).

However, increased power was expected after training in the 15-30 Hz and 30-55

Hz power bands as a result of motor cortex rehabilitation. Although the effect for session

was not significant, indicating that there were no differences between the pre and

posttests, other results were evident. The coupled protocol group exhibited a higher

power in the beta band (15-30 Hz) compared to the weighted group when moving

unimanually. More specifically, the affected limb in the coupled protocol group showed

a higher power than the unaffected limb in the same group as well as the affected and









unaffected limbs of the weighted group. These findings do not indicate that the coupled

training protocol is definitively more effective than the weighted protocol, but the

specific subjects in the coupled group responded to that treatment better. Additional

sources of variability include age, movement ability before training, and time since

stroke; each may influence the efficacy of the training program. Further, the left

hemisphere has been linked to bilateral movement control, so damage to this side may

affect training performances (Wyke, 1971). In fact, 18 of 25 subjects in the present study

suffered a left CVA, which may explain the absence of the statistically pretest-posttest

effect.

For the low gamma band (30-55 Hz), findings revealed a high correspondence with

the current hypotheses and previous publications. The unaffected limb exhibited higher

power than the affected limb, indicating that the unimpaired limb is able to harness larger

motor neurons in its recruitment strategy. The recruitment of larger motor neurons

theoretically leads to a larger force production, which is supported by this research. The

most force is produced by the unimpaired limb when moving unimanually in the posttest;

therefore inferring that the nonparetic hand tends to decrease its force production when

coupled with the paretic hand. Research has shown that the nonparetic limb matches the

movement time and trajectory of the paretic limb when moving together (Rose &

Winstein, 2004). In addition, the current results support previous research (Van Dieen et

al., 2003) in that the nonparetic limb attempted to match the paretic limb in force

production. Although the affected limb did not exhibit a significant change after the

coupled bimanual training protocol, in Figure 3-4, a trend for increased power in the low

gamma band exists for the affected limb in the bimanual and unimanual conditions,









though a steeper slope is exhibited in the bimanual condition. According to the

hypotheses, the affected limb was expected to display an increase in power with a larger

increase in the bimanual condition. Though Tukey's HSD post-hoc analyses did not

reveal a significant increase, a trend for increased power in the low gamma band in the

affected limb is evident. This trend for increased in power in the low gamma band

translates into the ability to recruit larger motor neurons and therefore the potential to

produce larger amounts of force.

In an effort to explain the results revealed in the four power bands from 1-55 Hz,

the entire continuous power spectrum was examined (personal communication, J.

Garland, March 31, 2006). Instead of averaging power for each subject across a specific

power band, the continuous power spectrum may reveal phenomenon not evident in

averaging. After examining each subject's power spectrum, results were inconclusive.

There were no trends for an increase or decrease in power in either limb. Subjects

showed increases in both limbs (Figures 4-1 and 4-2), decreases in both limbs (Figures 4-

3 and 4-4), and increases in one limb coupled with a decrease in the opposite limb

(Figure 4-5 and 4-6). Although each subject's power spectrum displays a lower power in

the affected limb compared to the unaffected limb in the pretest, each subject responded

differently to the training which is clearly supported by the various power spectrum

results displayed in Figures 4-1, 4-2, 4-3, 4-4, 4-5, and 4-6.










Pretest
Subject 5003


0.14
0.12
0.1
0.08 __ Affected Limb
0.06 ----Unaffected Limb

0.04
0.02- .


1 5 9 13 17 21 25 29 33 37 41 45 49 53

Figure 4-1. Power spectrum. Frequency (Hz) on x-axis. Power (utV2) on y-axis.


Posttest
Subject 5003

0.14
0.12 -- II
I t I % ,
0.1 I / -,
0.08 I Affected Limb
0.06 I ---- Unaffected Limb
I-
0.04- A A
0.02 -


1 5 9 13 1721 2529333741 454953

Figure 4-2. Power spectrum. Frequency (Hz) on x-axis. Power (tV2) on y-axis.






























