Citation
Neuromuscular neutral zone responses to static load variations

Material Information

Title:
Neuromuscular neutral zone responses to static load variations
Creator:
Le, Brook B
Publication Date:
Language:
English
Physical Description:
xi, 72 leaves : illustrations ; 28 cm

Subjects

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

Notes

Bibliography:
Includes bibliographical references (leaves 67-72).
General Note:
Department of Electrical Engineering
Statement of Responsibility:
by Brook B. Le.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
435526827 ( OCLC )
ocn435526827
Classification:
LD1193.E54 2009m L4 ( lcc )

Full Text
NEUROMUSCULAR NEUTRAL ZONE RESPONSES TO
STATIC LOAD VARIATIONS
by
Brook B. Le
B.S., University of Colorado at Denver, 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
2009


This thesis for the Master of Science
degree by
Brook B. Le
has been approved
by
c


Le, Brook B. (M.S., Electrical Engineering)
Neuromuscular Neutral Zone Responses to Static Load Variations
Thesis directed by Associate Professor Miloje Radenkovic
ABSTRACT
Two experiments were conducted, with identical protocols. In the first
experiment, ten in vivo feline models were loaded with a 20N force and subjected
to a simulated work schedule of six work-rest cycles in a 10:10 ratio (10 minutes
loading, 10 minutes rest), followed by 7 hours of recovery. EMG from the
preparations was monitored throughout the duration and later analyzed. An
identical protocol was used on the second group, only with a 60N loading force.
In both groups, results determined that muscular activity was diminished in the
first 2-3 hours post-loading, and creep in the viscoelastic tissues was significant
throughout the duration of the experiment in both cases. A compensatory
mechanism was observed in the 60N case that was not observed in the 20N case.
The presence of the compensatory mechanism was detected in the 3rd hour of
recovery, which suggests that it plays an instrumental role in spine stability.
While the presence of a compensatory mechanism has been confirmed by these
experiments, the exact nature of this mechanism and its place in the overall spinal
feedback system are unknown and bear further investigation.
This abstract accurately represents the content of the candidate's thesis. I
recommend its publication.
Signed
Miloje Radenkovic


DEDICATION
I dedicate this paper to my Mom and Dad, who taught me to pursue my dreams, to
my brother and sisters, who helped me reach them, and to all the rest of my
friends and family, who gave me their support every step of the way.


ACKNOWLEDGEMENT
I would like to thank my family for giving me encouragement and support
throughout this endeavor. You always made the road smooth when it got rough.
I would like to express my utmost thanks to Dr. Moshe Solomonow for providing
me with the opportunity to learn and study under his guidance, for all he has
taught me, and for all the patience he showed while I worked for him. You have a
personal touch that goes beyond the lab, and it has been an honor working for
you.
I would like to thank Dr. Bradley Davidson for all of his help on this project.
Your insights were informative and invaluable, and your work ethic has been an
inspiration.
I would like to thank Dr. Lu for teaching me about the anatomical aspects of this
work, and Dr. Zhou for lending his expertise in engineering and coding. Thank
you for sharing your knowledge and experience with me. It has been a pleasure
working with you.
I would also like to thank my committee members, Dr. Miloje Radenkovic and
Professor Robert Grabbe, for taking the time to review this work, and without
whom this would have never been possible. Thank you for giving me some of the
best learning experiences Ive ever had.


TABLE OF CONTENTS
List of Figures...........................................................ix
List of Tables............................................................xi
Chapter
1. Introduction............................................................1
1.1 Hypothesis.............................................................3
1.2 Significance of Study..................................................3
2. Physiological Background................................................5
2.1 Vertebral Column.......................................................5
2.2 Intervertebral Discs..................................................7
2.3 Facet Joints and Facet Capsules.......................................8
2.4 Ligaments.............................................................8
2.5 Multifidus Muscles...................................................10
2.6 Receptors............................................................12
2.7 Function of the Spine, Ligamento-Muscular Reflex.....................13
2.8 Spinal Stability.....................................................14
3. Electromyography (EMG).................................................16
3.1 Motor Unit Recruitment................................................17
3.2 EMG Peak Mean Absolute Value.........................................18
3.3 EMG Median Frequency.................................................18
vi


3.4 Recruitment Patterns..................................................19
4. Methods and Procedure.................................................20
4.1 Preparation..........................................................20
4.2 Instrumentation......................................................20
4.3 Experimental Protocol................................................21
4.4 Data Analysis........................................................22
4.5 DNNZ and TNNZ Definitions............................................23
4.6 Normalized EMG PMAV..................................................25
4.7 Median Frequency.....................................................25
4.8 Mean Creep...........................................................26
4.9 Mean Percentage Change from Baseline.................................27
4.10 Empirical Models....................................................27
4.11 Fitted Models.......................................................33
4.12 Statistics..........................................................39
5. Results...............................................................40
5.1 Displacement NNZs.....................................................40
5.2 Tension NNZs.........................................................44
5.3 Normalized Peak MAV..................................................49
5.4 Median Frequency.....................................................53
5.5 Mean Creep...........................................................56
vii


6. Discussion
58
6.1 Spinal Stability, Post-Loading......................................58
6.2 Comparison to Previous Studies......................................60
6.3 Validity of the Feline Model........................................63
7. Conclusion...........................................................65
References..............................................................67
viii


LIST OF FIGURES
Figure
2.1 Vertebral Structure of the Spinal Column. Image available
at: www.stjohnsmercv.org/. ./test/ortho/TP035.asp..................5
2.2 Lumbar vertebra from above and behind, (adapted from Grays
Anatomy of the Human Body online edition, 2000, modified
to enhance labels).................................................7
2.3 Posterior View of an Intervertebral Disc, Seated Atop a Single Vertebrae.
Image available at: www.mvhealthzone.com/.../28/000409.htm..........8
2.4 Median Sagittal Section of the Vertebral Column
Depicting Ligaments of the Spine Image available at:
www.coloradospineinstitute.com..................................9
2.5 Paraspinal Muscles. (Adapted from Grays Anatomy of the Human
Body online edition, 2000).........................................11
2.6 Simplified Ligamento-muscular Feedback System......................15
4.1 Experiment Protocol................................................21
4.2 Typical EMG Recording for Entire Length of Experiment..............22
4.3 Displacement and Tension NNZ Definitions...........................24
5.1 Mean Displacement NNZs (20N).......................................40
5.2 Mean Displacement NNZs (60N).......................................41
5.3 Mean Percentage Change from Baseline for 20N DNNZ..................42
5.4 Mean Percentage Change from Baseline for 60N DNNZ..................43
5.5 Mean Tension NNZs (20N)............................................44
IX


5.6 Mean Tension NNZs (60N)
45
5.7 Mean Percentage Change from Baseline for 20N TNNZ..................47
5.8 Mean Percentage Change from Baseline for 60N TNNZ..................48
5.9 Normalized Peak MAV (20N).........................................49
5.10 Normalized Peak MAV (60N)........................................50
5.11 Mean Percentage Change from Baseline for 20N MAV................51
5.12 Mean Percentage Change from Baseline for 60N MAV................52
5.13 Average Median Frequency for All Cats (20N).......................53
5.14 Average Median Frequency for All Cats (60N).......................54
5.15 Mean Percentage Change from Baseline for 20N MF.................55
5.16 Mean Percentage Change from Baseline for 60N MF.................56
5.17 Mean Percentage Change from Baseline of 20N Creep.................57
5.18 Mean Percentage Change from Baseline of 60N Creep.................57
x


LIST OF TABLES
Table
4.1 Definition/description of coefficients and terms of DNNZ(t)..........28
4.2 Definition/description of coefficients and terms of TNNZ(t)..........29
4.3 Definition/description of coefficients and terms of PMAV(t)..........30
4.4 Definition/description of coefficients and terms of MF(t)............32
4.5 Displacement neuromuscular neutral zones model parameters (20N)......34
4.6 Displacement neuromuscular neutral zones model parameters (60N)......34
4.7 Tension neuromuscular neutral zones model parameters (20N)...........35
4.8 Tension neuromuscular neutral zones model parameters (60N)...........36
4.9 Normalized peak MAV model parameters (20N)...........................37
4.10 Normalized peak MAV model parameters (60N).........................38
xi


1. Introduction
Low back pain is a disorder affecting millions of people worldwide. The
health and economic costs of this disorder are staggering. According to the
National Institute of Neurological Disorders and Stroke, low back pain is the
second most-common cause of job-related disability in America and a leading
contributor to missed work (NINDS 2008), with some $50 billion spent each year
on its diagnosis and treatment (NIH 2008). Even more billions are lost by
employers who seek to replace and train new workers, while injured workers are
subject to costly medical expenses, lost wages, and other disability expenses.
Nearly 8 out of 10 Americans will experience low back pain at some point
in their lives (NINDS 2008), and studies have shown that workers in labor-
intensive occupations involving repetitive motions or the lifting of heavy loads
have a higher risk of developing and experiencing low back disorders than does
the general population (Andersson 1981, Marras et al. 1995, Marras et al. 1993,
McGill 1997, Punnett et al. 1991, Xu et al. 1997).
While the exact causes of lower back pain are still unknown, ongoing
research has determined that some of the risk factors include prolonged static and
cyclic work cycles, short rest periods between loading cycles, and insufficient
duration of post-loading rest periods (Eversull et al. 2001, Gedalia et al. 1999,
Hoops et al. 2007, Le et al. 2007, Lu et al. 2004, Navar et al. 2006, Sbriccoli et al.
2004a, Sbriccoli et al. 2004b, Solomonow et al. 1999). Static work cycles involve
a loading of the spine such that the tissue of the spine is held flexed for a
pronounced period of time before it is relaxed, as exemplified by cotton pickers
who bend forward to pick cotton, and then remain in that position while
performing their work. Cyclic work cycles load the spine in such a way that the
tissues of the spine are stretched many times over a fixed time period, with a brief
1


maximal contraction in each stretching motion. Although cyclic loading shows
increased pathology over static loading (Hoops et al. 2007, Solomonow et al.
2008, Youssef et al. 2008), extended exposure of the lumbar spine to either
loading type results in the development of substantial creep and laxity in the
viscoelastic tissues which, in turn, alters the activation patterns of the spinal
musculature (Silverstein et al. 1986, Dickey et al. 2003, Le et al. 2007, Olson et
al. 2004, Solomonow et al. 2001).This has profound consequences on the stability
of the spine.
It was also discovered that recovery times for creep accumulated during
loading are longer than inducement times (Gedalia et al. 1999, van Dieen et al.
2003, Crisco et al. 1997, Ekstrom et al. 1996). Insufficient rest periods between
loading periods allow creep to accumulate over months and years until it becomes
a disorder known as Cumulative Trauma Disorder (CTD). CTD is a chronic
disorder that does not generally improve with medical therapies (Solomonow et
al. 2003b) and cannot be confirmed by diagnostic procedures (Courville et al.
2005). It is characterized by pain, weakness, limited range of motion, stiffness,
and spasms of the back muscles (Sbriccoli et al. 2004, Solomonow et al. 2003b,
Courville et al. 2005). It has been confirmed that workers subjected to cyclic,
static, and vibratory loading for extended periods of time are at risk for
developing CTD (Punnet et al. 1991, LaBry et al. 2004).
This study is the third in a series of tests investigating the spinal response
to static and cyclic loads of various magnitudes. The first study used cyclic loads
of moderate (40N) magnitude, the second static loads of moderate (40N)
magnitude. This study explores the spinal response to static loads of low (20N)
and high (60N) magnitudes. All experiments in this series of tests were conducted
with the same protocol in regard to loading type, i.e. the protocol used in this


experiment was exactly the same as that used in the static 40N experiment. In
both 40N tests, for both static and cyclic loads, creep was elicited in the lumbar
viscoelastic tissue, Neuromuscular Neutral Zones (NNZs) were enlarged
immediately post-loading, and a previously unidentified compensation
mechanism was evidenced (Youssef et al. 2008, Solomonow et al. 2008). Since
this experiment is a continuation of these previous studies, the results from this
one are expected to be similar to the other two.
1.1 Hypothesis
It is hypothesized that 60 minutes of static lumbar loading at low (20N)
and high (60N) magnitudes will elicit creep in the viscoelastic tissues and induce
enlargement of the Neuromuscular Neutral Zones (NNZs) immediately after
loading. It is further hypothesized that the viscoelastic tissues and NNZs will
return to their baseline measurements only after several hours have transpired. It
is predicted that significant changes will be observed in the EMG amplitude
and/or the EMG median frequency when the pre-loading data is compared to the
post-loading data. Based upon previous work, we expect to find no evidence of a
neuromuscular compensation mechanism in the case of the 20N loading, but we
do expect to find some type of compensation in the case of the 60N loading.
1.2 Significance of Study
Information obtained from this study may be able to provide insight into
the motor control compensation mechanism responsible for maintaining stability
of the lumbar levels of the spine after the performance of static loading. It may
also be used to assess the potential for injury in the lifting of heavy loads, and
suggest a means for its prevention while also providing insight into the prevention
3


of neuromuscular disorders such as CTD. This information may also be used as
baseline data for the design of safe work-scheduling. Additional studies would
need to be performed to transform the data obtained from the models used in this
study to that of a human model.
4


2. Physiological Background
2.1 Vertebral Column
The vertebral column extends from the base of the skull to the tip of the
coccyx. It is composed of 33 individual vertebrae, divided into five sections, as
shown in Figure 2.1. Beginning from the base of the skull and travelling
downward, these sections are named: 1) the cervical, 2) the thoracic, 3) the
lumbar, 4) the sacral, and 5) the coccygeal sections, respectively.
Spinal Column
with Vertebrae
/_
Cervical
Vertebrae
(7)
' c,-c7
Thoracic
Vertebrae
(12)
Ti-Ti2
Lumbar
Vertebrae
H5>
Li -1-5
Sacrum
(5 fused)
Coccyx
. (4 fused)
Figure 2.1 Vertebral Structure of the Spinal Column.
Image available at: www.stiohnsmercv.org/.../test/ortho/TP035.asp
The 7 vertebrae in the cervical section support the weight of the head.
Relatively flexible, they allow for flexion, extension, lateral, and axial movements
5


of this same area. Flexion is a bending in the forward, or anterior direction,
extension is a bending in the back, or posterior direction, lateral motion is
movement from side to side, and axial movement is a rotation about the spine.
The 12 vertebrae of the thoracic section support the trunk of the body and
serve as a stable point of attachment for the ribs and pelvic girdle.
The 5 vertebrae in the lumbar section support the entire upper body. These
are the largest vertebrae in the spine. They are subject to the heaviest loads and,
compared to the cervical vertebrae, are capable of only small flexion, extension,
lateral, and axial motions.
The 5 sacral vertebrae are fused together, as are the 4 coccygeal vertebrae.
Neither the sacral nor the coccygeal sections provide movement of the spine.
The focus of this paper is on the lumbar section of the vertebral column,
i.e. from lumbar level 1 (LI) to lumbar level 5 (L5). This section was chosen for
study as it is most prone to work-related injuries, due to the fact that it is often
subjected to high loads, stresses, and constantly changing force trajectories in
labor-intensive activities. A typical lumbar vertebra is shown in Figure 2.2. The
processes depicted in the figure are junction points where the muscles and
ligaments of the lower back connect to the spine.
6


Posterior spinous
Figure 2.2: Lumbar vertebra from above and behind, (adapted from Grays
Anatomy of the Human Body online edition, 2000, modified to enhance labels).
2.2 Intervertebral Discs
Intervertebral discs are viscoelastic structures located in between each
adjacent vertebra. Figure 2.3 shows their location and composition. The two
physical components of these discs are an outer ring composed of collagen fibers
called the annulus fibrosus, and an elastic gelatinous material that contains fluid
called the nucleus pulposus. The primary role of the discs is to maintain a strong
joint by providing a cushion for adjacent segments of the spine as they change
position relative to one another. Unlike the vertebrae themselves, the
intervertebral discs are pliable enough to deform as the spine undergoes articular
movements, yet they are stiff enough to resist excessive motion. They have been
shown to develop creep under load applied over time (Adams et al. 1990), which
reduces their effectiveness in maintaining spinal stability.
7


Annulus
flbrosus
*
&

t
Intervertebral disk
Nucleus
pulposus
Transverse
process
Spinal
cord
Spinous
process
Superior
articular facet
f'ADAM
Figure 2.3 Posterior View of an Intervertebral Disc, Seated Atop a Single
Vertebrae. Image available at: www.myhealthzone.com/.../28/000409.htm
2.3 Facet Joints and Facet Capsules
Two pairs of facet joints are located to either side of the spinous process of
each vertebra. These joints interlock with vertebra one level above and one level
below and limit the rotation and sliding of one vertebra relative to another. Facet
capsules are made up of collagenous membranes that surround and enclose each
facet joint. They also provide resistance to excessive motion. Under load, damage
to the joints and capsules diminishes the stabilizing role of these structures.
2.4 Ligaments
Ligaments are passive viscoelastic tissue composed of long, tightly
packed, elastic collagen fibers. They are connectors that link bone to bone. They
also promote guided joint movements and provide stability to the joint
(Solomonow et al. 2004). In the spine, ligaments attach to the walls and processes
8


of each vertebra and assist in holding the spinal column together. As seen in
Figure 2.4, there are many different ligaments in the spine. The one of primary
concern to this study is the supraspinous ligament as it is the first ligament to
experience tension and elongation during flexion (Adams et al. 1980). The
supraspinous ligament connects the tips of the posterior spinous processes of the
vertebrae from the seventh cervical vertebra to the sacrum. Under load, the
ligaments exhibit creep, stress-relaxation, and hysteresis, all of which are
dependent on frequency and strain-rate (Solomonow et al. 2004). These effects
cause laxity in the ligaments and reduce their overall effectiveness in preventing
exposure to injury.
Figure 2.4 Median Sagittal Section of the Vertebral Column Depicting Ligaments
of the Spine. Image available at: www.coloradospineinstitute.com
Ligamentum Flavum
, Intertransverse
! / Ligament
Posterior
Longitudinal
Ligament
-N. Anterior
Longitudinal
Ligament
9


2.5 Multifldus Muscles
Paraspinal muscles, shown in Figure 2.5, are the posterior muscles closest
to the spine. They help to stabilize the spine and allow it some degree of mobility.
Of particular interest to this study are the multifidus muscles, known collectively
as the multifidi. These muscles fill the grooves on either side of the spinous
processes of the vertebrae. They have a large cross-sectional area and a short fiber
length, which means they exert force only over small distances. They usually span
2 to 4 joint segments to connect the mamillary process of one vertebra to the
spinous process of the second, third, and fourth vertebra above it. The multifidi,
being closest to the spine, have a very short moment arm that develops minimal
torque but strong compressive forces for stability. Initial engagement of the
multifidi occurs as a direct reflexive response to stretching of ligaments and other
viscoelastic tissues of the lumbar spine. Hence, the role of the multifidi is to
provide the spine with stability rather than mobility (Kim et al. 2007).
10


Occipitnl houc
Figure 2.5 Paraspinal Muscles. (Adapted from Grays Anatomy of the Human
Body online edition, 2000)
11


2.6 Receptors
Information about the bodys environment is relayed to the brain through
neurons and inter-neurons existing in the central nervous system (CNS). These
neurons are activated by receptors in the body, which act as transducers to convert
various forms of environmental energy into action potentials in the neurons.
Five types of receptors exist in the human body. These are the
thermoreceptors that sense changes in temperature, nociceptors that sense pain,
photoreceptors that detect light in the retina of the eye, chemoreceptors that
respond to chemical stimulation in the mouth and nose, and mechanoreceptors
that respond to mechanical forces such as pressure and stretching. Of these five
receptors types, this paper is concerned only with the mechanoreceptors.
When stimulated by mechanical forces, i.e. pressure or stretching,
mechanoreceptors respond by firing nerve impulses, or action potentials, which
travel from the point of contact to the CNS, vis-a-vis the aforementioned neurons
and interneurons. The action potentials are assessed by different areas of the
Motor Control centers and an appropriate response is prescribed, which then
travels back through the CNS to the point of origin, where muscle activity is then
generated.
Five major types of mechanoreceptors have been identified. These are 1)
Meissners corpuscles, which adapt rapidly and are used primarily for fine touch
and detecting changes in temperature, 2) Golgi organs, which respond to extreme
forces, 3) Merkels discs, which adapt slowly to touch, 4) Ruffini corpuscles that
detect sustained pressure, and 5) Pacinian corpuscles, which are fast-adapting.
It has been shown that at least two types of mechanoreceptors are present
in the supraspinous ligament; these are the Pacinian corpuscles and Golgi organs
(Rhalmi et al. 1993). Pacinian corpuscles signal initiation and cessation of stimuli
12


within the CNS and provide continuous signaling for changes in tensile forces
(Rhalmi et al. 1993). The slow-adapting Golgi organs signal only during extreme
changes in motion and load (Rhalmi et al. 1993). Hence, ligaments and the other
passive tissues can also be considered to be sensory organs that relay information
to and from the CNS.
2.7 Function of the Spine, Ligamento-Muscular Reflex
One of the primary roles of the vertebral column is to protect the spinal
cord, which is the main pathway of the CNS. The spinal cord runs from the brain
to the coccyx through the vertebral foramen, which is a canal located within the
column itself. The spinal cord carries information from sensory organs to the
brain and responses from the brain to motor units via a complex network of inter-
neural pathways. Interruptions in these connections can cause a loss of sensation
and/or loss of motor control.
The other roles of the spine are to provide structural support and balance
to maintain an upright posture, and to allow flexible motion (Reeves et al. 2007).
It is not enough to merely carry out these tasks, however; the body must do so
without experiencing injury or pain (Reeves et al. 2007). Research has shown that
the spine is more resistant to injury when the spine is stable (Panjabi 1992), and
indicated that the body achieves this stability reflexively, through the employment
of a ligamento-muscular reflex (Stubbs et al. 1998). Subsequent research
determined that this reflexive feedback loop made use of both the passive and the
dynamic structures of the spine to achieve this stability (Panjabi 1992). Passive
structures are viscoelastic structures such as the discs, ligaments, and facet
capsules, and dynamic structures are muscles such as the multifidi. It has been
determined that the passive structures play a smaller role in spinal stability as
13


compared to the dynamic structures (Panjabi 1996, Solomonow et al. 1998,
Abumi et al. 1990, Anderson et al. 1985, Posner et al. 1982, Crisco et al. 1991,
Kaigleetal. 1995).
2.8 Spinal Stability
Spinal stability is achieved through the interaction of the passive and
dynamic structures of the back. Previous work has determined that under normal,
non-loaded conditions, the passive structures of the back (ligaments, discs,
capsules, etc.) are sufficient to maintain the stability of the spine (Wilke et al.
1995). Under these non-loaded conditions, the spine is in a neutral position, and is
relatively compliant to small perturbations about this position (Panjabi 1992). An
absence of muscular activity under these conditions was determined, leading to
the designation of these small displacement ranges as Neutral Zones (NZs)
(Panjabi 1992).
As long as the ligaments are stiff enough to support the spine, the NZs are
not exceeded and there is an absence of muscular activity (EMG). When the body
is sent into a trajectory, however, or is somehow loaded, the ligaments begin to
elongate. At some point the ligaments relegate their role as stabilizers to the
muscles of the lower back, which are one of an agonist-antagonist pair (the other
being the abdominal muscles) that contracts to maintain stability. Removal of the
load or a return to an unperturbed position results in a reverse process whereby
the stabilizing role played by the muscles is transferred back to the ligaments.
The interaction between the active and passive components of the spine
can be modeled as the closed-loop feedback system of Figure 2.6. The model is a
simplified diagram of the sensori-motor control system of the spine (Solomonow
14


et al. 1984) depicting the various interactions between the passive and dynamic
components.
Force
Length
Velocity
Acceleration
etc...
Figure 2.6 Simplified Ligamento-muscular Feedback System
Motor control, initiated in the cerebellum and motor cortex, is the input to
the system. The input is modified by various motor unit, musculoskeletal, and
viscoelastic dynamics present in the feed-forward path that serve to control the
force, length, velocity, acceleration, joint angle, etc. of spinal loads.
Measurements of these kinesthetic and proprioceptive outputs are sensed by
mechanoreceptors in the feedback loop and relayed to the motor control center of
the brain through inter-neurons in the spinal tract, which also connect to the
Central Nervous System through a sub-system that provides further modification
of the feedback signal. The fully processed signal is then compared to the
intended action and, if necessary, an appropriate response is formed. Continuous
15


feedback of this nature ensures that spinal stability under load is well-regulated by
the bodys Motor Control center.
3. Electromyography (EMG)
Electromyography is a technique used to record the activation of muscles
by monitoring their electrical activity. EMG recordings are usually made at rest
and during contraction.
At rest, skeletal muscle fiber has a membrane resting potential of
approximately -70 mV. When a movement is needed, electro-chemical
interactions raise the membrane potential of appointed muscle fibers to a
threshold level of approximately -55mV. Once this membrane potential is
reached, an action potential is generated that travels through muscle fibers and
causes the muscle to contract. An action potential is a pulse-like wave that
triggers muscle activity. It is important to realize that all action potentials have the
same magnitude.
The spatio-temporal summation of all the action potentials from all the
active motor units forms the EMG of the muscle and gives it its random shape.
EMG varies in frequency and amplitude depending on the intensity of the muscle
contraction. The contractions are sensed through the use of electrodes. There are
three types of electrodes: surface electrodes that lie flat on the surface of a body,
needle electrodes that are inserted into the body in an invasive procedure, and
wire electrodes that resemble needle electrodes, the difference being that wire
electrodes have thinner wires and non-fixed caps.
In these experiments, fine-wire bipolar electrodes were used since they are
capable of detecting EMG signals without interference from other muscles, as
16


would be the case if surface electrodes were used. They also have smaller
impedances than surface electrodes and greater flexibility than needle electrodes.
3.1 Motor Unit Recruitment
The basic motor unit consists of a single axon and the muscle fibers that it
innervates. Activation of the motor unit is initiated when receptors in the body are
exposed to a stimulus. The receptors transduce the information into an electrical
signal, which is carried by afferent nerves to the spinal cord. Interneurons in the
spinal cord direct this information, in an upward fashion, to the brain, and load the
information into the cerebral and motor cortices, where the information is
processed and an appropriate response is formed. The response information
travels back down the spinal cord to the axon, via efferent nerves, and triggers an
electrical response, which causes innervation of the muscle fibers. Smaller
muscles, used for precision movements, have a small innervation ratio, which
means that the axons of the smaller muscles are connected to fewer muscle fibers.
Large muscles, used for gross movements, have a large innervation ratio. This
means that the axons of the larger muscles are connected to many muscle fibers.
Axons are independent of each other and cannot innervate any muscle fibers aside
from the ones to which it is connected.
Studies have shown that motor neurons are recruited in a manner called
orderly recruitment, and it has likewise been shown that they obey the all-or-
none principle. Orderly recruitment means that motor units are recruited in
order, from smallest to largest, as the need for greater force arises. The all-or-
none principle states that the response of the muscle fiber is not dependent on the
strength of the stimulus. That is to say, if the stimulus is above a certain threshold
level, an action potential will be generated (complete response), and if the
17


stimulus is below the threshold level, no potential will be generated (no response).
Stimuli above the threshold do not generate stronger action potentials, as any
generated action potentials already give a complete response when the threshold
level has been met.
Two important measures are used to study the pattern of motor unit
recruitment. These are the EMG Peak Mean Absolute Value (PMAV) and the
EMG Median Frequency (MF).
3.2 EMG Peak Mean Absolute Value
The Peak Mean Absolute Value of the EMG record shows the relative
strength of muscular contractions. Plots of the EMG PMAV are generated by
detecting the magnitude of sensed action potentials. Higher values indicate higher
forces in a non-fatigued muscle.
3.3 EMG Median Frequency
The amplitude of the spatio-temporal summation of action potentials is
also affected by changes in the firing rate of motor units, which can be studied
through examination of EMG Median Frequency. Many studies have shown that
EMG frequencies are related to the average conduction velocity of active motor
units (Bellemere et al. 1979, Givens et al. 1978, Kadefors et al. 1968, Lindstrom
et al. 1970, Magora et al. 1976). It was also shown that this relationship is linear
(Stulen et al. 1981), with EMG frequencies increasing with increasing conduction
velocities. Since an increase in the conduction velocity leads to an increase in the
median frequency, which in turn indicates newly recruited, larger motor units, an
increase in the median frequency indicates an orderly recruitment of the motor
18


units. Thus, it is confirmed that the frequency component of the power spectral
density (PSD) can be used as an indicator of the EMG signal (Stulen et al. 1981).
3.4 Recruitment Patterns
Examination of these two measures suggests that the body obtains stronger
EMG, manifest as an increase in force, through any the following: by increasing
the firing rate of existing motor units in the pool of active motor units, by
recruiting more units into the active motor unit pool through motor unit
recruitment (i.e. by activating larger motor units) under a constant firing rate, or
by recruiting more units and increasing the firing rate simultaneously. Differing
patterns of recruitment are observed in each case. Analysis of the EMG provides
insight into which pattern of recruitment was taking place in the experiments.
19


4. Methods and Procedure
4.1 Preparation
Ten in vivo feline models were used in the 20N experiment, and eight
were used in the 60N experiment. The felines were used in a protocol approved
by the Internal Animal Care and Use Committee (IACUC). Each preparation was
injected with a 2mL dose of chloralose, and the skin over the spine was dissected
from the thoracic level to the sacrum to expose the dorso-lumbar fascia and the
supraspinous ligament. After dissection, the specimens were placed in a prone
position on the plate of a Bionix 858 Material Testing System (MTS). External
fixators were applied to the LI and L7 dorsal processes to prevent lumbar
interaction with the sacral and the thoracic structures. One end of an S-shaped
stainless steel hook was inserted into the supraspinous ligament above the L3-4
dorsal process, and the other end was attached to the force transducer of the MTS.
4.2 Instrumentation
Three pairs of fine stainless steel wire EMG electrodes were inserted about 3-
4 mm apart into the multifidus muscles of the L3-4, L4-5, and L5-6 dorsal
processes. A ground electrode was inserted into the gluteus muscle of the right
leg. Each electrode pair was fed into a differential EMG amplifier with a 110-db
CMRR, gain up to 200K, and a band-pass filter in the range of 6-500 Hz. The
output of a force transducer was sampled at 1000 Hz and stored in a computer for
later analysis.
20


4.3 Experimental Protocol
The MTS was first operated manually to pull the ligament taut, then
programmed to deliver three 4-sec cycles, each of which was followed by a 1
minute rest interval, followed by six cycles of work-rest administered in a 10:10
ratio (10 minutes of work followed by 10 minutes of rest), followed by test cycles
appearing at 10 minutes, 20 minutes, 30 minutes, and every hour thereafter during
the recovery period, for a total recovery time of 7 hrs. The procedure was exactly
the same for the second experiment, except that there were only 8 specimens, and
the loading forces applied were 20N in the first experiment and 60N in the
second. A schematic of the protocol is shown in Figure 4.1.
R1 R2 R3 *-1 ^ *"3 *-N
Pre-static Static loading period 7 hour recovery period
loading
t = 0.25 Hz
P1.P2.P3 Pre-static loading (0.25 Hz, 20 N, 60 N)
L,.L Ln Static load (10 min, 20 N, 60 N)
R1.R2'-.RN-1 Rest (10 min, no load)
N Number of static load repetitions
tt2 t9 Single test cycle during recovery period (0.25 Hz, 20 N, 60 N)
Figure 4.1 Experiment Protocol
21


4.4 Data Analysis
Figure 4.2 shows a typical EMG recording for the entire length of an
experiment. The first three rows of the figure depict the EMG recorded in each of
the lumbar levels L3-4, L4-5, and L5-6. The fourth row indicates the measured
displacement, and the last row shows the load magnitude. The first three entries in
each row are the baseline values, which were pooled together to form a composite
baseline measurement. The next six entries represent the six work-rest periods, in
a ratio of 10:10 (10 minutes work, 10 minutes rest), and the entries from 120
minutes onward indicate the rest or recovery period, as outlined in Fig. 4.1.
Spikes seen in the work and recovery periods are spasms, or reflexive muscle
contractions. Also, inspection of the displacement row reveals that creep
accumulated during the work-rest periods recovers only gradually during the rest
period. EMG was elicited in the recovery period through single-cycle tests. These
records were down-sampled and subsequently analyzed.
60N, STATIC, 6x10x10
g JCJ J
E
:}<( WM "HIM
: tH+§ W IN W| t * + H t H
liUUiU A A A
1111 ill 1.1
lAjlili.
I

60 -J .
:ii.
Ji. A
I
baseline 0 10 20 30 40 50 60 70 60 90 100 120 140 170 230 290 350 410 470 530
Time (min)
Figure 4.2 Typical EMG Recording for Entire Length of Experiment
22


4.5 DNNZ and TNNZ Definitions
Figure 4.3 depicts the down-sampled record of a typical single-cycle test.
In each of these tests, the first 500 msec of each EMG record prior to the initiation
of 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.
Similarly, the last EMG value to exceed five times the baseline noise was denoted
as the EMG cessation threshold in this channel, and its corresponding tension and
displacement values during the relaxation phase were designated as Tension and
Displacement Neuromuscular Neutral Zones Offset. The arrows in the plot
represent the EMG initiation and cessation thresholds.
23


Tension (N) Disp (mm) EMG (mV)
DNNZ on
DNNZ off
TNNZ off
TNNZ on
Time (sec)
Figure 4.3 Displacement and Tension NNZ Definitions
20
10
0
80
40
0
24


4.6 Normalized EMG PMAV
The strength of the EMG is dependent on muscle size and frequency of
firing rate. The EMG Peak Mean Absolute Value (PMAV) shows the relative
strength of muscular contractions.
The PMAV was analyzed through application of a moving average filter
to the single-cycle test records. The signal record of the EMG for each of the three
lumbar levels was full-wave rectified and smoothed with a low-pass filter with a
200-ms time constant. To clarify, the first three values of the records signal were
averaged together and plotted as one point, the window was moved to the next
three values which were averaged and plotted as well, and the process was
repeated until the whole EMG plot was sampled. This has a smoothing effect on
the data and helps to eliminate some of the sharp spikes in the EMG record, the
most likely cause of which are spasms and/or other artifacts.
4.7 Median Frequency
The Power Spectral Density (PSD) shows the frequency of firing rate of
the motor neurons. Hence, information provided by the PSD plot was used to
determine if the firing rate increased or decreased in response to loading.
In each single-cycle test of each EMG channel, a 0.5 second window
centered about the maximum tension was obtained. The data elicited from this 0.5
second window was zero padded (by pre- and appending zeroes to the initial data)
to yield 512, or 29 points of data. This set of data points was then multiplied by a
tukey window, which is an adjustable windowing function. The tukey window
was chosen specifically since it has the benefits of being similar to a rectangular
window (which means it didnt attenuate a significant part of the signal, which
25


can occur when you use a hamming or hanning window). Then the PSD of the
signal was computed by applying the fast Fourier transform to the data. (Note:
algorithms that implement the FFT on the computer are optimized for data sets
that are a multiple of 2; this is why we zero-padded the initial data to obtain 29
data points).
Plots of the PSD revealed that 60 Hz noise, and its harmonics, from the
power line were present. Hence, we applied the baseline noise spectrum
subtraction method to our processed data in order to filter out the 60 Hz noise and
its harmonics. To implement this method, we calculated the PSD of a 0.5 second
window from the baseline noise and then subtracted it from the PSD of our
processed signal. Using this final set of data, we were able to calculate the median
frequency, which is the frequency that divides the area of the power spectrum
signal exactly in half.
4.8 Mean Creep
To calculate the mean creep, the peak displacement of each channel, for
each single-cycle test, was detected. The data from the first three single-cycle
tests were averaged and the mean (SD) was designated the displacement
baseline value, or zero percent creep. After establishing this baseline, the
percentage of creep was calculated using the following equation:
creep =
h.t*ioo
L;
, where Lf is the value of the displacement at each specific single cycle test, and U
is the displacement baseline value. The resulting values of the creep in the
26


recovery period in each of the preparations were pooled, and the mean (SD) was
calculated and plotted as a function of time.
4.9 Mean Percentage Change from Baseline
After the baseline values were established, the percentage change from
these values over the full time-course was calculated for each specimen in each
data set (20N and 60N). This mean percent of change from baseline was evaluated
for each of the DNNZs (onset and offset), each of the TNNZs (onset and offset),
the normalized peak MAV, the median frequency (MF), and the Peak
Displacement. The percentages for each category of data were then averaged for
all preparations and plotted versus time.
4.10 Empirical Models
The mean ( SD) values of the DNNZ, TNNZ, and PMAV during
recovery for each lumbar level were fit with exponential-based models, as they
represent the classical response of viscoelastic tissues (Solomonow et al. 2000).
Levenberg-Marquardt nonlinear regression algorithms were used to generate the
best fit, optimized for the regression coefficient. No fits were applied to the
median frequency models due to a lack of significance in those models. The
following tables present the coefficients, along with their definitions and
descriptions, derived from the fits.
27


The time-course of the DNNZ thresholds during the stretch phase and
relaxation phase of the test cycles during the recovery period were described by:
DNNZ(t) = D0 + D{e 1 r' J {t | 120 Table 4.1 describes the coefficients and terms of this equation.
Table 4.1 Definition/description of coefficients and terms of DNNZ(t)
Coefficients/Terms of DNNZ(t) Definition/Description
Do intercept of the displacement (mm)
D, amplitude of decay dominating end of recovery period (mm)
time of first recovery measurement (120 min)
exponential time constant of decay dominating end of recovery period (min)
D,e ^ r' > exponential decay exhibited during the recovery period
The time-course of the TNNZ thresholds during the stretch phase and
relaxation phase of the test cycles during the recovery period were described by:
TNNZ (t) = T0 +(t Tr )TLe 1 rj J +TMe 1 J {t| 120 28


Table 4.2 describes the coefficients and terms of this equation.
Table 4.2 Defmition/description of coefficients and terms of TNNZ(t)
Coefficients/Terms of TNNZ(t) Definition/Description
T0 the intercept of the tension (N)
affects rise amplitude of exponential dominating beginning of recovery period (N/sec)
amplitude of decay dominating end ofrecovery period (N)
time of first recovery measurement (120 min)
affects rates of rise and fall (sec)
*5 exponential lime constant of decay dominating end of recovery period (min)
1 1 1 1 *- *1 allows for transient rise at beginning of recovery period
4 ] TMe 1 r5 J exponential decay dominating end of recovery period
The time-course of the PMAV during the recovery period was described by:
PMAV(t) =
P+Peyr'> + P
* 0 rLC rM
I e
-(?)
p+pe^r<)+p
r0 T rLc T rt
\
f
l-e
-r.'A
+ (f'
Td)PHey!>) ,t>Td
J
{t | 120 Table 4.3 describes the coefficients and terms of this equation.
29


Table 4.3 Definition/description of coefficients and terms of PMAV(t)
Coefficients/Terms of PMAV(t) Definition/Desc ri ption
P, inlcrcept of the peak MAV (mV)
Pl amplitude of exponential decay dominating beginning of recovery period (mV)
Pm amplitude of exponential increase following decay in beginning of recovery period (mV)
time of first recovery measurement (120 min)
time of onset of hyperexcitability (min)
h Exponential time constant of exponential decay dominating beginning of recovery period (min)
TZ Exponential time constant of exponential increase following decay in beginning of recovery period (min)
*3 Exponential lime constant of hyperexcilability term dominating end of recovery period (min)
f ] PLe 1 r' J allows for exponential decay dominating beginning of recovery period
f -N1 l-e [ > V J allows for exponential increase following decay in beginning of recovery period
-(1 hyperexcitability term with delayed onset dominating end of recovery period (equal to zero when t 30


The time-course of the MF (Median Frequency), smoothed with a 3-point
moving average algorithm during the recovery period, was described by:
MF(t) =
f '
h + FLe[ r J + Fm \-e{T: 1
\ )
J'-'A f J
Fn + Fi* { T' + Fm \-e ^ J >
\ )

+ (t-Td)FHe l r.
[t| 120 Table 4.4 describes the coefficients and terms of this equation.
31


Table 4.4 Definition/description of coefficients and terms of MF(t)
Coefficients/Terms of MF(t) Definition/Description
intercept of the peak MF (Hz)
FL Amplitude of exponential decay dominating beginning of recovery period (Hz)
Amplitude of exponential increase following decay in beginning of recovery period (Hz)
time of first recovery measurement (120 min)
Tu lime of onset of hyperexcilability (min)
^1 exponential time constant of exponential decay dominating beginning of recovery period (min)
T2 exponential time constant of exponential increase following decay in beginning of recovery period (min)
h exponential time constant of hyperexcitability term dominating end of recovery period (min)
-f l F, e { T' > allows for exponential decay dominating beginning of recovery period
V / allows for exponential increase following decay in beginning of recovery period
hyperexcilability term with delayed onset dominating end of recovery period (equal to zero when t 32


4.11 Fitted Models
Best-fit models were superimposed on the experimental data. The models
consisted of exponential equations whose variables were derived using the
Marquardt-Levenberg algorithms. Accuracy of the models was determined from
the resulting r2 values. Results displayed r2 values ranging from 0.833502 to
0.97659 for the 20N DNNZs (Table 4.5) and 0.918289 to 0.988064 for the 60N
DNNZs (Table 4.6), 0.879629 to 0.959045 for the 20N TNNZs (Table 4.7) and
0.929794 to 0.979454 for the 60N TNNZs (Table 4.8), 0.83087 to 0.961463 for
the 20N normalized peak MAV models (Table 4.9), and 0.932746 to 0.98673 for
the 60N normalized peak MAV models (Table 4.10).
33


Table 4.5 DNNZ model parameters (20N)
DNNZ(t) = D0 + D{e
Stretch Relax
L3/L4 L4/L5 L5/L6 L3/L4 L4/L5 L5/L6
D 0 2 2.046 2.01 2.051 3.658 3.7
D, 5.579 4.727 5.009 7.682 5 4.861
T 1 220 220 220 420 420 423.3
2 r 0.955852 0.97659 0.941304 0.875437 0.901526 0.833502
Table 4.6 DNNZ model parameters (60N)
-f1
DNNZ(t) = D0 + Dle 1 r' }
Stretch Relax
L3/L4 L4/L5 L5/L6 L3/L4 L4/L5 L5/L6
D 0 4.291 3.599 3.5 12.47 12.38 12.61
D, 8.689 7.62 9.191 3.703 3.787 4.178
T I 165 180 165.7 140.5 143.2 140
2 r 0.984588 0.933735 0.988064 0.933837 0.918289 0.957015
34


Table 4.7 TNNZ model parameters (20N)
TNNZ(t) = T0+{t-Tr)TLe 1 >+TMe
Stretch Relax
L3/L4 L4/L5 L5/L6 L3/L4 L4/L5 L5/L6
T 0 4.393 3.201 3.397 7.581 8.242 8.438
T L 0.14 0.14 0.14 0.14 0.1246 0.11
T 2 30 30 30 90 90 90
T M 5.618 6.576 7.6 5.17 5 6.465
T 3 382 382 382.6 350 350 350
2 r 0.893339 0.957758 0.932843 0.959045 0.89216 0.879629
35


Table 4.8 TNNZ model parameters (60N)
JlZlL
TNNZ(t) = T0+{t-Tr)TLe 1 +TMe 1
Stretch Relax
L3/L4 L4/L5 L5/L6 L3/L4 L4/L5 L5/L6
T 0 3.661 2.5 2 23.94 22.63 25.18
T L 0.355 0.3766 0.32 0.3 0.65 0.624
T 2 57.47 60.51 55 65.53 41.75 40
T M 17 10.24 18.03 12.89 9 13.67
T 3 207.9 208.7 190 180 295 295
2 r 0.979454 0.958422 0.963953 0.938668 0.929794 0.940166
36


Table 4.9 Normalized peak MAY model parameters (20N)
PMAV(t) = PQ+P,e
r
+ PM
V
-e
l T r
\
L3/L4 L4/L5 L5/L6
Po -2 -1.757 -1.926
Pl 2.917 2.483 2.606
T 4 74.4 50.05 85
P M 3.378 2.89 3.211
T 5 120 80 117.1
T d 289.3 265.5 290.5
P H 0.001468 0.000776 0.002276
T 6 180 150 120
2 r 0.83087 0.961463 0.912192
37


Table 4.10 Normalized peak MAY model parameters (60N)
PMAV(t) = P0 + PLe y rj J+PM
v
-e
T> J +{t-Td)PHe
L3/L4 L4/L5 L5/L6
Po -8.539 -8.983 -8.011
Pi. 10 9.813 9.033
T 4 60.27 41.15 50.62
P M 10.07 10.36 9.58
T 5 70 45 60
T d 286.5 326.9 283
P H 0.008982 0.001991 0.003931
T 6 90.78 90 90
2 r 0.932746 0.98673 0.974107


4.12 Statistics
A two-way analysis of variance (ANOVA) was performed on the DNNZ
and TNNZ data with time and vertebrae set as independent variables, and
displacement onset, displacement offset, tension onset, and tension offset set as
the dependent variables. A one-way ANOVA was performed on the normalized
PMAV, MF, and Peak Displacement, with time set as the single independent
variable. The Students T test was applied to the results in order to determine
significance of variations, with significance set at p < 0.05. All data were assumed
to have normal distribution curves. In the cases where the data sets did not have a
well-defined normal distribution, the data set was transformed by applying an
appropriate transcendental transformation to the data set. All figures, however,
depict the original, untransformed data sets.
39


5. Results
5.1 Displacement NNZs
L3-4
Time (min)
L4-5 L5-6
Time (min) Time (min)
Figure 5.1 Mean Displacement NNZs (20N)
In the 20N experiment, the mean ( standard deviation) of the averaged
baseline, obtained from the pre-loading values, was 2.6374 mm ( 1.2036 mm)
for the L3/4 level, 1.9610 mm ( 0.8123 mm) for the L4/5 level, and 2.5497 mm
( 0.6224 mm) for the L5/6 level. During the stretch phase, the DNNZ values
were higher than the baseline value at the start of the recovery period, yielding
values of 6.9720 mm ( 2.7243 mm) for the L3/4 level, 6.7017 mm ( 2.009 mm)
for the L4/5 level, and 6.9283 mm ( 2.2121 mm) for the L5/6 level. The DNNZ
values decreased over the remaining recovery period, with final values of 3.0314
mm ( 1.4585 mm) for the L3/4 level, 2.9311 mm ( 1.5216 mm) for the L4/5
level, and 2.7867 ( 1.3278 mm) for the L5/6 level.
In the relaxation phase, the mean ( standard deviation) of the averaged
baseline values were 4.4393 mm ( 1.3766 mm) for the L3/4 level, 4.4077 mm (
40


1.0976 mm) for the L4/5 level, and 4.5853 mm ( 1.0153 mm) for the L5/6 level.
The DNNZ values in the relaxation phase were higher than the baseline
throughout the recovery period. Immediately after loading, the DNNZ values had
increased to 8.736 mm ( 2.3768 mm) for the L3/4 level, 8.6683 mm ( 2.4379
mm) for the L4/5 level, and 8.5933 mm ( 1.8957 mm) for the L5/6 level. These
values decreased throughout the remainder of the recovery period to yield final
values of 4.6233 mm ( 1.2567 mm) for the L3/4 level, 5.3533 mm ( 1.3563
mm) for the L4/5 level, and 5.3317 mm ( 1.2369 mm) for the L5/6 level.
L3-4
Time (min)
L4-5
Time (min)
L5-6
0 100 200 300 400 500 600
Time (min)
Figure 5.2 Mean Displacement NNZs (60N)
The mean ( standard deviation) of the averaged baseline, obtained from
the pre-loading values, was 5.1844 mm ( 3.7342 mm) for the L3/4 level, 4.6044
mm ( 3.5126 mm) for the L4/5 level, and 5.1483 mm ( 3.2349 mm) for the
L5/6 level. During the stretch phase, the DNNZs showed a significant increase
over baseline at the start of the recovery period, yielding values of 12.5457 mm (
3.9827 mm) for the L3/4 level, 10.0143 mm ( 4.0738 mm) for the L4/5 level,
and 12.72 mm ( 3.6817 mm) for the L5/6 level. The DNNZs decreased over the
41


remaining recovery period, with final values of 4.86 mm ( 2.7519 mm) for the
L3/4 level, 3.875 mm ( 1.6899 mm) for the L4/5 level, and 4.22 mm ( 1.9609
mm) for the L5/6 level.
In the relaxation phase, the mean ( standard deviation) of the averaged
baseline values were 8.8322 mm ( 3.6161 mm) for the L3/4 level, 12.28 mm (
7.0528 mm) for the L4/5 level, and 8.8939 mm ( 4.02 mm) for the L5/6 level.
The DNNZs in the relaxation phase showed significant increase over baseline
throughout the recovery period. At the beginning of the recovery period, the
DNNZs had increased to 16.3443 mm ( 4.6119 mm) for the L3/4 level, 15.825
mm ( 4.1012 mm) for the L4/5 level, and 16.615 mm ( 4.3493 mm) for the
L5/6 level. The values decreased throughout the remainder of the recovery period
to yield final values of 13.0983 mm ( 3.7266 mm) for the L3/4 level, 13.0167
mm ( 4.3128 mm) for the L4/5 level, and 12.965 mm ( 4.4754 mm) for the
L5/6 level.
The following figure shows the mean percentage change from baseline for
the 20N DNNZs. This measure was computed as:
|(recovery value baseline value)/baseline value|*100
L3-4 L4-5 L5-6
0 100 200 300 400 500 600 0 100 200 300 400 500 600 0 100 200 300 400 500 600
Time (min) Time (min) Time (min)
Figure 5.3 Mean Percentage Change from Baseline for 20N DNNZ
42


Through statistical analysis, i.e. the Students T test, it was revealed that
all of the values through the fifth hour of recovery were significant in the L3-4
and L5-6 levels, for both phases, and all values were significant in the L4-5 level,
for both phases. No time/phase interactions were detected in any of the lumbar
levels. Significance was set at p < 0.05.
In this case, all displacements in the stretch phase experienced a greater
change than their counterparts in the relaxation phase, leading to an inversion of
the data presented in Figure 5.1.
The following Figure shows the mean percentage change from baseline for
the 60N DNNZs. This measure was computed as:
|(recovery value baseline value)/baseline value|*100
L3-4 L4-5
800
600
400
200
0
-200
0 100 200 300 400 500 600 0 100 200 300 400 500 600 0 100 200 300 400 500 600
Time (min) Time (min) Time (min)
n = 8 Strath Tt C Relax 800 600 n = 8 # Stretch O Relax
I 400
-.-slum- 200 0 *Ht It
L5-6
n = e
Sireich
O Relax
Figure 5.4 Mean Percentage Change from Baseline for 60N DNNZ
Only the first five values in each of the levels L3-4, L4-5, and L5-6 were
significant, for both phases. No time/phase interactions were detected in any of
the lumbar levels. Significance was set at p<0.05. In this case the values of the
43


stretch phase show a greater percentage change in displacement than do those of
the relaxation phase. However, this change is reduced by the third hour of
recovery, indicating that the ligament is staying relatively immobile.
5.2 Tension NNZs
L3-4 L4-5
Time (min)
Time (min)
L5-6
Time (min)
Figure 5.5 Mean Tension NNZs (20N)
In the stretch phase of the TNNZs, the averaged pre-tension baseline
values were 9.4267 N ( 3.891 N ) for the L3/4 level, 6.2347 N ( 1.99 N) for the
L4/5 level, and 8.298 N ( 1.8194 N) for the L5/6 level. In the recovery period,
the values of the TNNZs immediately following loading were 9.3240 N ( 4.1771
N) for the L3/4 level, 9.4783 N ( 4.0557 N) for the L4/5 level, and 10.9900 N (
3.8157 N) for the L5/6 level. An increase in the mean ( standard deviation) was
observed in each level up to the first hour of recovery, and then the mean (
standard deviation) decreased in an exponential manner, falling below baseline
values in the third, seventh, and fifth hour for the L3/4, L4/5, and L5/6 levels,
respectively. Further decrease until the ninth hour of recovery was observed,
44


yielding values of 6.7057 N ( 3.5763 N) for the L3/4 level, 6.0700 N ( 3.559 N)
for the L4/5 level, and 5.7156 N ( 2.272 N) for the L5/6 level.
In the relaxation phase, the (20N) TNNZs had an averaged pre-tension
baseline value of 12.8037 N ( 3.3801 N) for the L3/4 level, 11.3617 N ( 2.2487
N) for the L4/5 level, and 12.9853 N ( 3.2807 N) for the L5/6 level. No
significant changes were observed until the (later hours of recovery) in any of the
lumbar levels. In the recovery, 30 minutes after loading, the TNNZ values were
12.7080 N ( 3.7274 N) for the L3/4 level, 12.7733 N ( 4.6788 N) for the L4/5
level, and 15.8983 N ( 4.2851 N) for the L5/6 level. The TNNZ levels began to
decrease in an exponential manner after the fourth hour of recovery for all levels,
dropping below baseline values (in all cases) after the sixth hour of recovery, and
yielding final values of 9.3683 N ( 3.4854 N) for the L3/4 level, 11.7033 N (
3.6008 N) for the L4/5 level, and 11.9733 ( 3.0646 N) for the L5/6 level.
L3-4
L4-5
L5-6
Time (min) Time (min) Time (min)
Figure 5.6 Mean Tension NNZs (60N)
In the stretch phase of the TNNZs, the averaged pre-tension baseline
values were 18.5883 N ( 14.0127 N ) for the L3/4 level, 15.8767 N ( 12.2965
45


N) for the L4/5 level, and 18.225 N ( 11.9641 N) for the L5/6 level. In the
recovery period, the values of the TNNZs immediately following loading were
21.0271 N ( 4.2416 N) for the L3/4 level, 11.93 N ( 9.0776 N) for the L4/5
level, and 20.8325 N ( 8.2625 N) for the L5/6 level. An increase in the mean (
standard deviation) was observed in each level up to the first hour of recovery,
and then the mean ( standard deviation) decreased in an exponential manner,
falling below baseline values in the fourth hour for the L3/4, L4/5, and L5/6
levels. Further decrease until the ninth hour of recovery was observed, yielding
values of 6.7313 N ( 8.7726 N) for the L3/4 level, 3.2913 N ( 1.9268 N) for the
L4/5 level, and 4.0737 N ( 3.683 N) for the L5/6 level.
In the relaxation phase, the (60N) TNNZs had an averaged pre-tension
baseline value of 32.9817 N ( 23.0523 N) for the L3/4 level, 37.201 N (
18.8936 N) for the L4/5 level, and 34.7333 N ( 24.343 N) for the L5/6 level. No
significant changes were observed until the (later hours of recovery) in any of the
lumbar levels. In the recovery, 30 minutes after loading, the TNNZ values were
37.7343 N ( 12.7883 N) for the L3/4 level, 29.605 N ( 12.8772 N) for the L4/5
level, and 37.7575 N ( 11.8767 N) for the L5/6 level. The TNNZ levels began to
decrease in an exponential manner, dropping below baseline values after the
fourth, fifth, and fourth hour of recovery for the L3/4, L4/5, and L5/6 levels,
respectively, and yielding final values of 27.765 N ( 10.809 N) for the L3/4
level, 26.2383 N ( 10.0824 N) for the L4/5 level, and 27.845 N ( 13.4121 N)
for the L5/6 level.
46


The following figure shows the mean percentage change from baseline for
the 20N TNNZs:
L3-4
L4-5 L5-6
Time (min) Time (min)
Figure 5.7 Mean Percentage Change from Baseline for 20N TNNZ
The diagrams reveal that the tension in the ligament remained relatively
equal in both phases, for each of the lumbar levels. This is seen through the points
and the somewhat equal spacing in the standard deviations of each phase.
Significance was found in the first half hour, the second hour, and the sixth hour
in the L3-4 level, while all data in the first three hours were significant in the L4-5
level. All data except that of the third, fourth, and fifth hours of recovery were
significant in the L5-6 level, for both phases. Significance was set at p<0.05. No
time/phase interaction was detected in any of the lumbar levels.
47


The following figure shows the mean percentage change from baseline for
the 60N TNNZs:
L3-4
L4-5
L5-6
0 100 200 300 400 600 $00
Time (min)
Figure 5.8 Mean Percentage Change from Baseline for 60N TNNZ
The diagrams reveal that the tension in the ligament experienced a greater
change in the relaxation phase, as compared to the stretch phase, for each of the
lumbar levels. In the L3-4 and L5-6 levels, this was true throughout the whole
recovery period, whereas this change occurred during the third hour of recovery
in the L4-5 level. Significance was found in the first half hour, the first hour, and
the fifth, sixth, and seventh hours in the L3-4 level, for both phases, in the first
half hour and the fourth through seventh hours of the L4-5 levels, for both phases,
and in the third through seventh hours of the L5-6 levels. Significance was set at
p<0.05. No time/phase interaction was discovered in any of the lumbar levels.
48


5.3 Normalized Peak MAV
The Peak Mean Absolute Value (PMAV) of the EMG was recorded and
analyzed to determine the relative size of the active motor units throughout the
experiment.
L3-4 L4-5 L5-6
Time(min) Time(min) Time(min)
Figure 5.9 Normalized Peak MAV (20N)
The above figure shows the time-course of the Peak Mean Absolute Value
(PMAV) during the recovery part of the 20N experiment. The values were
normalized with respect to the baseline values of each channel. Statistical analysis
showed that there was a pattern of values below the baseline, followed by an
upward trend after the second hour of recovery for the L3-4 level, and after the
third hour of recovery for the L4-5 and L5-6 levels. By the fourth hour of
recovery, the PMAV had returned to baseline levels, and then showed a trend of
increasing over the next three hours of recovery.
49


L3-4
L4-5
L5-6
Figure 5.10 Normalized Peak MAV (60N)
The above figure shows the time-course of the Peak Mean Absolute Value
(PMAV) during the recovery part of the 60N experiment. All values were
normalized with respect to the baseline. In the L3-4 level, the PMAV was about
160% higher than the baseline immediately after loading, followed by a decrease
to baseline levels within the next two hours. This, in turn, was followed by an
increase in PMAV to almost 200% of the baseline levels at its highest, with a
slight downward trend in hours five through seven. In the L4-5 level, the PMAV
was lower than baseline by 20% for the first hour, and then it increased from the
second hour to the sixth hour, where it attained a maximum of almost 160%
above the baseline value. In the L5-6 level, the PMAV was just above baseline
immediately after loading, and showed a decrease to about 50% of baseline in the
first half hour after loading. By the second hour, the PMAV was at baseline, and
then showed an increase in subsequent hours, up to a maximum of almost 180%
above baseline.
50


The following figure shows the mean percentage change from baseline for
the 20N PMAV:
L3-4 L4-5 L5-6
Figure 5.11 Mean Percentage Change from Baseline for 20N MAY
51


The following figure shows the mean percentage change from baseline for
the 60N PMAV:
L3-4
L4-5
L5-6
0)
c
(0
CD
E
o
0)
cn
c
(Q
O
Time(min)
Time(min)
Time(min)
Figure 5.12 Mean Percentage Change from Baseline for 60N MAV
Examination of the Mean Percentage Change from Baseline for these
values indicates that the subjects loaded with 60N had a higher mean percentage
change from baseline. This, in turn, indicates that more motor units were recruited
in the 60N case than in the 20N case.
52


5.4 Median Frequency
L3-4
L4-5
L5-6
Time(min)
Time(min)
Time(min)
Figure 5.13 Average Median Frequency for All Cats (20N)
In the L3-4 level, the values stayed fairly constant during the first five
hours of recovery, showing an increase at the sixth and seventh hours to only 1 %
above baseline. In the L4-5 level, the median frequency showed a slight decrease
in the first half hour, stayed at that level until the third hour, and then returned to
baseline value by the fourth hour and stayed at that level throughout the
remainder of the experiment. In the L5-6 level, the median frequency was about
2% above baseline for the first hour of the experiment, decreased to baseline
levels during the second and third hours, and then increased to about 2% above
baseline in the remaining hours of recovery.
53


L3-4
L4-5
L5-6
N
I
>.
O
c
0
13
CJ
0
C
0
0
0
0
Q_
Time(min) Time(min) Time(min)
Figure 5.14 Average Median Frequency for All Cats (60N)
The above figure shows the Median Frequency for the 60N case. In the
L3-4 level, the median frequency was about 5% below baseline immediately after
recovery, and gradually increased to just below baseline levels by the third hour
of recovery. In the L4-5 level, the median frequency was about 5% below
baseline immediately after recovery, then dipped to about 15% below baseline by
the second hour. From the second hour of recovery to the fourth, the median
frequency increased, to about 10% above baseline by the fourth hour. From the
fifth hour to the seventh, the median frequency was more or less constant,
displaying values that were approximately 10% above baseline. In the L5-6 level,
the median frequency showed no change from baseline during the first hour of
recovery, and then gradually increased to about 20% above the baseline value
during the second to third hours, held constant during the fourth to fifth hours, and
then decreased slightly during hours six and seven, ending up at about 1.9%
54


above baseline in the seventh hour. No significance was found for any of the
points in this 60N case, although each of the channels displayed a trend of initial
decrease, followed by substantial increase over the baseline values.
The following figure shows the mean percentage change from baseline for
the 20N MFs:
L3-4 L4-5 L5-6
Time(min) Time(min) Time(min)
Figure 5.15 Mean Percentage Change from Baseline for 20N MF
55


The following figure shows the mean percentage change from baseline for
the 60N MFs:
L3-4
L4-5
L5-6
Time(min) Time(min) Time(min)
Figure 5.16 Mean Percentage Change from Baseline for 60N MF
5.5 Mean Creep
The pooled and averaged baseline value for all cats in each experiment
was used to denote zero creep. For both the 20N and 60N experiment, creep was
present throughout the entire recovery period, with its highest values appearing
immediately after loading. Subsequently, the creep gradually decreased over time.
This is seen in Figure 5.17 for the 20N case and Figure 5.18 in the 60N case. It
was observed that the creep induced in the 20N experiments recovered more fully
than did the creep in the 60N case, suggesting that creep elicited from heavy static
loads remains in the viscoelastic tissues far longer than it does for light static
loading.
56


Figure 5.17 Mean Percentage Change from Baseline of 20N Creep
Figure 5.18 Mean Percentage Change from Baseline of 60N Creep
57


6. Discussion
Results indicate that overall, regardless of load magnitude, during the first
1-2 hours post-work the lumbar spine is not protected by the ligaments or
muscles, increasing the potential for spinal instability and injury. Also, the
presence of a compensation mechanism, separate from the ligamento-reflexive
feedback loop, was detected in the 60N loading case. Under light static loading,
this compensation mechanism is suppressed as the lumbar level of the back is not
sufficiently taxed to the point that the spine is unstable. Under high magnitude
loading this compensation mechanism triggers early in the recovery period and
acts to reduce spinal instability.
6.1 Spinal Stability, Post-Loading
Consistent with previous studies, creep was significant for the duration of
the recovery period in both experiments, indicating that seven hours of rest is
insufficient time for full recovery of the viscoelastic structures. Creep was more
pronounced in the 60N case as compared to the 20N case. In the latter case, the
average peak displacement approached baseline levels by experiments end, but in
the 60N case the average peak displacement was still greatly increased over
baseline levels. This would indicate that the ligament experiences greater trauma
when exposed to heavy loads as compared to lighter loads. Laxity in the
viscoelastic structures during this period results in increased risk of injury.
DNNZ time constants in the 60N exponential model were smaller than
those of the 20N model, indicating that the 60N DNNZs narrowed more quickly
than those of the 20N case. By the fifth hour of recovery, the 60N DNNZs had
returned to baseline, and the significance plot revealed no further changes in the
succeeding hours, implying the presence of muscular activity. In the 20N case,
58


DNNZs remained elevated from baseline levels until the seventh hour of
recovery, indication that muscle activity was suppressed throughout the recovery
period.
Time constants in the TNNZ empirical models showed that the 60N
TNNZs widened more quickly, and narrowed more quickly, than did those of the
20N case. The corresponding significance plots revealed that the 60N TNNZs are
below baseline measurements after the second and succeeding hours, whereas the
20N TNNZs are within baseline values until the fourth or fifth hour. The narrower
TNNZs in the 60N case are another indication of increased muscle activity.
Plots of the EMG PMAV show that the contractions in the 60N loading
were generally stronger than those of the 20N case, with increased percentage
change from baseline values. This reflects the greater force needed by the spine to
resist heavier loads. Significance was found only within the first hour of recovery
for the 20N case, and barely any significance was seen in the 60N case, with the
exception of the L5-6 lumbar level. In this level muscular contractions were near
baseline at the start of recovery, and significantly higher near the end of the
recovery period.
EMG Median Frequency plots showed some significance at the end of the
20N recovery period, but overall the frequency did not deviate from baseline
levels, an indication that action potentials were being fired at a fairly constant
rate. No significance was found for the 60N median frequency, but the plot
showed a trend of increase into the seventh hour of recovery, a possible indication
of increased firing rate.
Hence, EMG analysis in both cases revealed a period of two to three hours
post-loading during which DNNZs and TNNZs were elevated above pre-loading
baseline values, indicating shorter periods of muscular activity. In the same two to
59


three hour period, a decrease in EMG amplitude was observed along with
decreased levels of the median frequency, implying that smaller motors units were
being fired at slower rates and suggesting that less muscle force was available to
the spine during this period, resulting in increased spinal instability and risk of
injury.
Subsequently, in the 60N case, the DNNZs and TNNZs became narrower
in the fourth hour of recovery, with an observed increase in both the MAV and
MF plots, suggesting that muscle activity was being generated in order to
compensate for the loss of stability. In the case of 20N loading, this compensation
mechanism was non-existent, presumably because the loading was of insufficient
strength to tax the stability of the spine. In the 60-N case, the compensation
mechanism triggered earlier in the recovery period, somewhere between the third
and fourth hour of recovery. This mechanism is believed to preserve lumbar
stability by compensating for the laxity developed by the passive lumbar
viscoelastic tissues.
6.2 Comparison to Previous Studies
Studies completed just prior to this one afford us with the means to
compare static and cyclic loading responses of low, mild, and high loads.
In both cyclic and static loading cases, responses from 20N loading
magnitude were similar. DNNZs in both the stretch and relaxation phases had the
same pattern of behavior, and were modeled with the same model. Time constants
in the cyclic case were slightly lower in L3-4 and L5-6 levels than those of the
static case, but those of the L4-5 case were nearly identical. Both cases showed
significance throughout much to all of the recovery period, and cyclic loading had
higher changes from baseline measures. TNNZs were also similar, with peak
60


values appearing around the first hour into recovery, followed by a decrease for
the remaining duration. Initial increase was faster in the static case, but decrease
was faster in the cyclic case. Significance was almost the same in both cases;
significance was found in the first two hours after loading in both cases, and in the
eighth and ninth hours for the static case. MAVs also had similar trend, first
dipping slightly below baseline and then increasing above baseline after the
second hour of recovery. MFs differed slightly. In the static case, MFs stayed
fairly constant throughout the duration, but in the cyclic case the trend was much
like that of the MAVs.
Results from the 40N tests suggested that cyclic loading is more
detrimental to the stability of the spine than is static loading. The DNNZs in the
stretch phase of the 40N cyclic study were found to have a greater percent of
change above baseline immediately after loading than did their static counterpart.
Also, the TNNZs in both phases showed higher increases over baseline
measurements in the cyclic loads than the static loads, and also took a longer time
to return to their baseline measurements. Differences in the EMG PMAV and MF
levels were also apparent in the 40N tests. In cyclic loading, the EMG PMAV was
at a minimum immediately after loading. In static loading, this minimum value
was not attained until one hour of recovery had already transpired. Also, the MF
data from cyclic loading was found to be significant, but the MF data from static
loading was not significant. This implies that the MF from static loading never
deviated too greatly from baseline measurements. Hence, the muscle recruitment
patterns in each case were found to be different. In the cyclic case, motor units
were first de-recruited, and then were recruited later in the recovery period,
indicating a need for the employment of additional muscle force. In the static
case, additional force was generated through increased firing rates of active motor
61


units, rather than the recruitment of more units. A compensatory mechanism was
detected in both cases, but it triggered later in the recovery period for cyclic
loading than it did for static loading.
Results from the 60N tests were similar to those of the respective 40N
cases. With respect to the 40N case, greater damage was caused due to higher
loading magnitudes, with cyclic loading having increased tissue damage and more
diminished muscle capacity than static loading. In the 60N tests, the percent
change from baseline of DNNZs and TNNZs in the cyclic case was more
pronounced than that of the static case, and it took longer for the cyclic TNNZs to
return to their baseline values. EMG PMAV In the cyclic case, minimal values of
the EMG PMAVE appeared immediately after loading, whereas a delay was
experienced in the static case. Median frequency was significant in the cyclic
case, but not in the static case, although both showed a pattern of increase
throughout the recovery period. Motor unit recruitment patterns were thus largely
unchanged from the 40N case, with stronger responses being elicited in the 60N
cases. Even at high magnitude loading, there was an absence of hyperexcitability
and neuromuscular disorder for static loads, while both were present in cyclic
loads, characterized by numerous spasms in their EMG plots. A compensation
mechanism was detected in both cases. The compensation mechanism triggered
around the same time in the 60N cases, earlier in the recovery period as compared
to the 40N tests.
Hence, the data from the tests indicate that cyclic loading is more
hazardous than static loading, due to shorter durations of protective reflexes and
longer periods of exposure to injury. It was also determined that greater tissue
damage was experienced in the cyclic case than in the static case. Thus, it was
62


determined that cyclic loading is more detrimental to spinal stability than is static
loading when considering the respective load magnitudes.
6.3 Validity of the Feline Model
Obvious differences are apparent between the physiology of a feline and a
human, necessitating the need to address the issue of relevance in this study. The
data obtained from these experiments were elicited from an in vivo feline model.
To validate the results, we need to determine whether or not the data is applicable
to human models.
Humans, being bipeds, balance themselves on two feet while felines,
being quadrupeds, walk on all fours. Despite the differences in load dispersions
caused by these pedal arrangements, with the corresponding restrictions placed on
the alignment of the gravity vector of the spine (the gravity vector of a human
spine is oriented straight down, whereas the gravity vector of the feline spine is at
90 degrees from that of humans), a recent study has shown that quadruped models
experience the same types of axial compression in the spine as do bipeds (Smit et
al. 2002).
Research has also shown that the spine of a feline has many biomechanical
similarities to that of a human (Smit et al. 2002). Like humans, the elemental
make-up of the passive structures of the feline spine consists of collagen
structures in the ligaments, facet capsules, and discs. These viscoelastic structures
of the feline contain mechanoreceptors (Hirsch 1963, Yahia 1988) and exhibit
results similar to those of humans when they are placed under load (i.e. they
exhibit creep, tension-relaxation, hysteresis, etc.). Stabilizing muscles close to the
spine are present in both species, and studies have confirmed the presence of a
ligamento-muscular reflex in both humans and felines (Solomonow et al. 1988).
63


Furthermore, the neuromuscular disorder associated with feline models has also
been confirmed in humans (Olson et al. 2006, Olson et al. 2004, Sbriccoli et al.
2005, Chu et al. 2003).
Differences in size between the two species would necessitate the scaling
of feline data to that of humans, which indicates necessary adjustments in the
empirical models. We would expect longer response times in humans due to the
greater distances of travel by action potentials. Ligaments in the human spine are
much also much larger than those in felines, suggesting that heavier loads are
required in order to initiate a response.
If these factors are taken into account, however, along with differences in
metabolic and hormonal characteristics, it is reasonable to assume that data
extrapolated from feline models has relevance to humans.
64


7. Conclusion
The results indicate that the spine is susceptible to injury 2 to 3 hours after
low or high static loading. Creep is present throughout the recovery period,
indicating that the passive viscoelastic structures are more lax than they are during
normal activity, contributing to a reduction of stability of the spine. Also, the
TNNZs and DNNZs are enlarged during the early part of recovery, indicating that
there is delayed or diminished muscle activity, which increases risk of injury to
the spine. The following conclusions were draw n from the data obtained in these
experiments:
1. A sequence of static loading at a 1:1 work-to-rest ratio, for a
moderate cumulative loading duration and a low load
significantly increases the displacement and tension
neuromuscular neutral zones. The NNZs are also significantly
increased for moderate cumulative loading duration and high
loads.
2. In the two to three hours immediately after loading, the lumbar
spine is exposed to significant reduction of stability control and
high risk of injury. The risk of injury and reduction of stability
is greater for higher loading.
3. A neuromuscular control compensation mechanism was found
to exist (different from the simple ligamento-muscular reflex)
in the 60N load. This mechanism triggers early in the recovery
period, within two to three hours, enhancing the role of the
65


musculature of the spine and allowing the viscoelastic tissues
to recover. This mechanism was not present in the 20N load.
The results of this study may contribute to the understanding of the motor
control mechanisms of the lumbar spine. These results may also be of use in the
design and implementation of safe work scheduling in the workplace, with the
intent of preventing injury to occupational workers. The results also suggest that
an effective lumbar belt can lend stability to the spine immediately after work,
while the system is still recovering.
66


REFERENCES
Abumi, K., Panjabi, M.M., Kramer, K.M., Duranceau, J., Oxland, T., Crisco, J.J.
(1990). Biomechanical evaluation of lumbar spinal stability after graded
facetectomies. Spine, 15(11): 1142-1147.
Adams, M.A., Dolan, P., Hutton, W.C., Porter, R. (1990) Diurnal changes in spinal
mechanics and their clinical significance. J Bone and Joint Surgery 72-B:266-270.
Adams, M.A., Hutton, W.C., and Stott, J.R (1980). The resistance to flexion of
the lumbar intervertebral joint. Spine, 5(3):245-253.
Andersson, C., Chaffin, D., Herrin, C., Mathews, L (1985). "A biomechanical
model of the lumbar spine. J Biomechanics, 18(8):571-584.
Andersson, G. B. (1981) Epidemiologic aspects on low-back pain in industry.
Spine, 53-56.
Bellemere, F. and Grassino, A. (1979). The fatigue and recovery of the human
diaphragm. Int. Congr. Electrophysiology and Kinesiology, 4,14-15.
Chu, D., LeBlanc, R., D'Ambrosia, P., D'Ambrosia, R., Baratta, R. V. and
Solomonow, M. (2003). Neuromuscular disorder in response to anterior cruciate
ligament creep. Clin Biomech (Bristol, Avon) 18:222-230
Courville, A., Sbriccoli, P., Zhou, B., Solomonow, M., Lu, Y., and Burger, E.L.
(2005). Short rest periods after static lumbar flexion are a risk factor for
cumulative low back disorder. J Electromyogr Kinesiol, 15:37-52.
Crisco, J.J. the 3rd and Panjabi, M.M.. (1991). The intersegmental and
multisegmental muscles of the lumbar spine. A biomechanical model comparing
lateral stabilizing potential. Spine, 16(7):793-799.
Crisco, J.J., Chelicani, S. Brown, R.K. and Wolfe, S.W. (1997). The effects of
exercise on ligamentous stiffness in the wrist. J Hand Surg [Am] 22:44-48.
Dickey, J. P., McNorton, S. and Potvin, J.R. (2003). Repeated spinal flexion
modulates the flexion-relaxation phenomenon. Clin Biomech (Bristol, Avon ,
18:783-789.
67


Ekstrom, L., Kaigle, A., Hult, E Holm, S., Rostedt, M. and Hansson, T. (1996).
Intervertebral disc response to cyclic loadingan animal model. Proc Inst Mech
Eng [H], 210:249-258.
Eversull, E., Solomonow, M., Zhou, B.H., Baratta, R.V. and Zhu, M.P. (2001).
Neuromuscular neutral zones sensitivity to lumbar displacement rate. Clin
Biomech (Bristol, Avon), 16:102-113.
Gedalia, U., Solomonow, M., Zhou, B.H., Baratta, R.V., Lu, Y. and Harris, M.
(1999). Biomechanics of increased exposure to lumbar injury caused by cyclic
loading. Part 2. Recovery of reflexive muscular stability with rest. Spine ,
24:2461-2467.
Givens, M.W. and Teeple, J.B.(1978). Myoelectric frequency changes in children
during static force production. Electroe Clin Neuroph, 45:173-177.
Gray's Anatomy On-line. (2000). Retrieved 2009, from Medi-Smart:
http://www.medi-smart.com/student-gray.htm
Hirsch, C., Inglemark, B., Miller, M. (1963). 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.,
33:1-17.
Hoops, H., Zhou, B., Lu, Y., Solomonow, M., and Patel, V. (2007). Short rest
between cyclic flexion periods is a risk factor for a lumbar disorder. Clin Biomech
(Bristol, Avon), 22:745-757.
Kadefors, R., Kaiser, E. and Petersen, 1. (1968). Dynamic spectrum analysis of
myo-potentials and with special reference to muscle fatigue. Electromyography,
8:39-74.
Kaigle, A.M., Holm, S.H., and Hansson, T.H. (1995). Experimental instability in
the lumbar spine. Spine, 20(4):421-430.
Kim, C. (2007). The Multifidus Muscle is the Strongest Stabilizer of the Lumbar
Spine. The Spine, 7(5):76S.
68


LaBry, R., Sbriccoli, P., Zhou, B.H. and Solomonow, M. (2004). Loner static
flextion duration elicits a neuromuscular disorder in the lumbar spine. J Applied
Physiology, 96:2005-2015.
Le, P., Solomonow, M., Zhou, B., Lu, Y. and Patel, V. (2007). Cyclic load
magnitude is a risk factor for a cumulative lower back disorder. J Occup Environ
Med, 49:375-387.
Li, L., Patel, N., Solomonow, D., Le, P., Hoops, H., Gerhardt, D., Johnson, K.,
Zhou, B., (2007). Lu, Y. and Solomonow, M. Neuromuscular response to cyclic
lumbar twisting. Human Factors, 49:820-829.
Lindstrom, L., Magnusson, R. and Petersen, I. (1970). Muscular fatigue and
action potential conduction velocity changes studied with frequency analysis of
EMG signals. Electromyography, 10:341-356.
Lu, D., Solomonow, M., Zhou, B., Baratta, R.V. and Li, L. (2004). Frequency-
dependent changes in neuromuscular responses to cyclic lumbar flexion. J
Biomech, 37:845-855.
Magora, A., Gonen, B., Eimerl, M.S. and Magora, F. (1976). Electrophysiological
manifestations of isometric contraction sustained to maximal fatigue in healthy
humans. Electromyo Clin Neuroph, 16:309-334.
Marras, W. S., Lavender, S.A., Leurgans, S.E., Fathallah, F.A., Ferguson, S.A.,
Allread, W.G. and Rajulu, S.L. (1995). Biomechanical risk factors for
occupationally related low back disorders. Ergonomics 38, 377-410.
Marras, W., Lavender, S. A., Rajulu, S. E., Allread, S. L., Fathallah, F. A., &
Ferguson, S. A. (1993). The role of dynamic three-dimensional trunk motion in
occupationally-related low back disorders. The effects of workplace factors, trunk
position, and trunk motion characteristics on risk of injury. Spine, 18:617-62.
McGill, S.M. (1997). The biomechanics of low back injury: implications on
current practice in industry and the clinic. J Biomech, 30:465-475.
NIH (2008). Retrieved from National Institute of Health, online:
http://www.nlm.nih.gov/medlineplus/backpain.html
69


NINDS (2009). Retrieved from National Institute of Neurological Disorders and
Stroke: http://www.ninds.nih.gov/disorders/backpain/detail backpain.htm
Navar, D., Zhou, B., Lu, Y. and Solomonow, M. (2006). High-repetition cyclic
loading is a risk factor for a lumbar disorder. Muscle and Nerve, 34:614-622.
Olson, M., Solomonow, M. and Li, L.(2006). Flexion-relaxation response to
gravity. J Biomechanics, 39(14):2545-2554.
Olson, M.W., Li, L. and Solomonow, M. (2004). Flexion-relaxation response to
cyclic lumbar flexion. Clin Biomech (Bristol, Avon), 19(8):769-776.
Panjabi, M. M. (1996). Low Back Pain and Spinal Stability, in Low Back Pain: A
Scientific and Clinical Overview, American Academy of Orthopedic Surgeons,
Rosemont, II, 367-384.
Panjabi, M.M. (1992). The stabilizing system of the spine. Part I. Function,
dysfunction, adaptation, and enhancement. J Spinal Disorders, 5(4):383-389,
discussion 397.
Panjabi, M.M. (1992). The stabilizing system of the spine. Part II. Neutral zone
and instability hypothesis. J Spinal Disorders, 5:390-396.
Panjabi, M.M. (2003). Clinical spinal instability and low back pain. J Electromyo
Kines, 13(4):371-379.
Posner, I., White, A., Edwards, T., and Hayes, W. (1982). A biomechanical
analysis of the lumbar and lumbosacral spine.
Punnett, L., Fine, L. J., Keyserling, W. M., Herrin, G. D., & Chaffin, D. B.
(1991). Back disorders and nonneutral trunk postures of automobile assembly
workers. Scand J Work Environ Health, 17:337-346.
Reeves, N.P., Narendra, K.S. and Cholewicki, J. (2007). Spine stability: the six
blind men and the elephant. Clin Biomech (Bristol, Avon), 22:266-274.
Rhalmi, W., Yahia, H., Newman, N., Isler, M. (1993). Immunohistochemical
study of nerves in lumbar spine ligaments. The Spine, 18(2):264-267.
70


Sbriccoli, P., Solomonow, M., Zhou, B. H., Lu, Y. and Sellards, R. (2005).
Neuromuscular response to cyclic loading of the anterior cruciate ligament. Am J
Sports Med 33:543-551.
Sbriccoli, P Solomonow, M., Zhou, B Baratta, R.V., Lu, Y., