Citation
Neuromuscular neutral zones response to static lumbar flexion

Material Information

Title:
Neuromuscular neutral zones response to static lumbar flexion
Creator:
Youssef, Jimmy E
Publication Date:
Language:
English
Physical Description:
xii, 67 leaves : ; 28 cm

Subjects

Subjects / Keywords:
Dead loads (Mechanics) ( lcsh )
Lumbar vertebrae ( lcsh )
Spine ( lcsh )
Dead loads (Mechanics) ( fast )
Lumbar vertebrae ( fast )
Spine ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 62-67).
Statement of Responsibility:
by Jimmy E. Youssef.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
268674717 ( OCLC )
ocn268674717
Classification:
LD1193.E54 2008m Y68 ( lcc )

Downloads

This item has the following downloads:


Full Text
NEUROMUSCULAR NEUTRAL ZONES RESPONSE TO STATIC LUMBAR
FLEXION
By
Jimmy E. Youssef B.S., University of Balamand, 2005
A thesis submitted to the University of Colorado Denver In partial fulfillment Of the requirements for the degree of Master of Science in Electrical Engineering
2008


This thesis for the Master of Science
Degree by Jimmy E. Youssef Has been approved By
Date


Youssef, Jimmy E. (M.S., Electrical Engineering) Neuromuscular neutral zones response to static lumbar flexion Thesis directed by Associate Professor Miloje Radenkovic
ABSTRACT
The objective of this thesis was to study the effect of prolonged static lumbar flexion at moderate load on the spines stability. Eight preparations of in vivo feline models were subjected to 40N static loading in a series of 6 periods of 10 minutes of work spaced by 10 minutes of rest, followed by a seven hours rest period. Reflexive multifidi electromyogram (EMG) initiation threshold in the stretch phase, EMG cessation threshold in the relaxation phase along with their corresponding displacement, and tension thresholds that trigger that reflexive muscular activity was recorded. A significant increase in the Displacement Neuromuscular Neutral Zones (NNZs) was observed after the static loading period followed by an exponential decrease to its normal value over 7 hours. Similarly, the Tension NNZ showed an increase followed by a decrease below baseline after 2 to 3 hours of recovery. A decrease in the peak MAV occurred and was then followed by an increase that exceeded the baseline. No variability in the EMG median frequency was detected throughout the recovery period. The results suggest that laxity in the ligaments and decreased reflexive muscular activity in the first 2 hours of recovery following the static loading period, leaves the spine unprotected and under risk of injury. During the remaining recovery time, a compensatory muscle activity takes over, lending an increased protection to the spine, and resulting in limited motion and muscle stiffness. Workers exposed to static loading of their spine should protect it for the first 2 hours after work since spinal stability is compromised in this period.
This abstract accurately represents the content of the candidates thesis. I recommend its publication.
Miloje Radenkovic


DEDICATION
I dedicate this thesis to my parents and family who always encouraged me to pursue knowledge and have shown me love and support my whole life.
I also dedicate this to my cousin Joe who showed me what real courage and strength are, and even though he is not with me anymore and never had the chance to pursue his dreams; I share this with him and hell always be in my heart.


ACKNOWLEDGMENT
I thank God for always being there for me, keeping me going throughout the stressful times, and leading me to success.
I want to express my most sincere and utmost gratitude to my parents, Edmond and Fayrouz. Mom and Dad, thank you for never failing to believe in me. Thanks for all your sacrifices and hard work to give me the best education and the best future. You have always put us, your children, first which I will never forget. Thanks Mom for your constant prayers and I hope that I will always make you proud.
I also want to thank my family in the United States that gave me support, a place to stay, and love. They have been so generous to me and given me a home away from home. Thanks Tony, Kellie, Taylor, Jon, and Diana.
I want to express great thanks to my advisor Dr Moshe Solomonow who was a great professor and mentor throughout this study. Thanks for always looking out for me and guiding me by giving me valuable lessons not only in academia but also in life. You always treated me like a part of your family; I really appreciate all the opportunities and the help you gave me.
To Dr. Yun Lu and Dr. Bing-He Zhou, thank you for sharing your knowledge with me. I thank Dr. Lu for teaching me a lot about anatomy and machining. Thank you Dr. Zhou for teaching me valuable information about data acquisition and signal processing. I also would like to thank Dr. Bradley Davidson for introducing me to data modeling, statistics and presentation skills.
Finally, I would like to thank Dr. Miloje Radenkovic and Dr. Robert Grabbe for taking the time to review my thesis. Thanks Dr. Mike for teaching me a lot about signal processing and introducing me to Dr. Solomonow which led to this great learning experience.


TABLE OF CONTENTS
Figures...................................................................ix
Tables....................................................................xii
Chapter
1. Introduction.........................................................1
1.1 Significance of Study................................................3
2. Physiological Background.............................................5
2.1 Vertebral Column.....................................................5
2.2 Intervertebral Discs.................................................7
2.3 Ligaments............................................................8
2.4 Mechanoreceptors....................................................11
2.5 Multifidus Muscles..................................................12
2.6 Spinal Stabilizing Feedback Control System......................... 14
3. Electromyography (EMG)..............................................17
3.1 Motor Unit Recruitment and EMG Median Frequency.................... 18
3.2 Zero Padding........................................................19
3.3 Windowing...........................................................20
4. Methods and Procedure...............................................23
4.1 Preparation.........................................................23
VI


4.2 Instrumentation....................................................25
4.3 Experimental Protocol..............................................26
4.4 Data Analysis......................................................28
4.4.1 Displacement NNZs and Tension NNZs.................................28
4.4.2 Mean Creep.........................................................30
4.4.3 Normalized peak MAV................................................30
4.4.4 Median Frequency...................................................31
4.4.5 Mean of Percentage Change from Baseline..........................33
4.4.6 Statistics.........................................................33
4.5 Models.............................................................34
5. Results.............................................................37
5.1 Displacement NNZs..................................................40
5.2 Tension NNZs.......................................................43
5.3 Normalized peak MAV................................................46
5.4 Mean Creep.........................................................48
5.5 Median Frequency...................................................49
5.6 Models.............................................................50
6. Analysis............................................................53
6.1 Spine Stability......................................................53
vii


LIST OF FIGURES
Figure
2.1 Spinal vertebrae (NIH 2007)...........................................6
2.2 Schematic of anterior flexion and extension...........................7
2.3 Intervertebral Disc (NIH 2006)........................................8
2.4 Behavior of the Creep of viscoelastic tissues under constant load.....9
2.5 Ligaments of the lumbar spine........................................10
2.6 Multifidus muscles (Adapted from Grays Anatomy of the
Human Body 2000)....................................................13
2.7 Simplified feedback control system of the spine (adapted from
Solomonow et al, 2001)............................................. 15
3.1 Commonly used window functions.....................................21
4.1 Schematic of the experimental setup showing the external fixations
placed on LI and L7, and the hook applied to the L4/5 ligament during rest period (a), and during static loading period (b)...............24
4.2 Stainless steel wire electrodes......................................25
4.3 Actual picture of the preparation showing the external fixators, the
electrodes, and the S shaped hook inserted in the L4/5 supraspinous ligament............................................................26
4.4 Schematic of the protocol that was followed in the experiment
with N=6 repetitions................................................27
4.5 Plot of the hysteresis that shows the tension versus displacement curve of a single cycle at 0.25-Hz frequency and 40-N amplitude.
IX


The black circle shows EMG initiation at the stretch phase. The
Boldface on the curve designates the period where EMG was
present and recorded throughout the cycle. The white circle
represents EMG cessation in the release phase. The unBold
part of the curve represents the NNZs.................................29
5.1 EMG responses from channels L3/4, L4/5, and L5/6 along with the displacement and tension channels for a single preparation
when subjected the experimental protocol..............................38
5.2 EMG responses from channels L3/4, L4/5, and L5/6 along with the displacement and tension channels for a single cycle test
(0.25-Hz, 40-N).......................................................39
5.3 Plots of the tension versus displacement (hysteresis) for a single
preparation at the pre-static loading period, after 1 hour of recovery, after 2 hours of recovery, and after 6 hours of recovery...............40
5.4 Mean displacement neuromuscular neutral zones for
N=8 preparations.......................................................42
5.5 Mean of percentage change from baseline for displacement NNZs
(onset and offset). The star symbol represents a statistically significant change from baseline values............................................43
5.6 Mean tension neuromuscular neutral zones for N-8
preparations...........................................................44
5.7 Mean of percentage change from baseline for tension NNZs
(onset and offset). The star symbol represents a statistically significant change from baseline values............................................45
5.8 Mean normalized peak MAV for N=8 preparations..........................47
5.9 Mean of percentage change from baseline for normalized peak MAV.
The star symbol represents a statistically significant change from baseline value.........................................................48
x


5.10 Mean of the creep for N=8 preparations. The star symbol represents
a statistically significant change from baseline value.....................49
5.11 Mean of 3-point averaged median frequency for N=8 preparations.............50
xi


LIST OF TABLES
Tables
5.1 Displacement neuromuscular neutral zones model parameters......51
5.2 Tension neuromuscular neutral zones model parameters...........52
5.3 Normalized peak MAY model parameters...........................52
xii


1.
Introduction
Low back pain is a widespread problem, affecting roughly 80 percent of Americans at some point in their lives. It's considered the fifth most common reason for visiting a doctor, the most common cause of work-related disability in people under age 45, and one of the first reasons for missing work [1,2], Low back injuries mostly occur in the workplace, especially where workers are exposed to intense physical activities as part of their daily occupational routine [3]. In such cases, the injury usually occurs after the work is completed while performing simple actions. And the cost for this problem is estimated to be a staggering $50 Billion yearly.
Previous reports showed that workers subjected to occupational activities requiring sessions of cyclic or prolonged static lumbar flexion (such as mechanics, farm workers, concrete and roofing workers, etc.), report with a high rate of low back disorders [2, 4-7] which is considered one of the most costly areas of general musculoskeletal disorders based on medical expenses, disability payments, and lost wages. Previous literature also proved that exposing the lumbar spine to high magnitude static loading periods or to moderate and high magnitude cyclic loading periods elicited creep of the viscoelastic components which led to a cumulative trauma disorder (CTD) resulting in pain, muscular


weakness, and stiffness [8-13]. High load magnitudes [11, 14], longer loading durations [15], high number of repetitions for a given period [16], and shorter periods of rest in between [17] are high risk factors for static and cyclic flexion. It has also been established that cyclic loading is more challenging than static loading on the viscoelastic tissues [11],
A significant part of low back pain problem is known to be of electromechanical origin and often referred to as spinal instability. Spinal instability could be explained as an excessive motion between vertebrae with high risk of injury caused by elicited creep in the ligaments. The passive viscoelastic structures (e.g. ligaments, discs, and capsules) have a minor role in maintaining the stability of the spine [18-22], whereas the major role goes to active forces generated from reflexive muscular contractions [23-25]. It was proven that electrical or mechanical stimulation to the lumbar viscoelastic tissues triggered a muscle activity in the multifidus muscles confirming the presence of a ligamento-muscular reflex [26, 27], It has also been established that the spine is relatively compliant for small perturbations about the neutral position [28]. These small displacement ranges within which stability is not disturbed, and where muscular contraction wasnt required were designated as Neuromuscular Neutral Zones
2


(NNZ) [29]. During these neutral zones, no muscular activity (EMG) will be observed, and the ligaments are stiff enough to support the spine.
Previous studies showed that the neuromuscular neutral zones become narrower with the increase of the frequency of flexion [30, 31] in order to maintain spine stability, and that they have a pattern of change across the lumbar spine [29]. It also showed the effect of static and cyclic lumbar loading on low back pain. What hasn't been investigated yet is the effect of prolonged static lumbar loading on the neuromuscular neutral zones, thus on spine stability. This project will explore the behavior of neuromuscular neutral zones in response to moderate static lumbar loading of 40N on in order to simulate a labor work environment. We hypothesize that a prolonged moderate static load over a moderate duration will result in increased neuromuscular neutral zones in the first few hours of rest due to the laxity in the viscoelastic tissues leaving the spine unprotected and under risk of injury.
1.1 Significance of Study
The results from this study may provide experimental and biomechanical information about spine instability and may present guidelines for designing safe work schedules to prevent spinal injury. This will require further research to
3


extrapolate obtained data from the feline model to a human model. It will also provide a new insight into the ligamento-muscular feedback loop that maintains the spine stability.
4


2. Physiological Background
It's very important to understand the anatomy and physiology of the lower back and its stabilizing system in order to understand the effect of static loading on the spine's stability. This section describes the anatomy of the lower back specifically the lumbar spine area to help understand its structure and function including some of the elements such as vertebral column, intervertebral joints, ligaments, multifidus muscles, and the stabilizing feedback control system of the spine. It also briefly explains motor unit recruitment and its relationship with muscle contraction.
2.1 Vertebral Column
The vertebral column, also called the spine or backbone, extends from the base of the skull to the tip of the coccyx. Its composed of 33 vertebrae; 7 cervical, 12 thoracic, 5 lumbar, 5 sacral (fused together) and 4 coccygeal (fused together) (Figure 2.1). The spine is flexible, yet extremely tough and serves to support the back through a full range of motion (anterior and lateral flexion, and extension, and rotation). Anterior Flexion is a forward movement (anterior bending) whereas extension is a backward movement (posterior bending), and rotation is a twisting of the vertebral column (Figure 2.2).
5


7 Cervical vertebrae
12 Thoracic vertebrae
5 Lumbar vertebrae
Figure 2.1 Spinal vertebrae (NIH 2007)
The vertebral column protects the spinal cord, which runs from the brain through the canal in the middle of the vertebral column (vertebral foramen), and also serves as a stable point of attachment for the ribs, the pelvic girdle, and the muscles of the trunk [32], The area of focus is the lumbar section from lumbar level 1 to lumbar level 5 (L1-L5) where the vertebrae are the largest segments and the strongest. The lumbar spine is most frequently involved in back pain because it withstands most of the pressure from body weight and is subjected to the largest forces and stresses along the spine.
6


Figure 2.2 Schematic of anterior flexion and extension 2.2 Intervertebral Discs
Intervertebral discs are located in between every two adjacent vertebrae and form one fourth of the spine's length. They consist of an outer ring made of collagen fibers and referred to as anulus fibrosus, and an inner elastic gelatinous material that contains fluid and is known as nucleus pulposus (Figure 2.3). The structure of the discs allows them to act as cushions, change shape while permitting various movements of the vertebral column, but still resist excessive motion in order to form a strong joint. The disc is viscoelastic and develops creep
7


under load applied over time. The intervertebral discs also serve as shock absorbers [33], and loose fluid content and height when injured [9],
Figure 2.3 Intervertebral Disc (NIH 2006)
2.3 Ligaments
A ligament is a short band or sheet of tough fibrous dense tissue composed mainly of long tightly packed and elastic collagen fibers. Ligaments are connective tissue so their major functions are to connect bones to each other, guide joint movements, and provide stability to the joint [34, 35],
Ligaments follow the properties of viscoelastic tissues such as creep, tension-relaxation, hysteresis, arid frequency dependence [36]. They adjust their length according to the position of the vertebral column during flexion and
Spinous
process
8


extension of the lumbar spine. They also gradually lengthen up to a finite maximum when under constant tension, and then recover exponentially after cessation of the tension and this is referred to as Creep (Figure 2.4).
Figure 2.4 Behavior of the Creep of viscoelastic tissues under constant load
Creep is described as the percentage of the elongation divided by the initial displacement as shown in the following equation.
Creep =----------------------*100% (2.1)
Initial displacement
Ligaments are considered as passive tissue, thus they do not contract voluntarily to generate a force. Instead, they become stiffer with increasing applied loads to prevent excessive displacement of the bones and damage to the joint [36]. The tension in a ligament decreases with time when under a constant
9


elongation, and this behavior is referred to as tension-relaxation. Ligaments also display a hysteresis loop, where the tension-displacement curve is different during stretch (loading) and relaxation (unloading) phases resulting in net internal energy loss [34],
The spine has numerous ligaments that connect individual vertebrae such as the anterior longitudinal ligament, the posterior longitudinal ligament, transverse ligament, ligamentum flavum, facet capsulary ligament, interspinous ligament, and supraspinous ligament. The focus of study will be on the supraspinous ligament since its the first ligament to fall under tension and elongation during forward flexion [37]. The supraspinous ligament runs between the tips of the spinous processes of adjacent vertebrae (Figure 2.5).
A'lW-v-ftr
Figure 2.5 Ligaments of the lumbar spine
10


2.4 Mechanoreceptors
The human body contains five kinds of receptors: thermoreceptors, nociceptors, photoreceptors, chemoreceptors, and mechanoreceptors [32]. Thermoreceptors are responsible for detecting changes in temperature, nociceptors detect pain, photoreceptors detect light in retina of the eye, chemoreceptors respond to chemicals in the mouth and smell in the nose, and mechanoreceptors respond to mechanical pressure or stretching by firing nerve impulses (action potentials) that travel though neural pathways to the central nervous system and then result in a muscle activity, which is known as ligamento-muscular reflex.
There are five main types of mechanoreceptors: Meissners corpuscles, Golgi organs, Merkels discs, Ruffini corpuscles, and Pacinian corpuscles. The Meissners corpuscles are rapidly adapting receptors for fine touch and detecting changes in texture, whereas Merkels discs are slowly adapting receptors and detect sustained touch and pressure. Pacinian corpuscles are fast adapting receptors with low threshold, and response to deep pressure and rapid variations, and Ruffinis corpuscles detect sustained pressure [32], The Golgi organs respond to extreme force ranges [38].


Ligaments are known to contain mechanoreceptors, thus they are considered as sensory organs and they relay information to the nervous system. The supraspinous ligament was shown to contain 2 types of mechanoreceptors: Type II and III [39], Type II receptors are mostly fast adapting; they signal the initiation and cessation of stimuli to the central nervous system, and provide continuous signaling for changes in tensile forces. Whereas type III mechanoreceptors are slow adapting and signal only during extreme changes in motion and load.
2.5 Multifidus Muscles
Paraspinal muscles are the posterior muscles immediately next to the vertebral column. They allow mobility and stability to the spine. Our interest in this study is specifically the multifidus muscles that fill up the groove on both sides (left and right) of the spinous processes of the vertebrae as shown in (Figure 2.6). They usually span 2 to 4 joint segments and connect the mamillary process of one vertebra to the spinous process of the second, third, and fourth vertebra above [40], This allows them to exert compressive forces between vertebrae with higher role in extension.
12


The multifidi are the deepest muscles in the back and the closest to the vertebral column. They have a large cross-sectional area and short fiber length, thus their main role is to limit excessive motion across individual motion segments in order to provide stability in the lumbar spine region rather than mobility [41]. Multifidus muscles are also known to be the major intervertebral stabilizing muscles of the spine.
Figure 2.6 Multifidus muscles (Adapted from Grays Anatomy of the Human Body 2000)
13


2.6 Spinal Stabilizing Feedback Control System
The spinal stabilizing control system consists of three subsystems [42].
The passive subsystem consists of the vertebrae and the viscoelastic tissues (ligaments, discs, facet capsules, etc.). The active subsystem consists of the active muscles surrounding the spine. The neural subsystem includes the nerves and the Central Nervous System (CNS) and monitors the transducer feedback signals from the passive subsystem for the relay of loading, displacement, and proprioception [42] and reacts to it by directing the active subsystem to maintain the needed spinal stability.
Spine stability is clinically defined as the ability of the spine to bear loads, allow movement, and still avoid injury and pain [43]. This definition implies that a dynamically stable spine should not have excessive intevertebral motion in response to a perturbation and should return rapidly to its undisturbed position. The term stability is used in this study for its clinical meaning that is often confused with the mechanical term robustness. For clarity, a system that is clinically less stable refers to a system that is mechanically less robust towards becoming unstable. As stated earlier, the spine stability is maintained partly by the passive viscoelastic tissues that generate tension expressed as resistance to destabilizing motion and mainly by the active forces generated from the
14


musculature through a reflexive system that involves sensorimotor feedback. This gives rise to the feedback control system of the spine shown in figure 2.7.
force length Velocity Acceleration etc..
Figure 2.7 Simplified feedback control system of the spine (adapted from Solomonow et al, 2001)
This figure depicts the major anatomic structures associated with the forward and feedback loops of the spine stabilizing system [30], In the forward loop, the central nervous system (controller) activates the motorunits dynamics, the musculoskeletal dynamics, and the viscoelastic dynamics to regulate the force, length, velocity, acceleration, etc. The feedback loop also called ligamento-muscular reflex is the focus of our study and starts with a perturbation in the spine, then the mechanoreceptors sense a deformation in the viscoelastic tissues
15


which leads to an afferent neuron. The afferent neuron signal travels to the spinal cord, passes through intemeurons, and finally reaches the multifidi and generates a muscle contraction thus stabilizing the spine.
Its important to recall that a feline model is being used in this research. Even though differences between the feline model and humans are evident, scientific studies have showed that feline models are appropriate for this type of investigations as will be discussed later in the report. Similar behavior of the viscoelastic tissues and the stabilizing system of the spine is expected.
16


3. Electromyography (EMG)
Electromyography (EMG) is the measurement and recording of electrical muscle activity. Recordings are usually made at rest and during contraction. A skeletal muscle fiber has a membrane resting potential of -70 mV during rest. Muscles contract when stimulated by action potentials traveling through motor neurons along muscle fibers. An action potential is a pulse-like wave that triggers a muscle activity. Action potentials have the same magnitude and are generated when a threshold of usually -55 mV membrane potential is reached.
The spatio-temporal summation of all the action potentials from all the active motor units (MUAPs) constitutes the electromyogram of the muscle [44] and forms its random shape. The electromyogram varies in frequency and amplitude depending on the intensity of the muscle contraction.
There are three kinds of electrodes, surface electrodes, needle electrodes, and wire electrodes. In this study, bipolar fine wire electrodes were inserted in deep muscles as they are capable of picking up EMG signals at rest and during contraction, only from the multifidus muscles and without crosstalk interference from superficial muscles. Wire electrodes were used for their increased flexibility compared with needle electrodes and for their small impedance compared with surface electrodes.
17


3.1 Motor Unit Recruitment and EMG Median Frequency
A motor unit is the connection between an efferent neuron coming from the spinal cord and a series of muscle fibers. Fewer muscle fibers in a motor unit results in a fine motor control, whereas more muscle fibers in a motor unit results in a stronger muscle contraction.
Muscle force and strength are regulated by many processes such as motor unit recruitment and increase of action potential firing rate. Motor unit recruitment is the process of selectively activating motor units. Motor units usually get recruited in an orderly pattern from smaller motor units progressively to larger motor units with the increase in the muscle force necessity. Once all motor units are fully active, maximal force is reached by the increase of the firing rate. The smaller motor units have a small diameter and are resistant to fatigue, but generate small force increments, whereas larger motor units have a larger diameter (larger conduction velocity) and fatigue fast, but generate large force increments. Our study of interest is the multifidus muscles that are considered slow twitch muscles and consist of mainly smaller motor units.
Median frequency (MF) is the frequency that divides the power spectra of the EMG signal in half. Previous studies have shown that the increase in the motor unit recruitment results in an increase in the average conduction velocity
18


(CV), and also causes a linear increase in the median frequency [45], showing that the average conduction velocity is linearly proportional to the MF [46]. This proves that an increase in the motor unit recruitment indicates an increase in the MF. Thus, the MF can be an indicator for motor unit recruitment and behavior of the muscles due to different loading conditions.
After sustaining long work periods, the larger and faster motor units start to fatigue, thus resulting in a decrease in the MF and an increase of EMG amplitude. The decrease in the MF indicates the deactivation of larger motor units in a process known as orderly derecruitment, and the increase of the EMG amplitude indicates the increase of the firing rate of the still active smaller motor units. In the absence of muscle fatigue, the EMG amplitude and MF are related, where a higher requirement of muscle force will be satisfied by an increase of the EMG amplitude and an increase in the MF. In this study, muscle fatigue is not expected to occur, and EMG peak Mean Absolute Value is expected to be related to the EMG median frequency.
3.2 Zero Padding
Zero padding consists of extending a dataset with zeros to a power-of-two length. It is often applied as a step in the signal processing associated with the
19


spectral analysis. Some FFT algorithms require zero padding in order to run faster and more efficiently. Zero padding also enables the power spectrum to be evaluated over a fuller range of frequencies. However, it can't increase resolution, but it only shows a better display by providing a high density spectrum [47].
3.3 Windowing
A window function is a function that has a zero value out of a specific interval. Window functions are usually multiplied by data signals and result in their product inside the specific interval, and in a zero-valued function outside the interval. For instance, a rectangular function has a constant value inside the interval and a zero value outside the interval. It can be an arbitrary finite-duration window and can be of unit amplitude to preserve the original signal's amplitude. But an issue arises for the rectangular window since it introduces discontinuities at the end of the data signal and it initiates side lobes into its Fourier transform because of the fast switching rectangular function. Then, some different window functions were introduced and were able to smoothly attenuate the signals amplitude to zero with time, which diminishes the power of the side lobes in the frequency domain (Hanning window). Each window function has its advantages and disadvantages and each is more effective for different types of signals. Some
20


windows improve the frequency resolution but attenuate the original amplitude of the signal, and some windows save the original signal's amplitude but introduce side lobes to its Fourier transform [48], Some of the most popular windows are Hamming, Barlett, Hanning, Tukey, and Blackman window (Figure 3.1).
"me domain
Figure 3.1 Commonly used window functions
The difference between all the previously mentioned windows is the ratio of taper to constant sections (r). The Tukey window was selected in this research since it doesn't attenuate the amplitude of the whole signal because of its unit
21


amplitude unlike the Hanning window, and since it doesn't introduce really high frequency components because of the presence of the ramp unlike the rectangular window. The Tukey window has a ratio of taper to constants sections r = 0.25, and is described by following equation (equation 3.1).
Mr(n) =
1 (2n n-\
1 +COS n
2 [r N-1 J
, n < ^(N -1)+1 -(N-\)+\ 1 + cos
2n 2n n-\
\ r
r N-1
71
N--(N-\) (3.1)
22


4.
Methods and Procedure
4.1 Preparation
Eight adult cats (3.347 0.51 kg) were anesthetized with alpha chloralose (60mg/kg), and used in a protocol approved by the Institutional Animal Care and Use Committee (IACUC). Alpha chloralose does not inhibit the reflexive muscle activity. The skin over the spine was dissected from the thoracic level to the sacral level, and then retracted to expose the intact dorsolumbar fascia. An S-shaped stainless steel hook was inserted around the middle of the supraspinous ligament between L4 and L5. The preparation was then placed prone in a rigid stainless steel frame with external fixations on the posterior processes of LI and L7, which allowed the isolation of the lumbar level from thoracic and sacral levels, and avoided mechanical interaction. The frames position was then adjusted to achieve a vertical alignment of the S shaped hook with the actuator of a Bionix 858 Material Testing System. A gauze pad was soaked with sterile water and placed over the incision area to prevent the exposed skin from drying. Then,
Saline fluid was regulated by a Flo-Gard Volumetric Infusion Pump to release lOml/kg/hour through an Intravenous (IV) line to prevent the tissue from dehydrating. The Arterial oxygen saturation (SPCE) and the pulse rate were
23


monitored by a Rad-5 Pulse Oximeter through a non invasive sensor placed on the preparations tongue. Ventilation was given by a Harvard Inspira Advanced Safety Ventilator. The temperature was maintained by a Gaymar heating pad and monitored by a WellchAllyn thermometer. Finally, records of the temperature, SPO2, and the pulse rate were taken every 15 minutes throughout the experiment. A schematic of the experiment is shown in figure 4.1.
Load
Load
Figure 4.1 Schematic of the experimental setup showing the external fixations placed on LI and L7, and the hook applied to the L4/5 ligament during rest period (a), and during static loading period (b).
24


4.2 Instrumentation
Six pairs of stainless steel wire EMG electrodes (Figure 4.2) were inserted 6-8 mm to the right of the midline into the multifidus muscles of Ll/2, L2/3,
L3/4, L4/5, L5/6, and L6/7. A ground electrode was inserted in the gluteus muscle. Each electrode pair constituted the input to a differential EMG amplifier with a 110-dB common mode rejection ratio, gain capability of up to 200,000, and a band pass filter in the range of 20-500 Hz. The EMG response from all six channels was monitored on oscilloscopes, recorded and stored into a computer at a sampling rate of 1000 Hz. The signal sampling rate was chosen based on the Nyquist theorem (e.g. 2 x highest frequency of the filter) to avoid aliasing. The S-shaped stainless steel hook was then connected to the actuator of the Bionix 858 Material Testing System (MTS Inc., St. Paul, MN). The load was applied by the vertical actuator via a computer controlled loading system. Vertical displacement of the actuator and the load cell output incorporated in it were also sampled into the computer all along with the EMG data. An actual picture of the experiment is shown in figure 4.3
25


Figure 4.2 Stainless steel wire electrodes
Figure 4.3 Actual picture of the preparation showing the external fixators, the electrodes, and the S shaped hook inserted in the L4/5 supraspinous ligament.
4.3 Experimental Protocol
Initially, three sinusoidal cycles at a frequency of 0.25 Hz and 40-N of vertical load were applied to the lumbar spine through the actuator of the Bionix 858, each followed by a 10 minutes rest period. The 10 minutes rest periods were included to allow recovery of any tension-relaxation that may have developed in the viscoelastic structures [10, 49, 50]. The 40N load is considered to be a
26


moderate load that would elicit a moderate lumbar flexion [26]. Then, the preparation was subjected to a static load of 40-N for 10 minutes. Next, the load was released for a 10 minutes rest period. This last process was applied 6 times for a total of 2 hours (6X 10:10). The six working periods were followed by a 7 hours recovery period (rest). During this recovery period, single 0.25 Hz (4-second) sinusoidal cycle tests at 40-N were applied to assess vertical displacement, Tension, and EMG. Tests were applied after 10, 20, and 30 minutes of rest, and every hour thereafter (Figure 4.4).
Pre-static Static loading period 7 hour recovery period
loading
f = 0.25 Hz
P1.P2.P3 Pre-static loading (0.25 Hz, 40 N)
L1,L2,...,Ln Static load (10 min, 40 N)
Ri,R2, Rn-, Rest (10 min, no load)
N Number of static load repetitions
Ti.T2.-.T9 Single test cycle during recovery period (0.25 Hz, 40 N)
Figure 4.4 Schematic of the protocol that was followed in the experiment with N=6 repetitions.
27


4.4 Data Analysis
EMG from all seven lumbar levels, displacement, and tension data were recorded and stored as sixteen seconds intervals in the computer. Then, a high pass filter with a cut-off frequency of 20 Hz was applied to the EMG data to prevent any movement artifact that is usually a brief electrode shift caused by muscle contractions. Next, the displacement and tension data were subsequently filtered through a low pass filter with a cut-off frequency of 5 Hz.
4.4.1 Displacement NNZs and Tension NNZs
In each of the single cycle tests, the first 500 msec of each EMG record which was prior to the initiation of the tension were used as a point of reference for baseline signal noise. After this period, the first point along the signal to exceed five times the baseline level was denoted as the EMG initiation threshold in this channel. The corresponding tension and displacement values during the stretch phase of this cycle were designated as Tension and Displacement Neuromuscular Neutral Zones (Onset). Figure 4.5 is a plot of the hysteresis that shows the tension versus displacement curve and where EMG was present on the curve in boldface. Similarly, the last EMG value to exceed five times the baseline noise was denoted as the EMG cessation threshold in this channel, and its
28


6. Analysis
The major finding of this research consisted of the fact that a moderate static load of the lumbar spine causes laxity and loss of reflexive stabilizing forces from the multifidus muscles resulting in a large risk of spine injury in the first two hours of rest after a sequence of six sessions of 10 min of 1:1 work to rest ratios (for a total of 2-h). This risk ends during the fourth hour of rest when the ligamento-muscular reflex provides the lumbar spine with a compensatory muscle activity that takes over stability function by increasing the muscular activity via a feedback loop. This mechanism serves well to preserve lumbar stability and compensates for the laxity developing in the lumbar viscoelastic tissues.
6.1 Spine Stability
Ligaments have a very small contribution toward maintaining the spines stability [18-21], whereas the major part is achieved by active multifidus muscle forces generated by the ligamento-muscular reflexes [26, 27, 33, 52, 53]. This reflex arc is triggered by load applications to the supraspinous ligament which excites its sensory mechanoreceptors and then activates the multifidus muscles at the level of deformation as well as one level below or above [26], Previous studies showed that prolonged flexion of the lumbar spine results in tension-
53


relaxation and laxity of its passive viscoelastic structures [13, 54], The laxity and creep induced in the viscoelastic tissues of the spine desensitize the mechanoreceptors resulting in a diminished reflexive muscular activity and allowing full exposure to instability and injury [13]. It was also shown that the muscles seem to compensate for the loss of tension in the lumbar viscoelastic tissues by an increased EMG activity later in the recovery period [12, 27, 50. 55].
6.2 Motor Unit Recruitment
A stronger reflexive muscle activity can be established by either increased firing rate by the same active motor units, or by motor unit recruitment (e.g. activating larger motor units) at the same firing rate, or by both. Increase in the conduction velocity leads to an increase in the median frequency, indicating newly recruited, large motor unit. It has been proven in previous studies that the mean power frequency of the EMG spectrum and the muscle fiber conduction velocity are linearly related [46, 56], It also has been shown that there is a linear relationship between the average conduction velocity and the median frequency [46]. Then, Previous studies demonstrated that the frequency component of the EMG power spectrum can be used as an indicator of motor unit recruitment [57], Summing up all these facts, it was shown that an increase in the median frequency
54


indicates an orderly recruitment of the motor units [10, 45]. And that means that an increase in the average conduction velocity indicates activation of larger and faster motor units, and therefore an orderly recruitment of motor units that leads to an increase in the median frequency. In this study, no change in the median frequency was observed; therefore no significant change in the motor unit recruitment occurred. It can be concluded, therefore, that the increase in the EMG amplitude is due to increase in the firing rate of the already active motor units.
6.3 Creep
Previous studies stated that the viscoelastic creep loading increases the laxity of the intervertebral joint resulting in the injury to the spine and low pack disorder [8-10, 33], In this study, the loading period developed laxity in the ligaments and creep in the lumbar spine structure. This resulted in the combination of enlarged displacement neuromuscular neutral zones, enlarged tension neuromuscular neutral zones, and decreased peak MAV eliciting a delayed reflexive muscle activity, leaving the spine unprotected. Therefore, in the first 1 to 2 hours of recovery, low muscular activity was present and was triggered only after reaching high displacement and tension values. Even after the 7 hours of recovery, significant creep in the passive viscoelastic structures was still
55


present. The recovery period was not sufficient enough for full recovery of the creep in the viscoelastic tissues, unlike what was demonstrated in the latest studies that involved a similar load/rest schedule but with different load magnitudes or with cyclic loading. It was shown by a previous report that 40-N static loading did not result in delayed hyperexcitability, whereas only high static loads (60-N) displayed delayed hyperexcitability and an inflammatory response [14]. Also, it was proven that moderate (40-N) to high (60-N) magnitudes of cyclic loading could trigger a delayed hyperexcitability [11], Even though cyclic loading seems to be more challenging to the viscoelastic tissues than static loading [11], a recent study showed that the creep disappeared at the end of the recovery period for moderate cyclic loading [58] whereas it was still present for static loading in this study. That can be explained in a way that moderate 40-N cyclic loading was riskier and more damaging to the viscoelastic tissues which caused increased muscular activity and stiffness in the spine that prevented creep from rising in order to avoid further damage to the spine. Comparing with this study, the moderate 40-N static loading wasn't intense or damaging enough and didnt require increased muscular stiffness in the lumbar spine.
56


6.4 Compensation Mechanism
In this study, the presence of the compensation mechanism was sufficiently indicated by decreased Tension neuromuscular neutral zones and increased EMG peak MAV. This confirms that the muscular activity was triggered earlier at low tension values and with higher EMG amplitude to compensate for the laxity in the ligaments in order to prevent damages in the spine. This compensatory muscle reaction was achieved by increased firing rate of active motor units with no significant change in the median frequency; e.g. no change in the motor unit recruitment. Static loading with 40-N magnitude was shown to be a moderate loading that didn't cause any damage to the spine and didnt display any hypexcitability [14]. Thus, it was probably an increase in the firing rate of the multifidus muscles that generated higher EMG amplitude through temporal and spatial summation, and that activation of larger motor units wasn't necessary for the prevailing conditions. Comparing with a recent report involving 40-N cyclic loading that displayed an inflammatory response resulting in a hyperexcitability, the compensatory mechanism was accomplished by higher EMG amplitude and by an increased median frequency indicating activation of larger and faster motor units [58], This compensatory mechanism presented by an increased electromyogram peak MAV might be a new motor control mechanism
57


that wasnt exposed or discussed before and that separates itself from the actual hyperexcitability phenomenon experienced in previous reports that is usually defined as an inflammatory response to the micro-damage of the viscoelastic tissues.
6.5 Humans Vs Cats
It's important to recall that the data obtained in the experiments was derived from the feline model (quadruped). One needs to assure that the same processes occur in humans (biped) to confirm the validity of the final results.
A recent study showed biomechanical similarity of the spine between quadruped and bipeds confirming that the quadruped can be a valuable animal model for spine research [59], It was also shown that the neuromuscular system of the feline and the human are similar [60]. The collagen structures of the ligaments, facet capsules, and discs for humans are very similar to cats and also contain mechanoreceptors [61,62], Previous studies also proved that the ligamento-muscular reflex exists also in humans, and provided support to the similarity of the responses between humans and felines [12, 26, 63-66],
Even though the difference in size between cats and humans is evident, the general pattern of responses is expected to be the same but with differences in the
58


loading periods, resting periods, number of repetitions, time constants in the mathematical models, and loading magnitude; where the loading magnitude would be addressed to as low, moderate, and high loads for humans.
59


7.
Conclusion
This study provided experimental evidence showing the effect of a total of 2 hours moderate static loading and rest with a work to rest ratio of 1:1 on the neuromuscular neutral zones revealing more information about spine instability in a work environment. It also displayed a biomechanical and physiological pattern for the normalized peak MAV and the median frequency in order to track the reflexive multifidus muscles activity. Some scaling of the data derived from feline quadruped still needs to be done in order to apply the results to humans.
As a summary, the combination of information gathered from this investigation leads to the following conclusions:
1. Moderate static loads applied to the lumbar spine results in corruption to the ligamento-muscular reflex due to creep in the lumbar spine, and shows enlarged neuromuscular neutral zones placing the spine under full exposure to instability and injury in the first couple of hours of rest.
2. A Compensation mechanism takes over in the later hours of rest resulting in narrow neuromuscular neutral zones, muscle stiffness, and increased muscle activity to compensate for the laxity in the ligaments and prevent spine injury.
60


3. The compensatory muscle activity is probably attained by a higher EMG firing rate that results in higher EMG amplitude, but with no variations in the motor unit recruitment.
4. The compensatory mechanism is probably a new motor control mechanism that hasn't been investigated yet.
5. Seven hours of rest isn't enough for creep recovery in the viscoelastic tissues of the spine.
So we can conclude that workers should be watchful in the first couple of hours of rest after a long day of work periods to prevent spine injury, since the spine will be susceptible to instability. Finally, an extension of this research would be to further investigate the neuromuscular neutral zones behavior for different load magnitudes and for cyclic loading, and to identify the newly mentioned compensatory mechanism that occurs in the recovery period.
61


References
1. NINDS (2003) Low Back Pain Fact Sheet. Volume,
2. NIOSH, National Institute for Occupational Safety and Health: A Critical Review of Epidemiologic Evidence for Work-Related Musculoskeletal Disorders of the Neck, Upper Extremity, and Low Back. 1997, Musculoskeletal Disorders and Workplace Factors.
3. NRC, National Research Council: Musculoskeletal Disorders and the Workplace: Low Back and Upper Extremities. 2001: Institue of Medicine.
4. McGill, S.M., The biomechanics of low back injury: implications on current practice in industry and the clinic. J Biomech, 1997. 30(5): p. 465-75.
5. Andersson, G.B., Epidemiologic aspects on low-back pain in industry.
Spine, 1981.6(1): p. 53-60.
6. Damkot, D.K., et al., The relationship between work history, work environment and low-back pain in men. Spine, 1984. 9(4): p. 395-9.
7. Anderson, J.A., E.O. Otun, and B.J. Sweetman, Occupational hazards and low back pain. Rev Environ Health, 1987. 7(1-2): p. 121-60.
8. Adams, M.A. and P. Dolan, Time-dependent changes in the lumbar spine's resistance to bending. Clin Biomech (Bristol, Avon), 1996. 11(4): p. 194-200.
9. Adams, M.A., et al., Diurnal changes in spinal mechanics and their clinical significance. J Bone Joint Surg Br, 1990. 72(2): p. 266-70.
10. Gedalia, U., et al., Biomechanics of increased exposure to lumbar injury-caused by cyclic loading. Part 2. Recovery of reflexive muscular stability with rest. Spine, 1999. 24(23): p. 2461-7.
11. Le, P., et al., Cyclic load magnitude is a risk factor for a cumulative lower back disorder. J Occup Environ Med, 2007. 49(4): p. 375-87.
62


12. Solomonow, M., et al., Flexion-relaxation response to static lumbar flexion in males and females. Clin Biomech (Bristol, Avon). 2003. 18(4): p. 273-9.
13. Solomonow, M., et al., Biomechanics of increased exposure to lumbar injury caused by cyclic loading: Part 1. Loss of reflexive muscular stabilization. Spine, 1999. 24(23): p. 2426-34.
14. Sbriccoli, P., et al., Static load magnitude is a risk factor in the development of cumulative low back disorder. Muscle Nerve, 2004. 29(2): p. 300-8.
15. LaBry, R., et al., Longer static flexion duration elicits a neuromuscular disorder in the lumbar spine. J Appl Physiol, 2004. 96(5): p. 2005-15.
16. Sbriccoli, P., et al., Static load repetition is a risk factor in the development of lumbar cumulative musculoskeletal disorder. Spine, 2004. 29(23): p. 2643-53.
17. Courville, A., et al., Short rest periods after static lumbar flexion are a risk factor for cumulative low back disorder. J Electromyogr Kinesiol, 2005. 15(1): p. 37-52.
18. Abumi, K., et al., Biomechanical evaluation of lumbar spinal stability after graded facetectomies. Spine, 1990. 15(11): p. 1142-7.
19. Anderson, C.K., et al., A biomechanical model of the lumbosacral joint during lifting activities. J Biomech, 1985. 18(8): p. 571-84.
20. Posner I, W.A., Edwards W, et al.. A biomechanical analysis of the clinical stability of the lumbar and lumbosacral spine. 1982.
21. White A, P.M., Clinical Biomechanical of the Spine. 1978.
22. McGill. S.M. and R. W. Norman, Partitioning of the L4-L5 dynamic moment into disc, ligamentous, and muscular components during lifting. Spine, 1986. 11(7): p. 666-78.
23. Crisco, J.J., 3rd and M.M. Panjabi, The intersegmental and multisegmental muscles of the lumbar spine. A biomechanical model comparing lateral stabilizing potential. Spine, 1991. 16(7): p. 793-9.
63


24. Kaigle, A.M., S.H. Holm, and T.H. Hansson, Experimental instability in the lumbar spine. Spine, 1995. 20(4): p. 421-30.
25. Wilke, H.J., et al., Stability increase of the lumbar spine with different muscle groups. A biomechanical in vitro study. Spine, 1995. 20(2): p. 192-8.
26. Solomonow, M., et al., The ligamento-muscular stabilizing system of the spine. Spine, 1998. 23(23): p. 2552-62.
27. Stubbs, M., et al., Ligamento-muscular protective reflex in the lumbar spine of the feline. J Electromyogr Kinesiol, 1998. 8(4): p. 197-204.
28. Panjabi, M.M., Clinical spinal instability and low back pain. J Electromyogr Kinesiol, 2003. 13(4): p. 371-9.
29. Elizabeth Eversull, B.S., et al., Neuromuscular neutral zones sensitivity to lumbar displacement rate. Clin Biomech (Bristol, Avon), 2001. 16(2): p. 102-13.
30. Solomonow, M., et al., Neuromuscular neutral zones associated with viscoelastic hysteresis during cyclic lumbar flexion. Spine, 2001.26(14): p. E314-24.
31. Hagood, S., et al., The effect ofjoint velocity on the contribution of the antagonist musculature to knee stiffness and laxity. Am J Sports Med, 1990. 18(2): p. 182-7.
32. Tortora, G.J., Principles of Human Anatomy. Ninth ed. 2002.
33. Bogduk, N., Clinical Anatomy of the Lumbar Spine and Sacrum. Fourth ed. 2005.
34. Kumar, S., Biomechanics in Ergonomics. 1999.
35. Woo, S., Structure and Function of Ligaments and Tendons. Basic orthopaedic Biomechanics, 2004.
36. Solomonow, M., Ligaments: a source of work-related musculoskeletal disorders. J Electromyogr Kinesiol, 2004. 14(1): p. 49-60.
64


37. Adams, M.A., W.C. Hutton, and J.R. Stott, The resistance to flexion of the lumbar intervertebral joint. Spine, 1980. 5(3): p. 245-53.
38. Solomonow, M. and M. Krogsgaard, Sensorimotor control of knee stability.
A review. Scand J Med Sci Sports, 2001. 11(2): p. 64-80.
39. Rhalmi, S., et al., Immunohistochemical study of nerves in lumbar spine ligaments. Spine, 1993. 18(2): p. 264-7.
40. Gray, H., Gray's Anatomy of the Human Body.
41. Kim, C., The Multifidus Muscle is the Strongest Stabilizer of the Lumbar Spine. The Spine, 2007. 7(5): p. 76S.
42. Panjabi, M.M., The stabilizing system of the spine. Part I. Function, dysfunction, adaptation, and enhancement. J Spinal Disord, 1992. 5(4): p. 383-9; discussion 397.
43. Reeves, N.P., K.S. Narendra. and J. Cholewicki, Spine stability: the six blind men and the elephant. Clin Biomech (Bristol, Avon), 2007. 22(3): p. 266-74.
44. Rangayyan. R.M.. Biomedical signal analysis: A case-study approach. IEEE Press, 2002.
45. Solomonow, M., et al., Electromyogram power spectra frequencies associated with motor unit recruitment strategies. J Appl Physiol, 1990. 68(3): p. 1177-85.
46. Stulen, F.B. and C.J. DeLuca, Frequency parameters of the myoelectric signal as a measure of muscle conduction velocity. IEEE Trans Biomed Eng, 1981.28(7): p. 515-23.
47. Manolakis, D., V. Ingle, and S. Kogon, Statistical and Adaptive Signal Processing: Spectral Estimation, Signal Modeling, Adaptive Filtering and Array Processing 2005.
48. Oppenheim, A., and Schafer, R., Digital Signal Processing. 1975, Englewood Cliffs, NJ: Prentice-Hall.
65


49. Jackson, M., et al., Multifidus EMG and tension-relaxation recovery after prolonged static lumbar flexion. Spine, 2001.26(7): p. 715-23.
50. Solomonow, M., et al., Biexponential recovery model of lumbar viscoelastic laxity and reflexive muscular activity after prolonged cyclic loading. Clin Biomech (Bristol, Avon), 2000. 15(3): p. 167-75.
51. Baratta, R.V., et al., Methods to reduce the variability of EMG power spectrum estimates. J Electromyogr Kinesiol, 1998. 8(5): p. 279-85.
52. Indahl, A., et al., Electromyographic response of the porcine multifidus musculature after nerve stimulation. Spine, 1995. 20(24): p. 2652-8.
53. Indahl, A., et al., Interaction between the porcine lumbar intervertebral disc, zygapophysial joints, andparaspinal muscles. Spine, 1997. 22(24): p. 2834-40.
54. Williams, M., et al., Multifidus spasms elicited by prolonged lumbar flexion. Spine, 2000. 25(22): p. 2916-24.
55. Solomonow, M., et al., Muscular dysfunction elicited by creep of lumbar viscoelastic tissue. J Electromyogr Kinesiol, 2003. 13(4): p. 381-96.
56. Arendt-Nielsen, L. and K.R. Mills, The relationship betMeen mean power frequency of the EMG spectrum and muscle fibre conduction velocity. Electroencephalogr Clin Neurophysiol, 1985. 60(2): p. 130-4.
57. Moritani, T. and M. Muro, Motor unit activity and surface electromyogram power spectrum during increasing force of contraction. Eur J Appl Physiol Occup Physiol, 1987. 56(3): p. 260-5.
58. Solomonow, D., Neuromuscular Neutral Zones Response due to Cyclic Lumbar Flexion. (In Press), 2008.
59. Smit, T.H., The use of a quadruped as an in vivo model for the study of the spine biomechanical considerations. Eur Spine J, 2002. 11(2): p. 137-44.
60. Field, H., Taylor, M, An Atlas of Cat Anatomy. 1992.
66


61. Hirsch, C., B.E. Ingelmark, and M. Miller, The anatomical basis for low back pain. Studies on the presence of sensory nerve endings in ligamentous, capsular and intervertebral disc structures in the human lumbar spine. Acta Orthop Scand, 1963. 33: p. 1-17.
62. Yahia, L.H., N. Newman, and C.H. Rivard, Neurohistology of lumbar spine ligaments. Acta Orthop Scand, 1988. 59(5): p. 508-12.
63. Chu, D., et al.. Neuromuscular disorder in response to anterior cruciate ligament creep. Clin Biomech (Bristol, Avon), 2003. 18(3): p. 222-30.
64. Olson, M., M. Solomonow, and L. Li, Flexion-relaxation response to gravity. J Biomech, 2006. 39(14): p. 2545-54.
65. Olson, M.W., L. Li, and M. Solomonow, Flexion-relaxation response to cyclic lumbar flexion. Clin Biomech (Bristol, Avon), 2004. 19(8): p. 769-76.
66. Sbriccoli, P., et al., Neuromuscular response to cyclic loading of the anterior cruciate ligament. Am J Sports Med, 2005. 33(4): p. 543-51.
67


corresponding tension and displacement values during the relaxation phase were designated as Tension and Displacement Neuromuscular Neutral Zones (Offset). Figure 5.2 shows a typical record of tension, displacement, and EMG from channels L3/4, L4/5, and L5/6 for a single cycle test which helps to understand better how displacement NNZs and tension NNZs were computed. The arrows in the plot represent the EMG initiation and cessation thresholds.
01 31 2007 (0.25 Hz, 40N)
Displacement (mm)
Figure 4.5 Plot of the hysteresis that shows the tension versus displacement curve of a single cycle at 0.25-Hz frequency and 40-N amplitude. The black circle shows EMG initiation at the stretch phase. The Boldface on the curve designates the period where EMG was present and recorded throughout the cycle. The white circle represents EMG cessation in the release phase. The unBold part of the curve represents the NNZs.
29


4.4.2 Mean Creep
In each of the single cycle tests for each of the channels, the peak displacement was detected. Then, for the first three single cycle tests (pre-static loading), the displacement values were averaged and the mean (SD) was denoted as the displacement baseline value (0 percent creep). Following, the percentage creep was calculated using the following equation:
Creep = x 100%
A
Where D is the displacement value at each specific single cycle test, and D\ is the displacement baseline value. Then, the resulting values of the creep in the recovery period in each of the eight preparations were pooled, and then the mean (SD) were calculated and plotted as a function of time.
4.4.3 Normalized peak MAV
A full wave rectification was applied to the EMG recording of each of the single cycle tests for each of the channels, followed by a 200 points moving average with a forward shift of 10 points. Next, the peak amplitude of the processed EMG recordings was detected and designated as peak MAV. Then, the data for each preparation was normalized with respect to the value of the first
30


single cycle test in each data set. The normalized peak MAV was computed in order to show the maximal level of activity of the multifidus muscles throughout the experiment.
4.4.4 Median Frequency
In each of the single cycle tests for each of the channels, the 0.5 sec window of EMG data where maximum tension was present, was detected, and then zero padded by adding zero points to yield 512 points of data. Next, the EMG data was multiplied by a tukey window [48], The tukey window was specifically selected since it was the most similar to a rectangular window so it didn't attenuate a significant part of the signal compared with hamming or hanning windows. Then, the power spectral density was computed after applying the fast Fourier transform to the data. The baseline noise spectrum subtraction method [51] was applied to filter out 60 Hz noise and its harmonics from the signal. And that was achieved by calculating the power spectral density of a 0.5 sec window from the baseline noise then subtracting it from the power spectral density of the real EMG computed earlier. After that, the median frequency was calculated. The median frequency is the frequency that divides the power spectrum in half as shown in equation 1. Then, the median frequency data was run
31


through a three point moving average in order to smooth it. It was found that the increasing average conduction velocity during motor unit recruitment is the major contributor to variations in the electromyogram median frequency [45], So the changes in the median frequency could be an index to identify if the motor unit recruitment throughout the recovery period increased, decreased, or remained unchanged.
Where S(f) is the power density spectrum and f is the frequency
For the first three single cycle tests (pre-static loading) in each of the eight preparations, the Displacement NNZs were averaged and the mean (SD) was denoted as the baseline value. Similarly, the same process was applied to the Tension NNZs, normalized peak MAV, and median frequency. Then, the Displacement NNZs, Tension NNZs, peak MAV, and median frequency of the remaining tests (Recovery period) for each channel in each of the eight preparations were pooled, and then the mean (SD) were calculated and plotted as a function of time.
(4.1)
32


4.4.5 Mean of Percentage Change from Baseline
After assigning the baseline values, the percentage change from baseline values throughout the time was calculated in each of the eight preparations for each of Displacement NNZs (onset and offset), Tension NNZs (onset and offset), and normalized peak MAV. The average of the percentages for each category of data for the total of all the preparations was then computed and plotted versus time.
4.4.6 Statistics
A two-way analysis of variance (ANOVA) with repeated measures where time and vertebrae were set as independent variables, and where displacement onset, displacement offset, tension onset, tension offset, normalized peak MAV, and median frequency were set as dependent variables was applied to all the previous processed data to show the significance of the variations and to evaluate the effects of loading on the recovery period. Then, a one-way analysis of variance (ANOVA) with repeated measures was applied to the creep data with time as an independent variable. The Students T test was used to provide the details about the significance of variations. A normality test was first run to assure a normal distribution of the data. Significance was set at p<0.05.
33


4.5 Models
Displacement NNZs at onset and offset, Tension NNZs at onset and offset, and normalized peak MAV data in the 7 hours recovery period from all three channels were fitted to models that represent their pattern of variation in the recovery period and to compare it with the pre-static loading period. Modeling consisted of exponential equations plots using Marquardt-Levenberg algorithm.
Exponential functions were used since they represent the classical characteristics
2
of viscoelastic materials, r values were then derived to verify the accuracy of the best-fit models.
The model structure for the Displacement NNZs at onset and offset in the recovery period for each of the channels consisted of a single exponential function that can be described in the following equation:
DNNZ(t) = D0 + Dr + (D, -DR)e-('-r)lT' (4.2)
Where Do is the baseline value (pre-static loading period), Dr is the residual value, Dl is the first recorded value in the recovery period, and (D, DR) is the
decay amplitude of the exponential, t is the time, Tr is the time when the recovery period begins, and Tj is the time constant.
34


The Tension neuromuscular neutral zones followed a bi-exponential model as shown in the equation below:
WiVZ(/) = 70+(/-rs)rte'K)"! + TMeA'~TR)l^ (4.3)
Where 7o is the steady-state amplitude, Tl and Tm are the amplitudes of the exponential terms, and T? and Tj are respectively their correspondent time
constants. The term TMe~'~TR',T' represents the decreasing exponential.
Similarly, exponential functions were utilized in the modeling of the normalized peak MAV during the 7 hours recovery period. The model took the form shown in the following equation:
PeakMA V(t) = P0 + PLe'{,'TR)'u + PM (1 e~(,'lR)lT5) + (/ rd )PHe~(,~T)lT6 (4.4)
Where Po is the steady-state amplitude, Pi, Pm, and Ph are the amplitudes of the exponential terms, Tj and Ty and are respectively their correspondent time
constants, and T^ is the time delay where the delayed hyperexcitability starts. The
first term .P/e~,'~rs*/rj describes the exponential decay at the beginning of the recovery period that reaches its minimum at the first hour. The term PM{\ ) describes the steady-state recovery component. The term
35


(,t-rd)PHe rj)/r6 represents the delayed transient hyperexcitability. This term
starts affecting the equation only after the recovery exceeds the time delay T^.
36


5.
Results
In general, the recorded data throughout the experiment from all six channels demonstrated an exponential decay of the EMG activity during the six 10 min load periods, with a partial recovery during the six 10 min rest periods in between. The corresponding displacement during the 10 min loads showed an exponential increase caused by the laxity of the ligaments. During the recovery period, EMG activity increased gradually, while the displacement decreased exponentially to reach a near full recovery. Figure 5.1 shows a typical EMG recording from channels L3/4, L4/5, and L5/6 as well as tension and displacement. As emphasized by a red circle, during the third static loading period at channel L3/4, large amplitude spasms appeared with an EMG exponential decrease background. Spasms are sudden powerful contractions or stiffening of the muscles that occur as a response to tissue damage and are usually indicated by large EMG amplitudes.
37


Spasms
s 1 T £
11 03 2006 (Sne 0.25 Hz, static, 40N, 6x10-10)
f*4
HI -1 > 2 E
; + + +
ill ts ****+ + i i i +
lU -2 S' 3 E
(3 0 5
LU -3
P P P P p* t +-+- f + -1
1Q& ,_ f_ ,_ ,_ ,_ ,_ ,_ ,_
ioJ^AA A A A A A A A A
mm
0 10 20 30 40 50 60 70 80 90 100 110 120 130 150 170 200 260 320 380 440 500 560
Time (min)
Figure 5.1 EMG responses from channels L3/4, L4/5, and L5/6 along with the displacement and tension channels for a single preparation when subjected the experimental protocol.
And Figure 5.2 shows a typical record of tension, displacement, and EMG from channel L4/5 for a single cycle test. The superimposed curve in red represents the MAV of the EMG at this specific lumbar level. The Arrows in the top plot represent the EMG initiation and cessation thresholds. A projection of these arrows towards displacements and tensions graphs reveals their corresponding Displacement and Tension NNZs (onset and offset).
38
L-4/5 L-4/5 L- 5/3 L-4/5 L-3/4


11 17- 2006 (0.25Hz, 40N)
0.18
-0.18
_12
E
E, 8
Q.
b
o
60
§40
^20
0
12 8 4
0
60 40 20 0
5 6 7
Time (sec)
Figure 5.2 EMG responses from channels L3/4, L4/5, and L5/6 along with the displacement and tension channels for a single cycle test (0.25-Hz, 40-N).
Figure 5.3 shows the behavior of the hysteresis for a specific preparation at the pre-static loading period (plot 0) in comparison with its status after the first hour of recovery (plot l), after the second hour of recovery (plot 2), and after the sixth hour of recovery (plot 6). Maximum displacement took a value of 9.1 at the prestatic loading period (plot 0), and then displayed an increase to a maximum displacement value of 13.52 at the first hour of recovery (plot 1) which shows the laxity and the creep in the ligaments caused by the static loading period. Then, a
39


decrease in the maximum displacement started at the second hour of recovery (plot 2) and took a value of 12.67, and then resulted in a maximum displacement value of 11.42 at the sixth hour of recovery (plot 6).
01 31 2007 (0.25 Hz, 40N)
40
30
z
c
o
£ 20
I-
10
0
0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14
Displacement (mm) Displacement (mm) Displacement (mm)
Figure 5.3 Plots of the tension versus displacement (hysteresis) for a single preparation at the pre-static loading period, after 1 hour of recovery, after 2 hours of recovery, and after 6 hours of recovery.
5.1 Displacement NNZs
Figure 5.4 shows the mean Displacement NNZs in both the stretch and relaxation phases for all three lumbar levels L3/4, L4/5, and L5/6. The black circles indicate the mean Displacement NNZs at Onset throughout the time along with their standard deviation. The white circles represent the mean Displacement
40


NNZs at Offset along with their standard deviation. The first values at time zero in each of the graphs indicate the baseline values that were detected in the prestatic loading period. The superimposed curves in red represent the exponential models that were fitted to the data using Marquardt-Levenberg algorithm. The Students T test showed significant changes of the Displacement onset and offset NNZs over time. The Displacement NNZs at onset started at baseline values of 3.181(2.3), 2.575(1.75), and 3.167(2.24) for L3/4, L4/5, and L5/6, and displayed significantly increasing values to 8.33(2.04), 7.938(2.24), and 8.761(1.95) respectively immediately after the loading period which was equivalent to an average of 317.44 percent increase from baseline value as shown in Figure 5.5. Then, an exponential decrease in the mean Displacement onset NNZs occurred and reached final values of 3.263(1.78), 2.868(1.93). and 3.026(2.17) after the 7 hours recovery period at the respective levels, and that was equivalent to an average of 34 percent increase from baseline (Figure 5.5).
41


L4/5
L5/6
L3/4
lime (mm) Time (min) Time (min)
Figure 5.4 Mean displacement neuromuscular neutral zones for N=8 preparations
For the mean Displacement NNZs at offset, initial values of 7.871 (2.33), 7.72(2.33), and 7.727(2.38) took place respectively at lumbar levels L3/4,
L4/5, and L5/6. After the loading period, those values displayed a significant increase to 13.781 (1.82), 13.26(1.42), and 13.57(1.56) which was equivalent to an average of 85.9 percent increase from baseline value. An exponential decrease in the mean Displacement NNZs at offset was then observed and reached values of 7.618(2.64), 7.071 (2.5), and 7.52(2.86) after the 7 hours recovery period for the respective levels, and that was equivalent to an average of -2.575 percent difference from baseline value.
42


DNNZ
Figure 5.5 Mean of percentage change from baseline for displacement NNZs (onset and offset). The star symbol represents a statistically significant change from baseline values.
5.2 Tension NNZs
A log based 10 transformation was necessary to apply to the Tension NNZs at Onset data in order to achieve normal distribution. Statistics then showed significant variations of Tension NNZs at onset and offset over time. Figure 5.6 shows the mean Tension NNZs in both the stretch and relaxation phases for all three lumbar levels L3/4, L4/5, and L5/6. The black circles indicate the mean Tension NNZs at onset over time along with their standard deviation. The white circles represent the mean Tension NNZs at offset along with their standard deviation.
43


L3/4 L4/5 L5/6
Time (min) Time (min) Time (min)
Figure 5.6 Mean tension neuromuscular neutral zones for N=8 preparations
Initial baseline values of the Tension NNZs onset started at 9.864(5.76), 7.849(4.81), and 9.186(5.23) for L3/4, L4/5, and L5/6, then displayed insignificantly increasing values to 10.516(5.07), 9.581(5.03), and 11.742(5.36) respectively immediately after the loading period which was equivalent to an average of 42.98 percent increase from baseline value (Figure 5.7). A further significant increase took place afterwards and reached values of 10.961 (4.97), 12.067(9.39), and 14.842(10.64) respectively during the first hour of recovery which was equivalent to an average of 60.96 percent change from baseline (Figure 5.7). Following, an exponential decrease in the mean Tension NNZs at onset occurred and took significantly lower values than baseline
44


during the third hour of recovery. Those values were 9.545(5.68), 5.791(3.04), and 5.822(3.01) and that was equivalent to an average of -9.67 percent difference from baseline value (Figure 5.7). A further decrease resulted in final values of 4.755(3.23), 3.623(3.33), and 4.033(4.02) after the 7 hours recovery period at the respective levels which was equivalent to an average of -51.61 percent difference from baseline as shown in Figure 5.7.
TNNZ
Time (hr)
Figure 5.7 Mean of percentage change from baseline for tension NNZs (onset and offset). The star symbol represents a statistically significant change from baseline values.
The Tension NNZs at offset started at baseline values of 23.415(10.61), 21.314(9.33), and 21.888(10.05) for L3/4, L4/5, and L5/6, and displayed significantly increasing values to 27.895(8.83), 23.025(8.4), and 25.48(7.15)
45


respectively immediately after the loading period which was equivalent to an average of 29.83 percent increase from baseline value (Figure 5.7). A further increase took place afterwards and reached values of 26.03(8.28), 27.86(10.4), and 28.463( 10.94) respectively during the first hour of recovery which was equivalent to an average of 35.6 percent increase from baseline (Figure 5.7).
Then, an exponential decrease in the mean Tension NNZs at offset occurred and took significantly lower values than baseline during the fourth hour of recovery. Those values were 19.858(10.34), 18.207(9.62), and 16.805(9.2) and that was equivalent to an average of -21.05 percent difference from baseline value (Figure 5.7). A further decrease resulted in final values of 12.727(9.47), 9.276(6.86), and 11,495(7.66) after the 7 hours recovery period at the respective levels which was equivalent to an average of -48.56 percent change from baseline (Figure 5.7).
5.3 Normalized peak MAV
Student's T test's results showed significant changes of the normalized peak MAV over time. Figure 5.8 shows the mean normalized peak MAV along with their standard deviation for all three lumbar levels L3/4, L4/5, and L5/6 versus the time. Initial baseline values started at 1 for lumbar levels L3/4, L4/5, and L5/6, then displayed insignificantly decreasing values to 0.941 (0.58),
46


0.866(0.25), and 0.758(0.51) respectively after the 2 hours loading period which was equivalent to an average of -14.44 percent difference from baseline value (Figure 5.9). Following, a significant exponential decrease in the mean normalized peak MAV occurred and took values of 0.643(0.26), 0.672(0.26), and 0.648(0.35) during the first hour of recovery and that was equivalent to an average of -33.72 percent change from baseline (Figure 5.9). An exponential increase took place afterwards and was then followed by a further significant delayed exponential increase that reached final values of 1.535(1.17),
1.541 (0.86), and 1,342(0.8) after the 7 hours recovery period at the respective lumbar levels which was equivalent to an average of 47.32 percent increase from baseline (Figure 5.9). Those final values were statistically different from baseline.
L3/4 14/5 L5/6
Time (min) Time (min) Time (min)
Figure 5.8 Mean normalized peak MAV forN=8 preparations
47



peak MAV
Figure 5.9 Mean of percentage change from baseline for normalized peak MAV. The star symbol represents a statistically significant change from baseline value.
5.4 Mean Creep
The initial baseline value was denoted as zero creep. Following, the creep demonstrated a significant increase after the 2 hours static loading period to reach a maximum value of 56.687%(23.33%). Then, the creep showed a significant decrease throughout the recovery period that reached a final value of 12.21%( 16.23%) at the seventh hour, and that was still significantly different from baseline as shown in Figure 5.10.
48


70
60
50 CL
CD 40 CD
^ 30 c CD
20 10 0
-10
0 100 200 300 400 500 600
Time(min)
Figure 5.10 Mean of the creep for N=8 preparations. The star symbol represents a statistically significant change from baseline value.
5.5 Median Frequency
After running the Students T test, results did not show any statistically significant changes in the median frequency over time. Figure 5.11 shows the mean of the 3-point averaged median frequency along w ith their standard deviation versus the time for all three lumbar levels. Initial baseline values for the median frequency started at 209.54(26.52), 209.63(26.15), and 193.19( 14.37) for lumbar levels L3/4, L4/5, and L5/6, then displayed insignificantly decreasing values to 201,66(42.99), 196.53(32.49), and 178.95(22.19) respectively after the 2 hours loading period. Then, an increase of the median frequency over time was observed and reached final values of 210.44(27.93), 202.75(11.71), and
Significant
change
49


192.38( 14.39) for the respective levels. None of those variations was statistically significant. The median frequency was initially computed in order to identify the motor unit recruitment status throughout the recovery period.
L3/4
L4/5
L5/6
o 220
C
Q)
cr 210
Q)
£ 200 TO T3 (1)
^ 190
< 180 Q.
m 170
'S
ra 160 0)
S
Time (min)
Time (min)
0 100 200 300 400 500 600
Time (min)
Figure 5.11 Mean of 3-point averaged median frequency forN=8 preparations
5.6 Models
Best-fit models were superimposed on the experimental data. Modeling consisted of exponential equations plots using Marquardt-Levenberg algorithm. In order to verify the accuracy of those models, r values were derived. Then, results displayed r2 values ranging from 0.9524 to 0.9876 for the displacement NNZs model (Table 5.1), from 0.8645 to 0.9927 for the tension NNZs model (Table 5.2), and from 0.9175 to 0.9848 for normalized peak MAY model (Table 5.3).
50


Time constants were noticed to be higher for Displacement NNZs at offset and Tension NNZs at offset comparing with Displacement NNZs at onset and Tension NNZs at onset respectively which indicates a slower recovery in the relaxation phase. Values of the time constants for Displacement NNZs at onset ranged from 127 to 170 min while they ranged from 214 to 232 min for the Displacement NNZs at offset.
Table 5.1 Displacement neuromuscular neutral zones model parameters
DNNZ(t) = D0+Dr+(Dl DR)e-[i,'rR)/rI]
Onset Offset
L3/L4 L4/L5 L5/L6 L3/L4 L4/L5 L5/L6
Do 3.18 2.57 3.16 7.87 7.72 7.72
Dr 0.082 0.293 -0.141 -0.252 -0.648 -0.207
Dl 4.95 5.36 5.75 5.65 5.56 5.75
Tl 170.5 127.6 126.9 217.0 231.6 213.6
0.9813 0.9876 0.9767 0.9709 0.9739 0.9524
51


Table 5.2 Tension neuromuscular neutral zones model parameters
TNNZ(t) = To+U-thJTi. e[(,-rR)lx2]+ TM e[u'TR)lr3]
Onset Offset
L3/L4 L4/L5 L5/L6 L3/L4 L4/L5 L5/L6
T0 5.58 3.52 4.49 0.00 0.00 0.00
Ti 0.123 0.358 0.480 0.156 0.225 0.253
T2 50.1 27.9 27.8 100.7 80.6 74.2
Tm 4.84 6.00 7.16 24.5 23.9 24.7
T3 140.0 144.9 106.4 643.9 485.7 498.2
r2 0.8645 0.9927 0.9831 0.8775 0.9468 0.8974
Table 5.3 Normalized peak MAY model parameters
Peak MA V(t) = P0+PL el(,'lR)l r4]+ PM( 1 el(MRm)+{t-rd) PH e[(Md)'z6]
L3/L4 L4/L5 L5/L6
Po -8.069 -7.64 -5.57
Pl 9.00 8.52 6.33
t4 40.0 58.7 37.7
Pm 8.99 8.81 6.57
Ts 43.8 65.1 42.0
T Ph 0.006 0.004 0.004
t6 287.4 230.0 233.5
r' 0.9835 0.9175 0.9848
52


Full Text

PAGE 1

NEUROMUSCULAR NEUTRAL ZONES RESPONSE TO STATIC LUMBAR FLEXION By Jimmy E. Youssef B.S., University of Balamand, 2005 A thesis submitted to the University of Colorado Denver In partial fulfillment Of the requirements for the degree of Master of Science in Electrical Engineering 2008

PAGE 2

This thesis for the Master of Science Degree by Jimmy E. Youssef Has been approved By 2 Date

PAGE 3

Youssef, Jimmy E. (M.S., Electrical Engineering) Neuromuscular neutral zones response to static lumbar flexion Thesis directed by Associate Professor Miloje Radenkovic ABSTRACT The objective of this thesis was to study the effect of prolonged static lumbar flexion at moderate load on the spine's stability. Eight preparations of in vivo feline models were subjected to 40N static loading in a series of 6 periods of 10 minutes of work spaced by 10 minutes of rest, followed by a seven hours rest period. Reflexive multifidi electromyogram (EMG) initiation threshold in the stretch phase, EMG cessation threshold in the relaxation phase along with their corresponding displacement, and tension thresholds that trigger that reflexive muscular activity was recorded. A significant increase in the Displacement Neuromuscular Neutral Zones (NNZs) was observed after the static loading period followed by an exponential decrease to its normal value over 7 hours. Similarly, the Tension NNZ showed an increase followed by a decrease below baseline after 2 to 3 hours of recovery. A decrease in the peak MA V occurred and was then followed by an increase that exceeded the baseline. No variability in the EMG median frequency was detected throughout the recovery period. The results suggest that laxity in the ligaments and decreased reflexive muscular activity in the first 2 hours of recovery following the static loading period, leaves the spine unprotected and under risk of injury. During the remaining recovery time, a compensatory muscle activity takes over, lending an increased protection to the spine, and resulting in limited motion and muscle stiffness. Workers exposed to static loading of their spine should protect it for the first 2 hours after work since spinal stability is compromised in this period. This abstract accurately represents the content ofthe candidate's thesis. I recommend its publication. _________ Miloje Radenkovic

PAGE 4

DEDICATION I dedicate this thesis to my parents and family who always encouraged me to pursue knowledge and have shown me love and support m y whole life. I also dedicate this to my cousin Joe who showed me what real courage and strength are and even though he is not with me anymore and never had the chance to pursue his dreams ; I share this with him and he 'll always be in my heart.

PAGE 5

ACKNOWLEDGMENT I thank God for always being there for me, keeping me going throughout the stressful times and leading me to success. I want to express my most sincere and utmost gratitude to my parents Edmond and Fayrouz. Mom and Dad thank you for never failing to believe in me. Thanks for all your sacrifices and hard work to give me the best education and the best future. You have always put us your children first which I will never forget. Thanks Mom for your constant prayers and I hope that I will always make you proud. I also want to thank my family in the United States that gave me support a place to stay and love. They have been so generous to me and given me a home away from home. Thanks Tony Kellie Taylor Jon, and Diana. I want to express great thanks to my advisor Dr Moshe Solomonow who was a great professor and mentor throughout this study Thanks for always looking out for me and guiding me by giving me valuable lessons not only in academia but also in life. You always treated me like a part of your family ; I really appreciate all the opportunities and the help you gave me To Dr. Yun Lu and Dr. Bing-He Zhou thank you for sharing your knowledge with me I thank Dr. Lu for teaching me a lot about anatomy and machining. Thank you Dr. Zhou for teaching me valuable information about data acquisition and signal processing. I also would like to thank Dr. Bradley Davidson for introducing me to data modeling statistics and presentation skills. Finally I would like to thank Dr. Miloje Radenkovic and Dr. Robert Grabbe for taking the time to review my thesis. Thanks Dr. Mike for teaching me a lot about signal processing and introducing me to Dr. Solomonow which led to this great learning experience.

PAGE 6

TAB LE OF CONTENTS Fig u res .... ...... ........................................................................... i x Tables .... .. .. .. .. .. ................ .. .................. .. ........................... xii Chapter I. Introduction ............................................................................ 1 1.1 Sig n ificance of Study ............................... .. ..... ................. ...... 3 2. Phys i ological Background .......... ..... ..... ....... ............... .. ..... .. ...... 5 2.1 Vertebral Column .................... .. ................... ............ .. .. .. .... 5 2.2 Intervertebral Discs ................................................................. 7 2 3 Ligaments .............. ............................... .. .. .............. ... .. 8 2.4 Mechanoreceptors ....... .. ...................................................... 11 2.5 Multifidus Muscles .. ....... .. ..... .. .. .......................................... 12 2 6 Spinal Stabilizing Feedback Control System .............. ................... 14 3. E lectr omyography (EMG) ......... .. ............ .................................. 17 3 1 Motor Unit Recruitment and EMG Median Frequency ...................... 18 3 2 Zero Padding .. .............. .. .. .. .................. ......................... 19 3.3 Windowing ....................................................................... 20 4 Methods and Procedure ............................................................. 23 4 1 Preparation .............................. ............... .. ........... ............. 23 VI

PAGE 7

4.2 Ins t rume n tat i o n ........... .. ....... ..... .. .. .. .. ........ ............. 25 4.3 Expe r ime n ta l Protocol. ............. ...... .. .. .. ....... .. .. .. ............ 26 4 4 Dat a Analysis ......... . . ............... ..... ............. ........ .. .. ... 28 4.4. 1 D i sp lace m e n t NNZs and Tension NNZs .. .......... .. .. ............ ... 28 4.4 2 Mea n C r eep ........ .. .. .. .. .......... ...... ...... . .. .. ..... ... .. 30 4.4 3 Normalize d p ea k MAV ............. . .. . .. .. .. ..................... 30 4.4.4 Med i a n Fre q uency ...... .. ............. .. .. .. ...... .... ................... 31 4.4 5 Mea n of Perce n tage Change from Base l ine ....... ...... .. ...... .... 33 4.4.6 St a tist i cs ............. ........................................................ 33 4 5 Models ......... .. . .. ................ .. . ...... .. .. . .. ....... .... 34 5 R esults .. ......... .. .. .. ............. .. ......... ................... ..... .... ...... 37 5.1 Di s p laceme n t NNZs ........... . ...... . .. ....... ...... .. .. ....... ... 40 5 2 Te n s i o n NNZs .. ....... .. ....... ..... .. ........ . .. .. . .......... ... 43 5 3 Normali zed peak MA V ..... . . .. ......... . ........... . . .. ... 46 5.4 Mea n Creep ........ ......... ....... ....... .. .. .. .. ................ ....... 48 5 5 M e di a n F r e qu e n cy ..... .. .. .. ...... .. ....... .. ...... ................ .. .... 49 5.6 M o d e l s .... ...... ................ ........ .. ..... .. ........ .. ................ ..... 50 6. A n a l ysis .............. ...... .. ..... ....... .. .. .. .. . .. ......................... 53 6.1 S pin e S t ability . ....... ........... ...... .. . .. ...... .. .. . ........ ... 53 VII

PAGE 8

LIST OF FIGURES Figure 2.1 Spinal vertebrae (NIH 2007) .......... ..................................... ..... 6 2.2 Schematic of anterior flexion and extension .. ........................... ... 7 2 3 Intervertebral D i sc (NiH 2006) .............. ......... .. ................. ....... 8 2.4 Behavior of the Creep of viscoelast i c tissues under constant load ...... ..... 9 2 5 Ligaments of the lumbar spine ........................................... ... 10 2 6 Mu l tifidus muscles (Adapted from Gray s Anatomy o f the Human Body 2000) ..... .................. ....................... ..... .. .. ..... 13 2.7 Simpl i fied feedback contro l system of the spine (adapted from Solomonow et aI, 200 I ) ......................................................... 15 3.1 Commonly used window functions .. .. .. ................................ 21 4.1 Schematic of t h e experimental setup showing the external fixations placed on L I and L 7 and the hook applied to the L4 / 5 ligament during rest period (a) and during static loading period (b) ................ ...... ... ... 24 4 2 Stainless steel wire electrodes .... ...................................... .. .... 25 4.3 Actual picture of the preparation showing the external fixators the electrodes and the S shaped hook inserted in the L4 / 5 supraspinous l igament. ........................ .. ......... .. ....... .......... ................. ... 26 4.4 Schematic of the protocol that was followed i n the experiment with N=6 repetitions ....................................................... ..... 27 4 5 Plot of the hysteresis that shows the tension versus displacement curve of a single cycle at 0.25-Hz frequenc y and 40-N amplitude. IX

PAGE 9

The black circle shows EMG initiation at the stretch phase. The Boldface on the curve designates the period where EMG was present and recorded throughout the cycle. The white circle represents EMG cessation in the release phase. The un Bold part of the curve represents the NNZs ........................................ 29 5.1 EMG responses from channels L3/ 4, L4 / 5 and L5/ 6 along with the displacement and tension channels for a single preparation when subjected the experimental protocol. ........ ............................ 38 5.2 EMG responses from channels L3 / 4 L4 / 5, and L5/ 6 along with the displacement and tension channels for a single cycle test (0.25-Hz 40-N) ..................... .............................................. 39 5.3 Plots of the tension versus displacement (hysteresis) for a single preparation at the pre-static loading period after I hour of recovery after 2 hours of recovery, and after 6 hours of recovery ..................... 40 5.4 Mean displacement neuromuscular neutral zones for N=8 preparations .................................................. ................ 42 5.5 Mean of percentage change from baseline for displacement NNZs (onset and offset). The star symbol represents a statistica lly significant change from baseline values .................................................. .. 43 5.6 Mean tension neuromuscular neutral zones for N=8 preparations ............ ............................................................ 44 5.7 Mean of percentage change from baseline for tension NNZs (onset and offset). The star symbol represents a statistically significant change from baseline values .................................................... 45 5.8 Mean normalized peak MA V for N = 8 preparations ......................... 47 5.9 Mean of percentage change from baseline for normalized peak MAV. The star symbol represents a statistically significant change from baseline value ............................................ ....................... 48 x

PAGE 10

5.10 Mean of the creep for N=8 preparations. The star symbol represents a statistically significant change from baseline value ........................ 49 5.11 Mean of 3 point averaged median frequency for N=8 p r eparations ....... 50 Xl

PAGE 11

LIST OF TABLES Tables 5 1 Displacement neuromuscular neutral zones model parameters ............ 51 5.2 Tension neuromuscular neutral zones model parameters .................... 52 5 3 Normalized peak MA V model parameters . ......... .. ............... ...... 52 XII

PAGE 12

1. Introduction Low back pain is a widespread problem affecting roughly 80 percent of Americans at some point in their lives. It's considered the fifth most common reason for visiting a doctor the most common cause of work-related disability in people under age 45 and one of the first reasons for missing work [1, 2]. Low back injuries mostly occur in the workplace, especially where workers are exposed to intense physical activities as part of their daily occupational routine [3]. In such cases the injury usually occurs after the work is completed while performing simple actions And the cost for this problem is estimated to be a staggering $50 Billion yearly. Previous reports showed that workers subjected to occupational activities requiring sessions of cyclic or prolonged static lumbar flexion (such as mechanics farm workers concrete and roofing workers etc.) report with a high rate of low back disorders [2, 4-7] which is considered one of the most costly areas of general musculoskeletal disorders based on medical expenses disability payments, and lost wages. Previous literature also proved that exposing the lumbar spine to high magnitude static loading periods or to moderate and high magnitude cyclic loading periods elicited creep of the viscoelastic components which led to a cumulative trauma disorder (CTO) resulting in pain muscular

PAGE 13

weakness, and stiffness [8-13]. High load magnitudes [11, 14], longer loading durations [15], high number of repetitions for a given period [16] and shorter periods of rest in between [17] are high risk factors for static and cyclic flexion. It has also been established that cyclic loading is more challenging than static loading on the viscoelastic tissues [II]. A significant part of low back pain problem is known to be of electromechanical origin and often referred to as spinal instability. Spinal instability could be explained as an excessive motion between vertebrae with high risk of injury caused by elicited creep in the ligaments The passive viscoelastic structures (e g. ligaments discs and capsules) have a minor role in maintaining the stability of the spine [18-22], whereas the major role goes to active forces generated from reflexive muscular contractions [23-25]. It was proven that electrical or mechanical stimulation to the lumbar viscoelastic tissues triggered a muscle activity in the multifidus muscles confirming the presence of a ligamento muscular reflex [26,27]. It has also been established that the spine is relatively compliant for small perturbations about the neutral position [28]. These small displacement ranges within which stability is not disturbed and where muscular contraction wasn t required were designated as Neuromuscular Neutral Zones 2

PAGE 14

(NNZ) [29]. During these neutral zones, no muscular activity (EMG) will be observed, and the ligaments are stiff enough to support the spi ne. Previous studies showed that the neuromuscular neutral zones become narrower with the increase of the frequency of flexion [30 31] in order to maintain spine stability and that they have a pattern of change across the lumbar spine [29]. It also showed the effect of static and cyclic lumbar loading on low back pain. What hasn't been investigated yet is the effect of prolonged static lumbar loading on the neuromuscular neutral zones, thus on spine stability. This project wi II explore the behavior of neuromuscular neutral zones in response to moderate static lumbar loading of 40N on in order to simulate a labor work environment. We hypothesize that a prolonged moderate static load over a moderate duration will result in increased neuromuscular neutral zones in the first few hours of rest due to the laxity in the viscoelastic tissues leaving the spine unprotected and under risk of injury. 1.1 Significance of Study The results f rom this study may provide experimental and biomechanical information about spine instability and may present guidelines for designing safe work schedules to prevent spinal injury. This will require further research to 3

PAGE 15

extrapolate obtained data from the feline model to a human model. It will also provide a new insight into the ligamento-muscular feedback loop that maintains the spine stability. 4

PAGE 16

2. Physiological Background It's very important to understand the anatomy and physiology of the lower back and its stabilizing system in order to understand the effect of static loading on the spine's stability. This section describes the anatomy of the lower back specifically the lumbar spine area to help understand its structure and function including some of the elements such as vertebral column intervertebral joints, ligaments multifidus muscles, and the stabilizing feedback control system of the spine. It also briefly explains motor unit recruitment and its relationship with muscle contraction 2.1 Vertebral Column The vertebral column, also called the spine or backbone, extends from the base of the skull to the tip of the coccyx. It s composed of33 vertebrae ; 7 cervical 12 thoracic 5 lumbar 5 sacral (fused together) and 4 coccygeal (fused together) (Figure 2.1). The spine is flexible yet extremely tough and serves to support the back through a full range of motion (anterior and lateral flexion and extension, and rotation). Anterior Flexion is a forward movement (anterior bending) whereas extension is a backward movement (posterior bending) and rotation is a twisting of the vertebral column (Figure 2.2). 5

PAGE 17

Z lihoracic vertebrae 5 L umba r vertebra e Figure 2.1 Spinal vertebrae (NIH 2007) The vertebral column protects the spinal cord which runs from the brain through the canal in the middle of the vertebral column (vertebral foramen) and also serves as a stable point of attachment for the ribs the pelvic girdle and the muscles ofthe trunk [32] The area of focus is the lumbar section from lumbar level 1 to lumbar level 5 (L l-LS) where the vertebrae are the largest segments and the strongest. The lumbar spine is most frequently involved in back pain because it withstands most of the pressure from body weight and is subjected to the largest forces and stresses along the spine 6

PAGE 18

Figure 2.2 Schematic of anterior flexion and extension 2.2 Intervertebral Discs Intervertebral discs are located in between every two adjacent vertebrae and form one fourth of the spine s length. They consist of an outer ring made of collagen fibers and referred to as anulus fibrosus and an inner elastic gelatinous material that contains fluid and is known as nucleus pulposus (Figure 2.3) The structure of the discs allows them to act as cushions change shape while permitting various movements of the vertebral column but still resist excessive motion in order to form a strong joint. The disc is viscoelastic and develops creep 7

PAGE 19

under load applied over time. The intervertebral discs also s erve as shock absorber s [33] and loose fluid content and height when injured [9]. Mnulus r '--X"b
PAGE 20

extension of the lumbar spine. They also gradually lengthen up to a finite maximum when under constant tension and then recover exponentially after cessation of the tension and this is referred to as Creep (Figure 2.4). T im e Figure 2.4 Behavior of the Creep of viscoelastic tissues under constant load Creep is described as the percentage of the elongation divided by the initial displacement as shown in the following equation C Elongation 1001 reep = x 1 0 Initial displacement (2.1) Ligaments are considered as passive tissue thus they do not contract voluntarily to generate a force. lnstead they become stiffer with increasing applied loads to prevent excessive displacement of the bones and damage to the joint [36]. The tension in a ligament decreases with time when under a constant 9

PAGE 21

elongation and this behavior is referred to as tension-relaxation. Ligaments also display a hysteresis loop where the tension-displacement curve is different during stretch (loading) and relaxation (unloading) phases resulting in net internal energy loss [34]. The spine has numerous ligaments that connect individual vertebrae such a s the anterior longitudinal ligament, the posterior longitudinal ligament, transverse ligament ligamentum flavum facet capsulary ligament, interspinous ligament and supraspinous ligament. The focus of study will be on the supraspinous ligament since it' s the first ligament to fall under tension and elongation during forward flexion [37]. The supraspinous ligament runs between the tips of the spinous processes of adjacent vertebrae (Figure 2.5). Figure 2.5 Ligaments of the lumbar spin e 10

PAGE 22

2.4 Mechanoreceptors The human body contains five kinds of receptors: thermoreceptors nociceptors photoreceptors chemoreceptors and mechanoreceptors [32]. Thermoreceptors are responsible for detecting changes in temperature nociceptors detect pain photoreceptors detect light in retina of the eye chemoreceptors respond to chemicals in the mouth and smell in the nose and mechanoreceptors respond to mechanical pressure or stretching by firing nerve impulses (action potentials) that travel though neural pathways to the central nervous system and then result in a muscle activity which is known as ligamento muscular reflex There are five main types of mechanoreceptors: Meissner s corpuscles Golgi organs Merkel s discs Ruffini corpuscles and Pacinian corpuscles. The Meissner s corpuscles are rapidly adapting receptors for fine touch and detecting changes in texture whereas Merkel s discs are slowly adapting receptors and detect sustained touch and pressure. Pacinian corpuscles are fast adapting receptors with low thresho l d and response to deep pressure and rapid variations, and Ruffini s corpuscles detect sustained pressure [32]. The Golgi organs respond to extreme force ranges [38] 1 I

PAGE 23

Ligaments are known to contain mechanoreceptors thus they are considered as sensory organs and they relay information to the nervous system. The supraspinous ligament was shown to contain 2 types of mechanoreceptors: Type II and III [39]. Type II receptors are mostly fast adapting; they signal the initiation and cessation of stimuli to the central nervous system, and provide continuous signaling for changes in tensile forces. Whereas type III mechanoreceptors are slow adapting and signal only during extreme changes in motion and load. 2.5 Multifidus Muscles Paraspinal muscles are the posterior muscles immediately next to the vertebral column They allow mobility and stability to the spine Our interest in this study is specifically the multifidus muscles that fill up the groove on both sides (left and right) of the spinous processes of the ve rtebrae as shown in (Figure 2.6). They usually span 2 to 4 joint segments and connect the mamillary process of one vertebra to the spinous process of the second, third and fourth vertebra above [40]. This allows them to exert compressive forces between vertebrae with higher role in extension. 12

PAGE 24

The multifidi are the deepest muscles in the back and the closest to the vertebral column. They have a large cross-sectional area and short fiber length, thus their main role is to limit excessive motion across individual motion segments in order to provide stability in the lumbar spine region rather than mobility [41] Multifidus muscles are also known to be the major intervertebral stabilizing muscles of the spine. Figure 2.6 Multifidus muscles (Adapted from Gray's Anatomy of the Human Body 2000) 13

PAGE 25

2.6 Spinal Stabilizing Feedback Control System The spinal stabilizing control system consists of three subsystems [42]. The passive subsystem consists of the vertebrae and the viscoelastic tissues (ligaments discs facet capsules etc.). The active subsystem consists of the active muscles surrounding the spine. The neural subsystem includes the nerves and the Central Nervous System (CNS) and monitors the transducer feedback signals from the passive subsystem for the relay of loading displacement, and proprioception [42] and reacts to it by directing the active subsystem to maintain the needed spinal stability. Spine stability is clinically defined as the ability of the spine to bear loads allow movement and still avoid injury and pain [43]. This definition implies that a dynamically stable spine should not have excessive intevertebral motion in response to a perturbation and should return rapidl y to its undisturbed position. The term stability is used in this study for its clinical meaning that is often confused with the mechanical term robustness. For clarity a system that is clinically less stable refers to a system that is mechanically less robust towards becoming unstable As stated earlier the spine stability is maintained partly by the passive viscoelastic tissues that generate tension expressed as resistance to destabilizing motion and mainly by the active forces generated from the 14

PAGE 26

musculature through a reflexive system that involves sensorimotor feedback This gives rise to the feedback control system of the spine shown in figure 2.7. eNS M otor Cor tlOI+ 0-M otorunits Dynamics eNS 1 o d u l ation Spmal JMerneu!"On D ynamics ToCNS Ij 1 !--Muscu l oskeleta l r--Dynamics CellOynamPcs Leng h Vetoclt,' A c cele rilti c n etc Viscoe l ast i c Dy n amics I I I I I Figure 2.7 Simplified feedback control system of the spine (adapted from Solomonow et ai, 200 I) This figure depicts the major anatomic structures associated with the forward and feedback loops of the spine stabilizing system [30]. In the forward loop the central nervous system (controller) activates the motorunits dynamics the musculoskeletal dynamics and the viscoelastic dynamics to regulate the force length velocity acceleration etc. The feedback loop also called ligamentomuscular reflex is the focus of our study and starts with a perturbation in the spine then the mechanoreceptors sense a deformation in the viscoelastic tissues 15

PAGE 27

which leads to an afferent neuron. The afferent neuron signal travels to the spinal cord, passes through interneurons and finally reaches the multifidi and generates a muscle contraction thus stabilizing the spine. It s important to recall that a feline model is being used in this research. Even though differences between the feline model and humans are evident scientific studies have showed that feline models are appropriate for this type of investigations as will be discussed later in the report. Similar behavior of the viscoelastic tissues and the stabilizing system of the spine is expected. 16

PAGE 28

3. Electromyography (EMG) Electromyography (EMG) is the measurement and recording of electrical muscle activity. Recordings are usually made at rest and during contraction. A skeletal muscle fiber has a membrane resting potential of -70 m V during rest. Muscles contract when stimulated by action potentials traveling through motor neurons along muscle fibers. An action potential is a pulse-like wave that triggers a muscle activity. Action potentials have the same magnitude and are generated when a threshold of usually -55 m V membrane potential is reached. The spatio-temporal summation of all the action potentials from all the active motor units (MUAPs) constitutes the electromyogram of the muscle [44] and forms its random shape. The electromyogram varies in frequency and amplitude depending on the intensity of the muscle contraction. There are three kinds of electrodes surface electrodes needle electrodes and wire electrodes In this study bipolar fine wire electrodes were inserted in deep muscles as they are capable of picking up EMG signals at rest and during contraction only from the multifidus muscles and without crosstalk interference from superficial muscles Wire electrodes were used for their increased flexibility compared with needle electrodes and for their small impedance compared with surface electrodes. 17

PAGE 29

3.1 Motor Unit Recruitment and EMG Median Frequency A motor unit is the connection between an efferent neuron coming from the spinal cord and a series of muscle fibers. Fewer muscle fibers in a motor unit results in a fine motor control, whereas more muscle fibers in a motor unit results in a stronger muscle contraction. Muscle force and strength are regulated by many processes such as motor unit recruitment and increase of action potential firing rate. Motor unit recruitment is the process of selectively activating motor units. Motor units usually get recruited in an orderly pattern from smaller motor units progressively to larger motor units with the increase in the muscle force necessity. Once all motor units are fully active, maximal force is reached by the increase of the firing rate. The smaller motor units have a small diameter and are resistant to fatigue, but generate small force increments whereas larger motor units have a larger diameter (larger conduction velocity) and fatigue fast but generate large force increments. OUf study of interest is the multifidus muscles that are considered slow twitch muscles and consist of mainly smaller motor units. Median frequency (MF) is the frequency that divides the power spectra of the EMG signal in half. Previous studies have shown that the increase in the motor unit recruitment results in an increase in the average conduction velocity 18

PAGE 30

(CV) and also causes a linear increase in the median frequency [45] showing that the average conduction velocity is linearly proportional to the MF [46]. This proves that an increase in the motor unit recruitment indicates an increase in the MF. Thus the MF can be an indicator for motor unit recruitment and behavior of the muscles due to different loading conditions After sustaining long work periods the larger and faster motor units start to fatigue, thus resulting in a decrease in the MF and an increase of EMG amplitude. The decrease in the MF indicates the deactivation of larger motor units in a process known as orderly derecruitment, and the increase of the EMG amplitude indicates the increase of the firing rate of the still active smaller motor units. In the absence of muscle fatigue the EMG amplitude and MF are related where a higher requirement of muscle force will be satisfied by an increase of the EMG amplitude and an increase in the MF. In this study muscle fatigue is not expected to occur and EMG peak Mean Absolute Value is expected to be related to the EMG median frequency. 3.2 Zero Padding Zero padding consists of extending a dataset with zeros to a power-of-two length It is often applied as a step in the signal processing associated with the 19

PAGE 31

spectral analysis. Some FFT algorithms require zero padding in order to run faster and more efficiently. Zero padding also enables the power spectrum to be evaluated over a fuller range of frequencies. However it can t increase resolution but it only shows a better display by providing a high density spectrum [47]. 3.3 Windowing A window function is a function that has a zero value out of a specific interval. Window functions are usually multiplied by data signals and result in their product inside the specific interval, and in a zero-valued function outside the interval. For instance, a rectangular function has a constant value inside the interval and a zero value outside the interval. It can be an arbitrary finite-duration window and can be of unit amplitude to preserve the original signal s amplitude. But an issue arises for the rectangular window since it introduces discontinuities at the end of the data signal and it initiates side lobes into its Fourier transform because of the fast switching rectangular function. Then, some different window functions were introduced and were able to smoothly attenuate the signal s amplitude to zero with time, which diminishes the power of the side lobes in the frequency domain (Hanning window). Each window function has its advantages and disadvantages and each is more effective for different types of signals. Some 20

PAGE 32

windows improve the frequency resolution but attenuate the original amplitude of the signal, and some windows save the original signal s amplitude but introduce side lobes to its Fourier transform [48]. Some of the most popular windows are Hamming Barlett Hanning Tukey and Blackman window (Figure 3.1). 0 8 0 6 0 4 0. 2 I i I I I I T \jKey f I etl r = 0 R ec t r = 0 .25 Tuk ey r = 1 Hannin g 2 0 30 Sa m ples Figure 3.1 Commonl y used window functions Reo \ <0 50 60 The difference between all the previously mentioned windows is the ratio of taper to constant sections (r). The Tukey window was selected in this research since it doesn t attenuate the amplitude of the whole signal because of its unit 21

PAGE 33

amplitude unlike the Hanning window, and since it doesn t introduce really high frequency components because of the presence of the ramp unlike the rectangular window. The Tukey window has a ratio of taper to constants sections r = 0 25 and is described by following equation (equation 3.1). 2 r N-I 2 w(n)= I (3.1) 2 2 _2Jr n-I_Jr) ] 2 r r N-I 2 22

PAGE 34

4. Methods and Procedure 4.1 Preparation Eight adult cats (3.347 0.51 kg) were anesthetized with alpha chloralose (60mg/kg), and used in a protocol approved by the Institutional Animal Care and Use Committee (IACUC). Alpha chloralose does not inhibit the reflexive muscle activity. The skin over the spine was dissected from the thoracic level to the sacral level and then retracted to expose the intact dorsolumbar fascia. An S-shaped stainless steel hook was inserted around the middle of the supraspinous ligament between L4 and L5. The preparation was then placed prone in a rigid stainless steel frame with external fixations on the posterior proce sses of L1 and L 7 which allowed the isolation of the lumbar level from thoracic and sacral levels and avoided mechanical interaction. The frame s position was then adjusted to achieve a vertical alignment of the S shaped hook with the actuator of a Bionix 858 Material Testing System. A gauze pad was soaked with sterile water and placed over the incision area to prevent the exposed skin from drying. Then, Saline fluid was regulated by a Flo-Gard Volumetric Infusion Pump to release 1 Oml/kg/hour through an Intravenous ( IV) line to prevent the tissue from dehydrating. The Arterial oxygen saturation (SP02) and the pulse rate were 23

PAGE 35

monitored by a Rad-5 Pulse Oximeter through a non invasive sensor placed on the preparation s tongue. Ventilation was given by a Harvard Inspira Advanced Safety Ventilator. The temperature was maintained by a Gaymar heating pad and monitored by a WellchAllyn thermometer. Finally records of the temperature SP02 and the pulse rate were taken every 15 minutes throughout the experiment. A sc hematic of the experiment is shown in figure 4.1. B Load Cell Figure 4.1 Schematic of the experimental setup showing the external fixations placed on L1 and L 7 and the hook applied to the L4 / 5 ligament during rest period (a) and during static loading period (b). 24

PAGE 36

4.2 Instrumentation Six pairs of stainless steel wire EMG electrodes (Figure 4.2) were inserted 6-8 mm to the right of the midline into the multifidus muscles ofL 112, L2/ 3 L3 / 4 L4/ 5 L5/ 6 and L6/7. A ground electrode was inserted in the gluteus muscle. Each electrode pair constituted the input to a differential EMG amplifier with a II O-dB common mode rejection ratio gain capability of up to 200 000 and a band pass filter in the range of20-500 Hz. The EMG response from all six channels was monitored on oscilloscopes recorded and stored into a computer at a sampling rate of 1000 Hz. The signal sampling rate was chosen based on the Nyquist theorem (e g 2 x highest frequency ofthe filter) to avoid aliasing. The S shaped stainless steel hook was then connected to the actuator of the Bionix 858 Material Testing System (MTS Inc. St. Paul MN) The load was applied by the vertical actuator via a computer controlled loading system Vertical displacement of the actuator and the load cell output incorporated in it were also sampled into the computer all along with the EMG data. An actual picture ofthe experiment is shown in figure 4.3 25

PAGE 37

Figure 4.2 Stainle s s steel wire electrodes Figure 4.3 Actual picture of the preparation showing the e x ternal fix ators the electrodes and the S shaped hook inserted in the L4/5 supraspinous ligament. 4.3 Experimental Protocol Initially three sinusoidal cycles at a frequency of 0.25 Hz and 40-N of vertical load were applied to the lumbar spine through the actuator of the Bionix 858 each followed by a 10 minutes rest period. The 10 minutes rest periods were included to allow recovery of any tension-relaxation that may have developed in the viscoelastic structures [10 49 50]. The 40N load is considered to be a 26

PAGE 38

moderate load that would elicit a moderate lumbar flexion [26]. Then the preparation was subjected to a static load of 40-N for 10 minutes. Next, the load was released for a 10 minutes rest period This last process was applied 6 times for a total of 2 hours (6x 1 0: 1 0). The six working periods were followed b y a 7 hours recovery period (rest). During this recovery period single 0.25 Hz (4second) sinusoidal cycle tests at 40-N were applied to assess vertical displacement Tension and EMG. Tests were applied after 10 20, and 30 minutes of rest and every hour thereafter (Figure 4.4) Pre-static loading f = 0 25 Hz P1.P2,P3 L1,L2,, LN R1, R2 .. RN 1 N T1, T2, .. Tg ...... ... nAAA.A.AAAA.A Static loading period Pre-static loading (0 25 Hz, 40 N) Static load (10 min 40 N) Rest (10 min no load) Number of static load repetitions 7 hour recovery period Single test cycle during recovery period (0 25 Hz 40 N) Figure 4.4 Schematic of the protocol that was followed in the experiment with N=6 repetitions. 27

PAGE 39

4.4 Data Analysis EMG from all seven lumbar levels displacement and tension data were recorded and stored as sixteen seconds intervals in the computer. Then a high pass filter with a cut-off frequency of20 Hz was applied to the EMG data to prevent any movement artifact that is usually a brief electrode shift caused by muscle contractions Next, the displacement and ten s ion data were subsequently filtered through a low pass filter with a cut-off frequency of 5 Hz. 4.4.1 Displacement NNZs and Tension NNZs In each of the single cycle tests the first 500 msec of each EMG record which was prior to the initiation of the tension were used as a point of reference for baseline signal noise. After this period, the first point along the signal to exceed five times the baseline level was denoted as th e EMG initiation threshold in this channel. The corresponding tension and displacement values during the stretch phase of this cycle were designated as Tension and Displacement Neuromuscular Neutral Zones (Onset). Figure 4.5 is a plot of the hysteresis that shows the tension versus displacement curve and where EMG was present on the curve in boldface. Similarly the last EMG value to exceed five times the baseline noise was denoted as the EMG cessation threshold in this channel and its 28

PAGE 40

6. Analysis The major finding of this research consisted of the fact that a moderate static load of the lumbar spine causes laxity and loss of reflexive stabilizing forces from the multifidus muscles resulting in a large risk of spine injury in the first two hours of rest after a sequence of six sessions of 10 min of 1 : I work to rest ratios (for a total of 2-h). This risk ends during the fourth hour of rest when the ligamento-muscular reflex provides the lumbar spine with a compensatory muscle activity that takes over stability function by increasing the muscular activity via a feedback loop. This mechanism serves well to preserve lumbar stability and compensates for the laxity developing in the lumbar viscoelastic tissues. 6.1 Spine Stability Ligaments have a very small contribution toward maintaining the spine's stability [18-21] whereas the major part is achieved by active multifidus muscle forces generated by the ligamento-muscular reflexes [26, 27, 33, 52, 53]. This reflex arc is triggered by load applications to the supraspinous ligament which excites its sensory mechanoreceptors and then activates the multifidus muscles at the level of deformation as well as one level below or above [26]. Previous studies showed that prolonged flexion of the lumbar spine results in tension53

PAGE 41

relaxation and laxity of its passive viscoelastic structures [13 54]. The laxity and creep induced in the viscoelastic tissues of the spine desensitize the mechanoreceptors resulting in a diminished reflexive muscular activity and allowing full exposure to instability and injury [13]. It was also shown that the muscles seem to compensate for the loss of tension in the lumbar viscoelastic tissues by an increased EMG activity later in the recovery period [12 27 50 55]. 6.2 Motor Unit Recruitment A stronger reflexive muscle activity can be established by either increased firing rate by the same active motor units or by motor unit recruitment (e.g. activating larger motor units) at the same firing rate or by both Increase in the conduction velocity leads to an increase in the median frequency, indicating newly recruited large motor unit. It has been proven in previous studies that the mean power frequency of the EMG spectrum and the muscle fiber conduction velocity are linearly related [46 56]. It also has been shown that there is a linear relationship between the average conduction velocity and the median frequency [46]. Then Previous studies demonstrated that the frequency component of the EMG power spectrum can be used as an indicator of motor unit recruitment [57]. Summing up all these facts it was shown that an increase in the median frequency 54

PAGE 42

indicates an orderly recruitment of the motor units [10, 45]. And that means that an increase in the average conduction velocity indicates activation of larger and faster motor units and therefore an orderly recruitment of motor units that leads to an increase in the median frequency. In this study no change in the median frequency was observed; therefore no significant change in the motor unit recruitment occurred. It can be concluded therefore that the increase in the EMG amplitude is due to increase in the firing rate of the already active motor units. 6.3 Creep Previous studies stated that the viscoelastic creep loading increases the laxity of the intervertebral joint resulting in the injury to the spine and low pack disorder [8-10,33]. In this study the loading period developed laxity in the ligaments and creep in the lumbar spine structure. This resulted in the combination of enlarged displacement neuromuscular neutral zones enlarged tension neuromuscular neutral zones and decrea s ed peak MA V eliciting a delayed reflexive muscle activity leaving the spine unprotected. Therefore in the first I to 2 hours of recovery low muscular activity was present and was triggered only after reaching high displacement and tension values. Even after the 7 hours of recovery significant creep in the passive viscoela s tic structures was still 55

PAGE 43

present. The recovery period was not sufficient enough for full recovery of the creep in the viscoelastic tissues unlike what was demonstrated in the latest studies that involved a similar load / rest schedule but with different load magnitudes or with cyclic loading It was shown by a previous report that 40-N static loading did not result in dela y ed hyperexcitabilit y, wherea s onl y high static loads (60-N) displayed delayed hyperexcitability and an inflammatory response [14]. Also it was proven that moderate (40-N) to high (60 -N ) magnitudes of cyclic loading could trigger a delayed hyperexcitability [11]. Even though cyclic loading seems to be more challenging to the viscoelastic tissues than static loading [11], a recent study showed that the creep disappeared at the end of the recovery period for moderate cyclic loading [58] wherea s it was still present for static loading in this study. That can be explained in a way that moderate 40-N cyclic loading was riskier and more damaging to the viscoelastic tissues which caused increased muscular activity and stiffness in the spine that prevented creep from rising in order to avoid further damage to the spine. Comparing with this study the moderate 40 N static loading wasn t intense or damaging enough and didn t require increased muscular stiffness in the lumbar spine 56

PAGE 44

6.4 Compensation Mechanism In this study the presence of the compensation mechanism was sufficiently indicated by decreased Tension neuromuscular neutral zones and increased EMG peak MA V. This confirms that the muscular activity was triggered earlier at low tension values and with higher EMG amplitude to compensate for the laxity in the ligaments in order to prevent damages in the spine This compensatory muscle reaction was achieved by increased firing rate of active motor units with no significant change in the median frequency; e.g. no change in the motor unit recruitment. Static loading with 40-N magnitude was shown to be a moderate loading that didn t cause any damage to the spine and didn t display any hypexcitability [14]. Thus it was probably an increase in the firing rate of the multifidus muscles that generated higher EMG amplitude through temporal and spatial summation, and that activation of larger motor units wasn t necessary for the prevailing conditions. Comparing with a recent report involving 40-N cyclic loading that displayed an inflammatory response resulting in a hyperexcitability the compensatory mechanism was accomplished by higher EMG amplitude and by an increased median frequency indicating activation of larger and faster motor units [58]. This compensatory mechanism presented by an increased electromyogram peak MA V might be a new motor control mechanism 57

PAGE 45

that wasn't exposed or discussed before and that separates itself from the actual hyperexcitability phenomenon experienced in previous reports that is usually defined as an inflammatory response to the micro-damage of the viscoelastic tissues. 6.S Humans Vs Cats It's important to recall that the data obtained in the experiments was derived from the feline model (quadruped). One needs to assure that the same processes occur in humans (biped) to confirm the validity of the final results. A recent study showed biomechanical similarity of the spine between quadruped and bipeds confirming that the quadruped can be a valuable animal model for spine research [59]. It was also shown that the neuromuscular system of the feline and the human are similar [60]. The collagen structures of the ligaments, facet capsules, and discs for humans are very similar to cats and also contain mechanoreceptors [61, 62]. Previous studies also proved that the ligamento-muscular reflex exists also in humans and provided support to the similarity of the responses between humans and felines [12 ,26,63-66]. Even though the difference in size between cats and humans is evident the general pattern of responses is expected to be the same but with differences in the 58

PAGE 46

loading periods resting periods number of repetitions time constants in the mathematical models and loading magnitude ; where the loading magnitude would be addressed to as low moderate and high loads for humans. 59

PAGE 47

7. Conclusion This study provided experimental evidence showing the effect of a total of 2 hours moderate static loading and rest with a work to rest ratio of I: I on the neuromuscular neutral zones revealing more information about spine instability in a work environment. It also displayed a biomechanical and physiological pattern for the normalized peak MA V and the median frequency in order to track the reflexive multifidus muscles activity. Some scaling of the data derived from feline quadruped still needs to be done in order to apply the results to humans. As a summary, the combination of information gathered from this investigation leads to the following conclusions: I. Moderate static loads applied to the lumbar spine results in corruption to the ligamento-muscular reflex due to creep in the lumbar spine and shows enlarged neuromuscular neutral zones placing the spine under full exposure to instability and injury in the first couple of hours of rest. 2 A Compensation mechanism takes over in the later hours of rest resulting in narrow neuromuscular neutral zones muscle stiffness and increased muscle activity to compensate for the laxity in the ligaments and prevent spine Injury. 60

PAGE 48

3 The compensatory muscle activity is probably attained by a higher EMG firing rate that results in higher EMG amplitude, but with no variations in the motor unit recruitment. 4. The compensatory mechanism is probably a new motor control mechanism that hasn t been investigated yet. 5. Seven hours ofrest isn t enough for creep recovery in the viscoelastic tissues of the spine. So we can conclude that workers should be watchful in the first couple of hours ofrest after a long day of work periods to prevent spine injury since the spine will be susceptible to instability. Finally an extension of this research would be to further investigate the neuromuscular neutral zones behavior for different load magnitudes and for cyclic loading, and to identify the newly mentioned compensatory mechanism that occurs in the reco v ery period 61

PAGE 49

References I. NINDS (2003) Low Back Pain Fact Sheet. Volume 2. N rOSH National institute for Occupational Safety and Health : A Critical Review of Epidemiologic Evidence for Work-Related Musculoskeletal Disorders of the Neck, Upper Extremity, and Low Back. 1997 Musculoskeletal Disorders and Workplace Factors. 3. NRC, National Research Council: Musculoskeletal Disorders and the Workplace: Low Back and Upper Extremities. 2001: Institue of Medicine. 4. McGi II, S.M., The biomechanics of low back injury: implications on current practice in industry and the clinic J Biomech 1997.30(5): p. 465-75. 5. Andersson G.B., Epidemiologic aspects on low-back pain in industry. Spine 1981. 6(1): p 53-60 6 Damkot, O.K. et aI., The relationship between work history work environment and low-back pain in men. Spine, 1984. 9(4): p. 395-9. 7. Anderson lA., .0. Otun and BJ. Sweetman, Occupational hazards and low back pain. Rev Environ Health 1987.7(1-2): p. 121-60. 8. Adams, M.A. and P. Dolan, Time-dependent changes in the lumbar spine's resistance to bending. Clin Biomech (Bristol Avon), 1996 11(4): p 194200. 9. Adams, M.A., et aI., Diurnal changes in spinal mechanics and their clinical significance. J Bone Joint Surg Br 1990. 72(2): p. 266-70. 10. Gedal ia U., et aI., Biomechanics of increased exposure to lumbar injury caused by cyclic loading. Part 2. Recovery of reflexive muscular stability with rest. Spine 1999.24(23): p 2461-7. 11. Le, P. et aI., Cyclic load magnitude is a risk factor for a cumulative lower back disorder. J Occup Environ Med, 2007. 49(4): p. 375-87. 62

PAGE 50

12. Solomonow, M et aI., Flexion-relaxation response to static lumbar flexion in males andfemales. Clin Biomech (Bristol Avon), 2003.18(4) : p. 273-9. 13. Solomonow, M., et aI., Biomechanics of increased exposure to lumbar injury caused by cyclic loading: Part 1. Loss ofreflexive muscular stabilization. Spine, 1999 24 (23): p. 2426-34. 14. Sbriccoli, P., et aI., Static load magnitude is a riskfactor in the development of cumulative low back disorder. Muscle Nerve, 2004. 29(2): p. 300-8. 15. LaBry, R., et aI., Longer static flexion duration elicits a neuromuscular disorder in the lumbar spi ne. J Appl Physiol 2004. 96 (5) : p. 2005-15. 16. Sbriccoli, P., et aI., Static load repetition is a risk factor in the development of lumbar cumulative musculoskeletal disorder. Spine, 2004. 29(23): p. 2643-53. 17. Courville A., et aI., Short rest periods after static lumbar flexion are a risk factor for cumulative low back disorder. J Electromyogr Kinesiol 2005. 15(1): p. 37-52. 18. Ab umi K., et aI. Biomechanical evaluation of lumbar spinal stability after gradedfacetectomies Spine 1990. 15(11) : p. 1142-7. 19. Anderson C.K et aI., A biomechanical model of the lumbosacral joint during lifting activities. J Biomech 1 985 18(8) : p 571-84. 20. Posner I W.A., Edwards W et aI., A biomechanical analysis of th e clinical stability of the lumbar and lumbo sac ral spine 1982. 21. White A P .M., C lini cal Biomechanical of the Spine 1978. 22. McGill S.M. and R.W. Norman, Partitioning of the L4-L5 dynamic moment into disc ligamentous, and mu scular components during lifting. Spine 1986. 11(7 ) : p. 666-78. 23. Crisco, J.1., 3rd and M.M. Panjabi The intersegmental and multisegmental muscles of the lumbar spine. A biomechanical model comparing lateral s tabilizing potential. Spine, 1991. 16(7): p 793-9. 63

PAGE 51

24. Kaigle A.M. S.H. Holm and T.H. Hansson Experimental instability in the lumbar spine. Spine 1995.20(4): p. 421-30. 25. Wilke, H.J. et aI. Stability increase of the lumbar spin e with different muscle groups A biomechanical in vitro s tud y. Spine 1995 20(2): p. 192-8. 26 Solomonow M. et aI., The ligamento-muscular s tabilizing system of th e spine Spine 1998. 23(23): p. 2552-62 27. Stubbs M., et aI. Ligamento-muscular protectiv e reflex in the lumbar spine ofthefeline. J Electromyogr Kinesiol 1998 8(4): p. 197-204. 28. Panjabi M.M., Clinical spinal ins tability and low back pain. J Electromyogr Kinesiol 2003.13(4): p. 371-9. 29. Elizabeth Eversull B.S. et aI., Ne uromuscular neutral z one s sen s itivity to lumbar di s pla c em ent rate. Clin Biomech (Bristol A von) 200 I. 16(2 ) : p. 102-13. 30. Solomonow M. et aI., N euromuscular neutral z one s a ssociated with viscoelastic hysteresis during cyclic lumbar flexion. Spine 200 I. 26(14): p. E314-24. 31. Hagood S. et aI., The effect of joint velocity on the contribution of the antagoni s t mu sc ulature to kne e s tiffne s s and laxity. Am J Sports Med 1990. 18(2): p. 182-7. 32. Tortora GJ., Prin c iples of Human Anatom y. Ninth ed 2002. 33. Bogduk N., C lini c al Anatom y of the Lumbar Spin e and Sacrum Fourth ed 2005. 34 Kumar S., Biom ec hanic s in Er g onomic s 1999 35. Woo S., Structure and Function of Ligaments and Tendon s Basic orthopaedic Biomechanics 2004. 36. Solomonow M. Ligaments: a s our ce of work r e lated musculo s k e letal disorder s J Electromyogr Kinesiol 2004 14(1) : p. 49-60. 64

PAGE 52

37 Adams, M.A., W.C. Hutton and J.R. Stott, The resistance tojlexion of the lumbar intervertebral joint. Spine, 1980.5(3): p. 245-53. 38. Solomonow, M. and M. Krogsgaard, Sensorimotor control of knee stability. A review. Scand J Med Sci Sports, 2001. 11(2): p. 64-80. 39. Rhalmi S., et aI., Immunohistochemical study of nerves in lumbar spine ligaments. Spine 1993. 18(2): p. 264-7. 40. Gray, H., Gray's Anatomy of the Human Body. 41. Kim, c., The Multifidus Muscle is the Strongest Stabilizer of the Lumbar Spine The Spine, 2007. 7(5) : p. 76S. 42. Panjabi M M., The stabilizing system of the spine. Part 1. Function, dysfunction, adaptation and enhancement. J Spinal Disord, 1992.5(4): p. 383-9; discussion 397. 43. Reeves N.P., K.S. Narendra and J. Cholewicki Spine stability: the six blind men and the elephant. Clin Biomech (Bristol, Avon), 2007.22(3): p. 266-74. 44. Rangayyan R.M. Biomedical signal analysis: A case-study approach. IEEE Press, 2002. 45. Solomonow M et aI., Electromyogram power spectra frequencies associated with motor unit recruitment strategies J Appl Physiol 1990 68(3) : p 1177-85. 46. Stu len F.B. and C.J. DeLuca Frequency parameters of the myoelectric signa l as a measure of muscle conduction velocity. IEEE Trans Biomed Eng, 1981. 28(7): p. 515-23. 47. Manolakis, D., V Ingle and S. Kogon Statistical and Adaptive Signal Processing: Spectral Estimation Signal Modeling Adaptive Filtering and Array Processing 2005. 48. Oppenheim, A., and Schafer R., Digital Signal Processing. 1975, Englewood Cliffs, NJ: Prentice-Hall. 65

PAGE 53

49. Jackson, M., et aI., Multifidus EMG and tension-relaxation recovery after prolonged static lumbar flexion. Spine, 2001. 26(7) : p. 715-23. 50. Solomonow M., et aI., Biexponential recovery model of lumbar viscoelastic laxity and reflexive muscular activity after prolonged cyclic loading. Clin Biomech (Bristo l Avon), 2000. 15 (3): p. 167-75 51. Baratta R. V., et aI., Methods to reduce the variability of EMG power spectrum estimates. J Electromyogr Kinesiol, 1998.8(5) : p. 279-85 52. Indahl A., et aI., Electromyographic response of the porcine multifidus musculature after nerve stimulation. Spine, 1995.20(24): p. 2652-8. 53. Indahl A., et aI., Interaction between the porcine lumbar intervertebral disc zygapophysialjoints, and paraspinal muscles Spine, 1997.22(24) : p. 283440. 54. Williams, M. et aI., Multifidus spasms elicited by prolonged lumbar flexion. Spine, 2000 25(22): p. 2916-24. 55. Solomonow M., et aI., Muscular dysfunction elicited by creep of lumbar viscoelastic tissue. J Electromyogr Kinesiol 2003 13(4): p. 381-96. 56. Arendt-Nielsen, L. and K.R. Mills The relationship benveen mean power frequency of the EMG spectrum and muscle fibre conduction velocity Electroencephalogr Clin Neurophysiol, 1985.60(2): p. 130-4 57. Moritani, T. and M. Muro, Motor unit activity and surface electromyogram power spectrum during increasingforce of contraction. Eur J Appl Physiol Occup Physiol, 1987.56(3) : p. 260-5. 58 Solomonow, D., Neuromuscular Neutra l Zones R esponse due to Cyclic Lumbar Flexion. (In Press) 2008. 59. Smit, T.H. The use of a quadruped as an in vivo modelfor the study of the spine biomechanical considerations. Eur Spine J, 2002. 11(2) : p. 137-44. 60. Field, H., Taylor, M, An Atlas of Cat Anatomy. 1992. 66

PAGE 54

61. Hirsch, C. B.E. Ingelmark, and M. Miller The anatomical basis for low back pain. Studies on the presence of sensor y nerve endings in ligamentous capsular and intervertebral di sc structures in the human lumbar s pine. Acta Orthop Scand 1963.33: p 1-17. 62. Yahia L.H. N Newman, and C.H Rivard N eurohistology of lumbar spine ligaments. Acta Orthop Scand 1988. 59(5): p. 508-12 63. Chu, D., et aI., Neuromuscular di s order in r es pon se to ant e rior c ru c iat e ligament cr e ep Clin Biomech (Bristol Avon) 2003.18(3): p. 222-30. 64. Olson M. M. Solomonow and L. Li, Fle x ion-relaxation re s pon s e to gravity J Biomech 2006. 39(14): p. 2545-54. 65 Olson M.W. L. Li, and M. Solomonow Flexion-relaxation respon s e to cyclic lumbar flexion. Clin Biomech (Bristol Avon) 2004. 19(8) : p. 769 76 66. Sbriccoli P et aI. Neuromuscular response to c y clic loading of the anterior cruciate ligament Am J Sports Med 2005. 33(4): p. 543-51. 67

PAGE 55

corresponding tension and displacement values during the relaxation phase were designated as Tension and Displacement Neuromuscular Neutral Zones (Offset) Figure 5.2 shows a typical record of tension displacement and EMG from channels L3 / 4 L4 / 5 and L5/ 6 for a single cycle test which helps to understand better how displacement NNZs and tension NNZs were computed The arrows in the plot represent the EMG initiation and cessation thresholds. c o 40 iii 20 c Q) t10 01 31 2007 (0.25 Hz, 40N) o 3 6 9 12 Displacement (mm) Figure 4.5 Plot of the hysteresis that shows the tension versus displacement curve of a single cycle at O.25-Hz frequency and 40-N amplitude The black circle shows EMG initiation at the stretch phase The Boldface on the curve designates the period where EMG was present and recorded throughout the cycle. The white circle represents EMG cessation in the release phase. The unBold part of the curve represents the NNZs. 29

PAGE 56

4.4.2 Mean Creep In each of the single cycle tests for each of the channels, the peak displacement was detected. Then for the first three single cycle tests (pre-static loading) the displacement values were averaged and the mean (SO) was denoted as the displacement baseline value (0 percent creep). Following, the percentage creep was calculated using the following equation: D-D Creep = ) x 1 00% D) Where D is the displacement value at each specific single cycle test and D J is the displacement ba se line value. Then, the resulting values of the creep in the recover y period in each of the eight preparati ons were pooled and then the mean (S O) were calculated and plotted as a function of time. 4.4 3 Normalized peak MA V A full wave rectification was applied to the EMG recording of each of the single cycle tests for each of the channels, followed by a 200 points moving average with a forward shift of 10 points. Next, the peak amplitude of the processed EMG recordings was detected and designated as peak MA V. Then, the data for each preparation was normalized with respect to the value of the first 30

PAGE 57

single cycle test in each data set. The normalized peak MAY was computed in order to show the maximal level of activity of the multifidus muscles throughout the experiment. 4.4.4 Median Frequency In each of the single cycle tests for each of the channels the 0.5 sec window of EMG data where maximum tension was present was detected and then zero padded by adding zero points to yield 512 points of data. Next, the EMG data was multiplied by a tukey window [48]. The tukey window was specifically selected since it was the most similar to a rectangular window so it didn t attenuate a significant part of the signal compared with hamming or hanning windows. Then, the power spectral density was computed after applying the fast Fourier transform to the data. The baseline noise spectrum subtraction method [51] was applied to filter out 60 Hz noise and its harmonics from the signal. And that was achieved by calculating the power spectral density of a 0.5 sec window from the baseline noise then subtracting it from the power spectral density of the real EMG computed earlier After that the median frequency was calculated. The median frequency is the frequency that divides the power spectrum in half as shown in equation I. Then the median frequency data was run 31

PAGE 58

through a three point moving average in order to smooth it. It was found that the increasing average conduction velocity during motor unit recruitment is the major contributor to variations in the electromyogram median frequency [45]. So the changes in the median frequency could be an index to identify if the motor unit recruitment throughout the recovery period increased decreased or remained unchanged. = fS(f)df = fS(f)df o MF 2 0 (4.1 ) Where Set) is the power density spectrum and f is the frequency For the first three single cycle tests (pre-static loading) in each of the eight preparations the Displacement NNZs were averaged and the mean (SD) was denoted as the baseline value. Similarly the same process was applied to the Tension NNZs, normalized peak MAY and median frequency. Then the Displacement NNZs Tension NNZs peak MA Y and median frequency of the remaining tests (Recovery period) for each channel in each of the eight preparations were pooled and then the mean (SD ) were calculated and plotted as a function of time. 32

PAGE 59

4.4.5 Mean of Percentage Change from Baseline After assigning the baseline values the percentage change from baseline values throughout the time was calculated in each of the eight preparations for each of Displacement NNZs (onset and offset) Tension NNZs (onset and offset) and normal ized peak MA V. The average of the percentages for each category of data for the total of all the preparations was then computed and plotted versus time. 4.4 6 Statistics A two-way analysis of variance CANOVA) with repeated measures where time and vertebrae were set as independent variables and where displacement onset displacement offset tension onset tension offset normalized peak MA V and median frequency were set as dependent variable s was applied to all the previous processed data to show the significance of the variations and to evaluate the effects of loading on the recovery period. Then a one-way analysis of variance CANOVA) with repeated measures was applied to the creep data with time as an independent variable. The Student's T test was used to provide the details about the significance of variations. A normality test was first run to assure a normal distribution of the data. Significance was set at p
PAGE 60

4.5 Models Displacement NNZs at onset and offset Tension NNZs at onset and offset and normalized peak MA V data in the 7 hours recovery period from all three channels were fitted to models that represent their pattern of variation in the recovery period and to compare it with the pre-static loading period Modeling consisted of exponential equations plots using Marquardt-Levenberg algorithm Exponential functions were used since they represent the classical characteristics of viscoelastic materials. r2 values were then derived to verify the accuracy of the best-fit models. The model structure for the Displacement NNZs at onset and offset in the recovery period for each of the channels consisted of a single exponential function that can be described in the following equation: DNNZ(t) = D + D + (D D )e-(I-TR ) IT\ o R L /I (4.2) Where D o is the baseline value (pre-static loading period) DR is the residual value D L is the first recorded value in the recovery period and (DL -D R ) is the decay amplitude of the exponential t is the time LR is the time when the recovery period begins and L 1 is the time constant. 34

PAGE 61

The Tension neuromuscular neutral zones followed a bi-exponential model as shown in the equation below: (4.3) Where T o is the steady-state ampl itude hand T M are the ampl itudes of the exponential terms and T 2 and T3 are respectively their correspondent time constants. The term T M e -(tr R ) / T represents the decreasing exponential. Similarly exponential functions were utilized in the modeling ofthe normalized peak MA V during the 7 hours recovery period The model took the form shown in the following equation: Where P o is the steady-state amplitude PL, PM, and PH are the amplitudes of the exponential terms, T 4 and TS and T6 are respectively their correspondent time constants and Td is the time delay where the dela yed hyperexcitability starts. The first term PL e -(t-rR) 1 r, describes the exponential decay at the beginning of the recovery period that reaches its minimum at the first hour. The term P M (I -e-(t-rR ) / r s ) describes the steady-state recover y compone nt The term 35

PAGE 62

(t -r d )PH e -(t-t" ) / t6 represents the delayed transient hyperexcitability This term starts affecting the equation only after the recovery exceeds the time delay rd36

PAGE 63

5. Results In general the recorded data throughout the experiment from all six channels demonstrated an exponential decay of the EMG activity during the six 10 min load periods with a partial recovery during the six 10 min rest periods in between. The corresponding displacement during the 10 min loads showed an exponential increase caused by the laxity of the ligaments During the recovery period EMG activity increased gradually, while the displacement decreased exponentially to reach a near full recovery Figure 5.1 shows a typical EMG recording from channels L3/ 4 L4 / 5 and L5/ 6 as well as tension and displacement. As emphasized by a red circle during the third static loading period at channel L3 / 4 large amplitude spasms appeared with an EMG exponential decrease background. Spasms are sudden powerful contractions or stiffening of the muscles that occur as a response to tissue damage and are usually indicated by large EMG amplitudes 37

PAGE 64

11 03 2006 {Sine 0 25 H z, sta tic, 4()N, 6x1(HO } o 10 20 30 4 0 50 60 70 80 90 100 1 10 120 130 150 170 200 260 320 360 SOD 560 l me(m i n ) Figure 5 1 EMG responses from channels L3/4 L4 / S and LS/ 6 along with the displacement and tension channels for a single preparation when subjected the e x perimental protocol. And Figure S.2 shows a typical record of tension displacement and EMG from channel L4 / S for a single cycle test. The superimposed curve in red represents the MA V of the EMG at this specific lumbar level. The Arrows in the top plot represent the EMG initiation and cessation thresholds. A projection of these arrows towards displacement s and tension s graphs reveals their corresponding Displacement and Tension NNZs (onset and offset). 38

PAGE 65

11 1 7 2006 (O.25H z, 40N) f: t-I __ __ --'" __ ---+-1.::: 1 2 E NNZ : DNNZ off E 8: : 8 --. Q. : : .!!1 4 : 4 o : o 0 Z 60 1 nlN'Zotl 60 :/ 4 0 c 20 [ : 20 O : 0 3 5 T ime ( s e c ) 6 7 Figure 5.2 EMG response s from channels L3/ 4 L4 / 5 and L5/ 6 along with the displacement and tension channels for a single cycle test (0.25-Hz 40-N). F igure 5.3 shows the behavior of the hysteresis for a s pecific preparation at the pre-static loading period (plot 0) in comparison with it s status after the first hour of recovery (plot 1), after the second hour of recovery (plot 2) and after the six th hour of recovery (plot 6). Maximum displacement took a value of9.1 at the prestatic loading period (plot 0) and then displayed an increase to a maximum displacement value of 13.52 at the first hour of recovery (plot 1) which shows the la x ity and the creep in the ligament s caused by the static loading period. Then a 39

PAGE 66

decrease in the maximum displacement started at the second hour of recovery (plot 2) and took a value of 12. 67 and th e n resulted in a maximum displacement value of 11.42 at the sixth hour of recovery (plot 6). 01 31 2007 (0.25 Hz, 40N) 0 0 2 0 6 40 40 3 0 c 0 20 2 0 2 0 Q) ..... : f-1 0 10 1 0 o 0 o 2 4 6 8 1 0 1 2 1 4 0 2 4 6 8 1 0 1 2 14 0 2 4 6 8 10 1 2 1 4 Displacement (mm) Displacement (mm) Displacement (mm) Figure 5.3 Plots of the tension versus displacement (hysteresis) for a single preparation at the pre-static loading period after 1 hour of recovery after 2 hour s of recovery, and after 6 hours of recovery 5.1 Displacement NNZs Figure 5.4 shows the mean Displacement NNZs in both the stretch and relaxation phases for all three lumbar levels L3/ 4 L4 / S and LS/ 6. The black circles indicate the mean Displacement NNZs at Onset throughout the time along with their standard deviation. The white circles represent the mean Displacement 40

PAGE 67

NNZs at Offset along with their standard deviation The first value s at time zero in each of the graphs indicate the baseline values that were detected in the pre static loading period. The superimposed curves in red represent the exponential models that were fitted to the data using Marquardt-Levenberg algorithm. The Student s T test showed significant changes of the Displacement onset and offset NNZs over time The Displacement NNZs at onset started at baseline value s of 3.181 ( 2.3) 2.S75( 1.75) and 3 1 67( 2.24) for L3/ 4 L4 / S and LS/ 6 and displayed significantly increasing values to 8.33 ( 2 04) 7.938( 2.24) and 8.761 ( 1.9S) respectively immediately after the loading period which was equivalent to an average of 317.44 percent increase from baseline value as shown in Figure S .S. Then an exponential decrease in the mean Displacement onset NNZs occurred and reached final values of 3 263 ( 1.78) 2.868( 1.93) and 3 026( 2.17) after the 7 hours recovery period at the respective lev els and that was equivalent to an average of 34 percent increase from ba s eline (Figure S.S). 41

PAGE 68

l3/4 L4i5 L5 / 6 2:0 20 20 E .s 15 1 S 1S N Z Z em 1 0 OJ E OJ 0 co 5 5 c5 {) 0 0 0 10 0 200 300 400 SOO 600 0 1 00 200 300< 400 SOO 600 0 100 200 300 400 SOO 600 Time (min) Time (min) Time ( m i n ) Figure 5.4 Mean displacement neuromuscular neutral zones for N = 8 preparations For the mean Displacement NNZs at offset, initial values of7.871(.33) 7 72(.33) and 7.727(.38) took place respectively at lumbar levels L3/ 4 L4/5 and L5/6. After the loading period those values displayed a significant increase to 13.781(.82), 13.26(.42), and 13.57(.56) which was equivalent to an average of85. 9 percent increase from baseline value. An exponential decrease in the mean Displacement NNZs at offset was then observed and reached values of7.618(.64) 7.071(.5) and 7.52(.86) after the 7 hours recovery period for the respective levels and that was equivalent to an average of -2.575 percent difference from baseline value. 42

PAGE 69

Q) .s: Qj '" '" co 300 E .g c: '" .c: (.) '" C 0 c: '" Q) :2' DNNZ N=B Stret c h o Relaxat i o n change . ... o 2 4 6 Time (hr) B Figure 5.5 Mean of percentage change from baseline for displacement NNZs (onset and offset) The star symbol represents a statistically significant change from baseline values. 5.2 Tension NNZs A log based 10 transformation was necessary to apply to the Tension NNZs at Onset data in order to achieve normal dist ribution Statistics then showed significant variations of Tension NNZs at onset and offset over time. Figure 5.6 shows the mean Tension NNZs in both the stretch and relaxation phases for all three lumbar levels L3/4 L4/5 and L5/ 6 The black circles indicate the mean Tension NNZs at onset over time along with their standard deviation. The white circles represent the mean Tension NNZs at offset along with their standard deviation. 43

PAGE 70

1.314 l4/ 5 L516 50 50 1----;:===::::;,50 Z N30 Z Z 5 20 "iii c 10 o 30 20 10 o a 1 00 200 300 40 0 5 00 6 00 0 100 20 0 300 40 0 500 600 0 1 00 200 300 4 00 SOO 600 T i m e ( m i n) Time (mi n ) T ime ( m i n ) Figure 5.6 Mean tension neuromuscular neutral zones for N=8 preparations Initial baseline values of the Tension NNZs onset started at 9.864(S.76), 7.849( 81), and 9.186(S.23) for L3/4 L4/S, and LS/6 then displayed insignificantly increasing values to 1O.S16(S.07), 9.S81(S.03), and 11.742(S.36) respectively immediately after the loading period which was equivalent to an average of 42.98 percent increase from baseline value (Figure S.7). A further significant increase took place afterwards and reached values of 10.961(.97), 12.067(.39) and 14.842(.64) respectively during the fITst hour of recovery which was equivalent to an average of 60.96 percent change from baseline (Figure S.7). Following an exponential decrease in the mean Tension NNZs at onset occurred and took significantly lower values than baseline 44

PAGE 71

during the third hour of recovery. Tho s e values were 9.545( 5.68), 5 .791 (.04) and 5.822(.0 I) and that was equivalent to an average of 9 67 percent difference from baseline value (F i gure 5 .7 ) A further decrease resulted in final values of 4.755(.23) 3.623(.33) and 4.033(.02) after the 7 hours recovery period at the respective l evels which was equivalent to an average of -51.61 percent difference f r om baseline as shown in Figure 5 .7. Q) c Qi 80 C/) ro III E 6 0 e Q) 40 0) c ro 20 .c. U Q) 0 0) ro 2 0 u Cii -40 Cl. c -6 0 ::2; 0 2 TNNZ 4 N = 8 Str etch o Relax ation Signi f ican t change 6 T i me (hr ) 8 Figure 5.7 Mean of percentage change from baseline for tension NNZs (onset and offset). T he star symbo l repre sents a statistically significant change from baseline values The Tension NNZs at offset started at baseline values of23.415(. 61) 21.314( 33) and 21. 888( 10.05) for L3/ 4 L4 / 5 and L5 / 6 and displayed significantly increasing va lue s to 27 .89 5( 8.83) 23.025( 8.4) and 25.48( 7.15 ) 45

PAGE 72

respectively immediately after the loading period which was equivalent to an average of29.83 percent increase from baseline value (Figure 5 7 ) A further increase took place afterwards and reached values of 26.03(.28) 27.86(IOA) and 28.463( 10.94) respectively during the first hour of recovery which was equivalent to an average of 35.6 percent increase from baseline (Figure 5 7). Then an exponential decrease in the mean Tension NNZs at offset occurred and took significantly lower values than baseline during the fourth hour of recovery. Those values were 19.858(.34) 18.207( 9.62) and 16.805(.2) and that was equivalent to an average of -21. 05 percent difference from baseline value (Figure 5 7) A further decrease resulted in final values of 12. 727(A7) 9.276(.86) and IIA95(.66) after the 7 hours recovery period at the respective levels which was equivalent to an average of -48.56 percent change from baseline (Figure 5.7) 5.3 Normalized peak MA V Student s T test s results showed significant changes of the normal ized peak MAY over time. Figure 5 8 shows the mean normalized peak MAY along with their standard deviation for all three lumbar levels L3 / 4, L4 / 5 and L5/ 6 versus the time. Initial baseline values started at I for lumbar levels L3/ 4 L4/ 5 and L5/ 6 then displa y ed insignificantly decreasing values to 0.941 ( 0.58) 46

PAGE 73

O.866(.25) and O. 758( O .5]) respectively after the 2 hours loading period which was equivalent to an average of -14.44 percent difference from baseline value (Figure 5 9). Following a significant exponential decrease in the mean normalized peak MA V occurred and took values ofO.643( O.26) O 672( O.26), and O.648( O.35) during the first hour of recovery and that was equivalent to an average of -33 72 percent change from baseline (Figure 5.9). An exponential increase took place afterwards and was then followed b y a further significant delayed exponential increase that reached final values of 1.535( 1.17) 1 .541 ( O 86) and J.342(O.8) after the 7 hours recovery period at the respective lumbar levels which was equivalent to an average of 47.32 percent increase from baseline (Figure 5.9) Those final values were statistically different from baseline L3I 4 l 4l 5 L516 1 8 .------------, 1.8 .----------r--,I. B .---------, > 1.4 (I} Cl.. -0 1 0 (I} E 0; 0 6 Z N = 8 N=8 1111=8 0 2 .&..-....--........ -..---.----..,....-......---102 .1..-....--........ -..---.--.--......---40 2 .&...-......---.-......... --.--.--......---4 o 1 00 200 300 400 500 600 0 100 200 300 400 SOO 600 o 100 200 300 400 sao 600 Time (min) Time ( m i n ) Time ( min) Figure 5.8 Mean normalized peak MA V for N = 8 preparation s 47

PAGE 74

Q) peak MAV S 80 ,=:==:==,-___ -, Q) I II) m CO 60 E .g Q) 4 0 Ol c m B 2 0 Q) 0 c Q) u iii -20 [L c -4 0 -'---r--..---r--.-----i ::2 o 2 4 6 B Time ( hr ) Figure 5.9 Mean of percentage change from baseline for normalized peak MAV. The star symbol represents a statisticall y significant change from baseline value. 5.4 Mean Creep The initial baseline value was denoted as zero creep. Following the creep demonstrated a significant increase after the 2 hours static loading period to reach a maximum value of 56.687%( 23 33 % ). Then the creep showed a significant decrease throughout the recovery period that reached a final value of 12.21 %( 16.23 %) at the seventh hour and that was still s ignificantl y different from baseline as shown in Figure 5.10. 48

PAGE 75

o 100 200 300 400 500 600 Time(min) Figure 5.10 Mean of the creep for N=8 preparations The star symbol repre se nts a statist icall y sign ific ant change from baseline value 5.5 Median Frequency After running the Student's T test, results did not show any statist icall y significant changes in the median frequency over time Figure S.ll shows the mean of the 3-point averaged median frequency along with their standard deviation versus the time for all three lumbar levels. Initial baseline values for the median frequency started at 209.S4( S2), 209 63(.IS), and 193.19( 14. 37) for lumbar levels L3/4, L4/S, and LSI6, then displayed insignificantly decreasing values to 201.66(.99), I 96. S3(.49) and I 78.9S(.1 9) respectively after the 2 hours loading period Then an increase of the median frequency over time was observed and reached final values of210.44(.93), 202.7S(.71), and 49

PAGE 76

192.38( 14 39) for the r espective level s None of tho s e variation s was statistically significant. The median frequenc y was initiall y computed in order to identif y the motor unit recruitment s tatus throughout the recov e r y period. L3/4 L4/5 L5/6 iJ' 220 22 0 220 c N=8 N=8 N = 8 Ql ::J rlrl 2 10 ........ .... ......... .. . 210 u.. [ I I 1 I ]I I ffi 2 00 200 200 'i5 r 111.1 jllr 1 90 1 90 1 9 0 Cl > ::1 1 60 1 60 160 a. (') 1 170 170 170 '0 ffi 1 60 1 60 1 60 Ql :2 0 100 2 0 0 300 400 500 6 0 0 0 100 2 00 3 00 4 00 50 0 600 0 100 200 300 400 500 60 0 Time (min) Time (min) T i me ( m i n ) Figure 5.11 Mean of 3-point avera g ed median frequency for N = 8 preparations 5.6 Models Best-fit models were superimpo s ed on the e x perimental data. M o del i n g consisted of exponential equations plots using Marquardt-Levenberg al g orithm. In order to v erify the accuracy of those models ,; va lue s were deri v ed The n res ults displayed r2 values rangin g from 0 95 2 4 to 0 9876 for the displacement NNZs model (Table 5.1), from 0.8645 to 0. 9 9 2 7 for the t e n s ion NNZs model ( Table 5.2) and from 0.9175 to 0 9848 for n o rmalized peak MAY model ( Table 5.3). 50

PAGE 77

Time constants were noticed to be higher for Displacement NNZs at offset and Tension NNZs at offset comparing with Displacement NNZs at onset and Tension N Zs at onset respectively which indicates a slower recovery in the relaxation phase. Values of the time constants for Displacement NNZs at onset ranged from 127 to 170 min while they ranged from 214 to 232 min for the Displacement NNZs at offset. Table 5.1 Displacement neuromuscular neutral zones model parameters Onset Offset L31L4 L41L5 L51L6 L31L4 L41L5 L5/L6 Do 3 .18 2.57 3 .16 7 87 7.72 7.72 DR 0.082 0.293 -0.141 -0 252 -0.648 -0.207 DL 4.95 5.36 5.75 5 65 5.56 5.75 TJ 170.5 127.6 126.9 217.0 231.6 213.6 ,; 0.9813 0.9876 0.9767 0.9709 0.9739 0.9524 51

PAGE 78

Table S.2 Tension neuromuscular neutral zones model parameters TNNZ(t) = To+(t-rrJTL e-[(/-rRJIr2]+ T M e-[(t-rRJIr3] Onset Offset L31L4 L4ILS LSIL6 L31L4 L4ILS LSIL6 To 5.58 3.52 4.49 0.00 0 00 0.00 TL 0.123 0.358 0.480 0.156 0.225 0.253 T2 50 1 27.9 27.8 100.7 80.6 74.2 TM 4.84 6.00 7.16 24.5 23.9 24.7 T3 140.0 144.9 106.4 643.9 485.7 498.2 l 0.8645 0.9927 0.9831 0.8775 0.9468 0.8974 Table S.3 Normalized peak MA V model parameters Peak MA V(t) = PO+PL e-[(t-rR)1 r./]+ P M (1-e-[(t-rR)/r5])+(t_Td) PH e-[(t-rd)lr6] L31L4 L4/LS LSIL6 Po -8.069 -7.64 -5.57 PL 9.00 8.52 6.33 T4 40 0 58.7 37.7 PM 8.99 8.81 6 57 TS 43.8 65.1 42.0 Td 330.3 339.4 308.5 PH 0.006 0.004 0.004 T6 287.4 230.0 233.5 0.9835 0.9175 0.9848 52