Figure 4-3.


Power spectrum. Frequency (Hz) on x-axis. Power (pV2) on y-axis.


Posttest
Subject 5005


0.014
0.012


u.u
0.00
0.00
0.00
0.00


1-
8-
6-
)4 A W MAA
V \-,


-


1 5 9 13 17 21 25 29 33 37 41 45 49 53
1 5 9 13 17 21 25 29 33 37 41 45 49 53


Figure 4-4. Power spectrum. Frequency (Hz) on x-axis. Power (pV2) on y-axis.


Pretest
Subject 5005


0.014-
0.012 1. A
0.01
0.008 & __Unaffected Limb
0. 006 ---- Affected Limb
0.004- -
0.002-


1 5 9 13 1721 25 29 333741 45 49 53


--Unaffected Limb
---- Affected Limb
































Power spectrum. Frequency (Hz) on x-axis. Power ([tV2) on y-axis.


Posttest
Subject 5007


0.06

0.05
rn r\Ai


u.u

0.0

0.0

0.0


: ,^;;
3

2 1' W
1 5


1 5 9 13 17 21 25 29 33 37 41 45 49 53


-- Unaffected Limb
----Affected Limb


Power spectrum. Frequency (Hz) on x-axis. Power (uV2) on y-axis.


Pretest
Subject 5007


0.06

0.05

0.04
OO__ Unaffected Limb
0.03 N
S---- Affected Limb
0.02 ,. -

0.01 .


1 5 9 13 17 21 25 29 33 37 41 45 49 53


Figure 4-5.


A I, ."


Figure 4-6.









Efficiency is the conservation of effort to accomplish a given task goal involving an

interaction with the environment and more specifically, a ratio of the effective work

output and the energy input into any system (Hatfield & Hillman, 2001). Kraemer (1994)

expounded on this idea of efficiency using a muscle endurance training example, stating

that the level of motor unit activation needed to maintain a given submaximal force

decreases as the skill is acquired. Additionally, Kraemer argued that improved endurance

performance may result in neural activity changes and lead to more efficient performance

with lower energy expenditure. In the field of motor unit recruitment, DeVries and

Housh (1994) used EMG measurements to show a gradual decrease in motor unit

recruitment as a result of resistance training. Even the simplest central nervous system

involves many neuromuscular events overlapping in time. Kelso, Tuller, and Harris

(1983) suggested a need for an organizing principle to direct the proper timing and

activation of several sets of muscles for a coordinated action to occur. The precise

facilitation and inhibition that must take place in a coordinated movement is not simply

several actions at one joint, but a synergy of muscles and joints constrained to act as a

single unit. In other words, a complex action with many degrees of freedom is simplified

to a more efficient model by the central nervous system. According to an extension of

the general adaptation syndrome (GAS) proposed by Hans Selye (1976), when an

organism is initially confronted with a specific task, it generally responds in an inefficient

manner by firing both relevant and irrelevant neural pathways. As skill level increases,

the neural networks are able to suppress irrelevant pathways and coordinate a skillful

movement. This learning process is termed "pruning", described as clipping away









ineffective pathways and clearly activating effective neural pathways which lead to

successful movements (Bell & Fox, 1996).

In the present pilot data, the average power decreases in the bimanual condition

which indicates that the system is not recruiting larger motor units to create more force.

Further, the system does not appear to be firing these motor units at a faster rate because

the MPF does not increase in the bimanual condition. Therefore, one interpretation is

that in the bimanual condition, the system is able to recruit more motor neurons of the

same size and fire at the same frequency as in the unimanual condition to create greater

force. In this method, smaller motor units are utilized to their fullest potential in an effort

to preserve energy because they are fatigue-resistant. Based on this analysis, it can be

concluded that after bimanual coordination training coupled with neuromuscular

stimulation, stroke subjects perform more efficiently by producing a greater amount of

force with less effort in a bimanual performance condition.

Overall, following a bimanual coordination protocol coupled with neuromuscular

stimulation, stroke subjects are able to produce greater force in both unimanual and

bimanual conditions. In addition, the system is more efficient when performing

bimanually as is evidenced by the decreased average power in the bimanual condition.

Although stroke subjects appear to respond uniquely to the training administered,

each group showed progress in the functional task, exhibiting a practical improvement.

The coupled training protocol group showed improved performance in the Box and Block

Test by increasing the average number of blocks moved by 8. Whereas, the weighted

protocol group improved its average number of blocks moved by 3. In addition, though

not statistically significant, there was an average increase when collapsing across subjects









in power bands 3 and 4, as well as MPF and force. These averages are shown in Table 4-

1.

Table 4-1. Averages collapsed across subjects and groups.

Dependent Measure Average Pretest Average Posttest Average Increase
Power Band 3 ([tV2) 0.00135715 0.00190377 +0.00054662

Power Band 4 ([tV2) 0.00350579 0.00456317 +0.00105738

Mean Peak Frequency (Hz) 37.93329836 38.02371529 +0.09041693

Force (kg) 4.76181822 4.78578040 +0.02396218


Each subject exhibited improvement in at least one category, number of blocks

moved, increased mean peak frequency, a change in power in the 30-55 Hz bands, or

amount of force produced. However, no single uniform characteristic of motor

improvement was shown by each stroke subject. Just as each stroke is unique and

variable, so is each patient's rehabilitation and adaptation strategy. The short training

period and less intense training schedule, may not yield physiological changes in the

central nervous system or the muscles of the wrist and finger extensors as could be

expected with a more strenuous schedule. Physiological adaptations poststroke because

of a training regimen may only be evident through highly intensive therapy for a longer

period of time which was not administered in the present studies (Oullette et al., 2004).

However, the improvements shown by each subject in the present protocol support the

efficacy of both the weighted and coupled therapy. Whether range of motion, flexibility,

strength, or a combination of these and other factors, is improved is unclear; however, the

current findings indicate that the training is affecting the stroke subjects positively and

this merits further research. In addition to discovering the type of training best suited for

individual stroke survivors, it is also important to utilize spectral analysis as a









noninvasive method of investigating motor unit recruitment. Although it is imperative to

develop uniform subject groups when employing spectral analysis to eliminate strong

variations which hide significant changes in the motor unit recruitment pattern of

individuals, the analysis technique lends itself well to human trials because it uses surface

EMG recordings. The use of spectral analysis to evaluate stroke motor recovery is

important because is investigates recovery not a functional level, but at a physiological

level. Employing spectral analysis in the investigation of motor unit recruitment changes

in the stroke population is an important research avenue deserving of future attention.














APPENDIX
PILOT DATA

Two subjects from a bimanual coordination training protocol were used as pilot

data. The data are presented for subjects 1102 and 2102 in both the unimanual and

bimanual conditions for pre- and post-tests. In the unimanual and bimanual conditions,

both subjects show an increase in average power in the beta and low gamma bands,

increased mean peak frequency, and increased force production in the paretic limb.

Increased average power is indicative of a larger reliance on Type II muscle fibers,

and thus an ability to recruit larger motor units (Bilodeau et al., 1995). Therefore, it can

be concluded that after participation in the bimanual training protocol, stroke subjects

improve their ability to capitalize on Henneman's size principle to produce larger force

values. That is, they were able to recruit larger motor units that are capable of greater

force output. This is a critical finding which demonstrates the effectiveness of the

training protocol. The increased average power is displayed in the beta and low gamma

frequency bands which are controlled by the motor cortex. Therefore, a logical

conclusion is that bimanual coordination exercises coupled with neuromuscular

stimulation activates and strengthens the motor cortex. Figures A-I, A-2, A-3, A-4

display the increased average power in the beta and low gamma bands for subjects 1102

and 2102 pre and posttraining in the unimanual and bimanual conditions.












1102 Unimanual Condition


0.0025



0.002



o-0.0015-
E Pretest
a-. m Posttest
0.001



0.0005



0o
1-3 Hz 5-12 Hz 15-30 Hz 30-55 Hz
Frequency Bands

Figure A-1. Average power (subject 1102) in pre and posttests in unimanual condition.


1102 Bimanual Condition


0.0014


0.0012


0.001


0.0008
o
o 0.0006


> 0.0004


0.0002


0


E Pretest
* Posttest


5-12 Hz 15-30 Hz
Frequency Bands


Average power (subject 1102) in pre and posttests in bimanual condition.


1-3 Hz


30-55 Hz
30-55 Hz


- -


Figure A-2.












2102 Unimanual Condition


0.014


0.012


0.01


0.008


0.006


0.004


0.002


0


3 Pretest
* Posttest


30-55 Hz


1-3 Hz


5-12 Hz 15-30 Hz
Frequency Bands


Figure A-3. Average power (subject 2102) in pre and posttests in unimanual condition.


2102 Bimanual Condition


0.006


0.005


0.004


0.003


0.002


0.001


0


3 Pretest
* Posttest


30-55 Hz


FTh~


5-12 Hz 15-30 Hz
Frequency Bands


1-3 Hz


Figure A-4. Average power (subject 2102) in pre and posttests in bimanual condition.


- -










Mean peak frequency (MPF) is indicative of the firing rate of individual motor

units (Bilodeau et al., 1995). Posttests reveal increased MPF for both subjects (see

Figures A-5, A-6, A-7, A-8). An increase in MPF also indicates more of a reliance on

Type II muscle fibers, just as increased average power shows. Hence, the same

conclusions can be deduced. After training, stroke subjects are not only using larger

motor units but also firing at higher frequencies which are the underlying mechanisms of

the increased force production.


1102 Unimanual Condition

40.4

40.3

40.2

>, 40.1 -
U
4 40
L.
39.9

39.8

S39.7

39.6

39.5
Pretest Posttest

Figure A-5. Mean peak frequency (MPF) for subject 1102 in pre and posttests in
unimanual condition.








































Figure A-6. Mean peak frequency (MPF) for subject 1102 pre and posttests in bimanual
condition.


2102 Unimanual Condition


35.4

35.2

N 35

S34.8

2 34.6
U-

34.4-

34.2

34

33.8
Pretest Posttest

Figure A-7. Mean peak frequency (MPF) for subject 2102 pre and posttests in unimanual
condition.


1102 Bimanual Condition


39.6

39.5

39.4
N
>, 39.3 -

S39.2
L.
_ 39.1

c39-

S38.9

38.8

38.7


Pretest


Posttest





































Figure A-8. Mean peak frequency (MPF) for subject 2102 pre and posttests in bimanual
condition.

Research has demonstrated that in both the unimanual and bimanual conditions,

stroke subjects exhibit increased average power in the beta and low gamma bands,

increased mean peak frequency, and increased force production, all as expected. When

comparing the post-tests of the unimanual and bimanual conditions, a decrease in power

coupled with an increase in force production is observed (Figures A-9 and A-10). MPF

in these conditions stays essentially equal and therefore will not be considered a factor for

increasing the force in the bimanual condition. The justification for an increase in force

accompanied by a decrease in average power of the beta and low gamma bands can be

explained through a model of efficiency.


2102 Bimanual Condition

40

39.5

39
N
"I
>, 38.5
U
S38-
L.
_ 37.5

,-37-

S36.5

36

35.5


Pretest


Posttest












1102 Posttest


4

3.5

3

F2.5

2
2-
cc
1.5




0.5

0
Unimanual Bimanual

Figure A-9. Mean force (subject 1102) in posttests in the unimanual vs. bimanual
conditions.



2102 Posttest

4

3.5

3

-2.5

,, 2
C
S1.5




0.5

0
Unimanual Bimanual

Figure A-10. Mean force (subject 2102) posttests in the unimanual vs. bimanual
conditions.
















LIST OF REFERENCES


Basmajian, J.V. and De Luca, C.V. Muscles alive. Baltimore, MD: Williams & Wilkins,
1985.

Beaumont E, Gardiner P. Effects of daily spontaneous running on the
electrophysiological properties of hindlimb motoneurones in rats. JPhysiol.
2002;540:129-38.

Bell, M. A., & Fox, N .A. Crawling experience is related to changes in cortical
organization during infancy: Evidence from EEG coherence. Dev Psychobiol.
1996;29,551-561.

Bilodeau M, Cincera M, Gervais S, Arsenault AB, Gravel D, Lepage Y, McKinley P.
Changes in the electromyographic spectrum power distribution caused by a
progressive increase in the force level. Eur JAppl Physiol Occup Physiol.
1995;71(2-3):113-23.

Brown P. Cortical drives to human muscle: the Piper and related rhythms. Prog
Neurobiol. 2000;60(1):97-108.

Cauraugh JH. Coupled rehabilitation protocols and neural plasticity: Upper extremity
improvements in chronic hemiparesis. Restor Neurol andNeurosci. 2002;22:337-
347.

Cauraugh JH, Kim SB. Two coupled motor recovery protocols are better than one:
Electromyogram-triggered neuromuscular stimulation and bilateral movements.
Stroke. 2002;33:1589.

Cauraugh JH, Summers JJ. Neural plasticity and bilateral movements: A rehabilitation
approach for chronic stroke. Prog Neurobiol. 2005;75(5):309-20.

Cunningham CL, Stoykov ME, Walter CB. Bilateral facilitation of motor control in
chronic hemiplegia. Acta Psychol. 2002;110: 321-337.

Cunningham CL, Stoykov MEP, Walter CB. Bilateral facilitation of motor control in
chronic hemiplegia. Acta Psychol. 2002; 110:321-337.

DeLuca CJ. Control properties of motor units. JExp Biol. 1985;115:125-36.

Denny-Brown D. On the nature of postural reflexes. Proc. R. Soc. Lond. B Biol. Sci.
1929;104:252-301.









DeVries HA, Housh TJ. Physiology of exercise 5th Edition. Madison, Wisconsin: Brown
and Benchmark, 1994.

Ertas M, Stalberg E, Falck B: Can the size principle be detected in conventional EMG
recordings? Muscle Nerve. 1995;18(4):435-439

Farina D, Macaluso A, Ferguson RA, and De Vito G. Effect of power, pedal rate, and
force on average muscle fiber conduction velocity during cycling. JAppl Physiol.
2004;97(6):2035-2041.

Farina D, Fosci M, Merletti, R. Motor unit recruitment strategies investigated by surface
EMG variables. JAppl Physiology. 2002;92,235-247.

Farmer SF, Bremner FD, Halliday DM, Rosenberg JR and Stephens JA. The frequency
content of common presynaptic inputs to motoneurones studied during voluntary
isometric contractions in man. JPhysiol. 1993a; 470:127-155.

Farmer SF, Bremner FD, Halliday DM, Rosenberg JR and Stephens JA. Changes in
motor-unit synchronisation following central nervous lesions in man. JPhysiol.
1993b;463:83-105.

Gardiner PF & Kernell D. The "fastness" of rat motoneurones: time-course of
afterhyperpolarization in relation to axonal conduction velocity and muscle unit
contractile speed. Pflugers Arch. 1990;415,762-766.

Gemperline JJ, Allen S, Walk D, Rymer WZ. Characteristics of motor unit discharge in
subjects with hemiparesis. Muscle Nerve 1995; 18:1101-14.


Granit R., Henatsch MD, and Steg G. Tonic and phasic ventral horn cells differentiated
by post-tetanic potentiation in cat extensors. Acta Physiol Scand. 1956;37:114-126.

Granit R, Kernell D, and Shortess KS. Quantitative aspects of repetitive firing of
mammalian motoneurons, caused by injected currents. JPhysiol. 1963; 168:911-
931.

Hatfield BD, Hillman CH. The psychophysiology of sport: A mechanistic understanding
of the psychology of superior performance. In R. Singer, H. Hausenblas, & C.
Janelle (Eds.), Handbook of sport psychology. New York: Wiley & Sons,
2001;362-386.

Henneman E, Somjen G, Carpenter DO: Functional significance of cell size in spinal
motoneurons. JNeurophysiol. 1965;28:560-80.

Hogrel JY. Clinical applications of surface electromyography in neuromuscular
disorders. Neurophysiol Clin. 2005;35(2-3):59-71.









Homberg V, Reiners K, Hefter H, and Freund HJ. The muscle activity spectrum: spectral
analysis of muscle force as an estimator of overall motor unit activity.
Electroencephalogr Clin Neurophysiol. 1986;63:209-222.

Kamen G, DeLuca CJ. Firing rate interactions among human Orbicularis Oris motor
units. Int JNeurosci. 1992;64:167-175.

Karlsson S, Yu J, and Akay M. Time-frequency analysis of myoelectric signals during
dynamic contractions: a comparative study. IEEE Trans BiomedEng. 2000;47:
228-237.

Kelso JAS, Tuller BH, Harris KS. A "dynamic pattern" perspective on the control and
coordination of movement. In P. MacNeilage (Ed.), The Production of speech.
New York: Springer-Verlag. 1983;137-173.

Kernell, D. Organized variability in the neuromuscular system: a survey of task-related
adaptations. Arch ItalBiol. 1992; 130:19-66.

Kraemer, W. J. (1994). General adaptations to resistance and endurance training
programs. In T. Baechle (Eds.), Essentials of strength training and conditioning
(pp. 127-150). Champaign: Human Kinetics.

Landau WM, Sahrmann SA. Preservation of directly stimulated muscle strength in
hemiplegia due to stroke. Arch Neurol. 2002;59(9): 1453-7.

Leonard CT. The Neuroscience of human movement. St Louis, Mo: Mosby; 1998.

Luff AR. and Atwood HL. Membrane properties and contraction of single muscle fibres
in the mouse. Am JPhysiol. 1972;222,1435-1440.

Mazevet D, Meunier S, Pradat-Diehl P, Marchand-Pauvert V, and Pierrot-Deseilligny E.
Changes in propriospinally mediated excitation of upper limb motoneurons in
stroke patients. Brain. 2003;126(4):988-1000.

McComas AJ, Sica RE, Upton AR, and Aguilera N. Functional changes in motoneurones
of hemiparetic patients. JNeurol Neurosurg Psychiatry. 1973;36:183-193.

McComas AJ, Fawcett PRW, Campbell MJ, Sica REP. Electrophysiological estimation
of the number of motor units within a human muscle. JNeurolNeurosurg
Psychiatry. 1971; 34:121-131.

Milner-Brown HS. Stein, RB, and Yemm R. Changes in firing rate of human motor units
during linearly changing voluntary contractions. JPhysiol. 1973;230:371-390.

Morales FR, Engelhardt JK, Soja PJ, Pereda AE, and Chase MH. Motoneuron properties
during motor inhibition produced by microinjection of carbachol into the pontine
reticular formation of the decerebrate cat. JNeurophysiol. 1987;57:1118-1129.






48


Moritani T, Muro M, and Nagata A. Intramuscular and surface electromyogram changes
during muscle fatigue. JAppl Physiol. 1986;60:1179-1185.

Mudie MH, Matyas TA. Can simultaneous bilateral movement involve the undamaged
hemisphere in reconstruction of neural networks damaged by stroke? Disabil
Rehabil. 2000;22:23-37

Myers LJ, Erim Z, and Lowery MM. Time and frequency domain methods for
quantifying common modulation of motor unit firing patterns. JNeuroengineering
Rehabil. 2004; 1:2.

Nudo RJ. Recovery after damage to motor cortical areas. Curr Opin Neurobiol.
1999;9:740-7.

Olson CB, Carpenter DO, and Henneman E. Orderly recruitment of muscle action
potentials. Arch Neurol. 1968;19:591-597.

Ouellette MM, LeBrasseur NK, Bean JF, Phillips E, Stein J, Frontera WR, Fielding
RA.Petajan JH. Stroke. 2004;35(6):1404-9.

Petaj an, JH. AAEM minimonograph #3: motor unit recruitment. Muscle Nerve.
1991;14(6):489-502.

Priez A, Duchene J, Goubel F. Duchenne muscular dystrophy quantification: a
multivariate analysis of surface EMG. MedBiolEng Comput. 1992;30:283-291.

Popovic D, Sinkjaer T. Control of Movement in the Physically Disabled. London, UK:
Springer-Verlag; 2001

Rose J, McGill KC. The motor unit in spastic cerebral palsy. Develop Med Child Neurol,
1998;40:270-277.

Rose DK, Winstein CJ. Bimanual training after stroke: are two hands better than one?
Top Stroke Rehabil. 2004; 11;20-30.

Salenius S, Portin K, Kajola M, Salmelin R and Hari R. Cortical control of human
motoneuron firing during isometric contraction. JNeurophysiol. 1997;77:3401-
3405.

Selye H. Stress in health and disease. Reading, MA: Butterworth. 1976.

Toulouse P, Carrault G, Le Rumeur E, Coatrieux JL: Surface electromyogram automatic
analysis and Guillain-Barr, syndrome follow up. Electromyogr Clin Neurophysiol.
1992;32(1-2):51-62.

Van Dieen JH, Futoshi 0, De Haan A. Reduced neural drive in bilateral exertions: A
performance-limiting factor. Med Sci Sports Exerc. 2003;35(1): 111-118.






49


Wakeling JM and Rozitis AI. Spectral properties of myoelectric signals from different
motor units in the leg extensor muscles. JExp Biol. 2004;207:2519-2528.

Walter CB, Swinnen JP. Kinetic attraction during bimanual coordination. JMot Behav.
1990;22(4):451-73.

Whitall J, Waller S, Silver K, Macko R. Repetitive bilateral arm training with rhythmic
auditory cueing improves motor function in chronic hemiparetic stroke. Stroke.
2000;31:2390-2395.

William M. Landau, MD; Shirley A. Sahrmann, PhD. Preservation of directly stimulated
muscle strength in hemiplegia due to stroke. Arch Neurol. 2002;59:1453-1457.

Wyke, M. The effects of brain lesions on the performance of bilateral arm movements.
Neuropsychologia. 1971;9:33-42.















BIOGRAPHICAL SKETCH

Lindsay McManus was born in North Carolina in 1982. She graduated with honors

from North Carolina State University with a Bachelor of Science degree in biomedical

engineering and biological engineering. During a summer internship in Cincinnati,

Lindsay joined a research team studying stroke and spinal cord injury recovery, which

eventually led her to the University of Florida in 2004. Lindsay has worked with Dr.

James Cauraugh in the Motor Behavior Laboratory investigating chronic stroke recovery

through bilateral training and EMG-neuromuscular stimulation. While at the University

of Florida, Lindsay also worked in the Biomechanics Laboratory with Dr. John Chow and

Dr. Mark Tillman. In addition, she was also an instructor for the undergraduate motor

learning/control course for four semesters. After graduating with a Master of Science

degree in August 2006, Lindsay plans to enter the field of medical devices and implants.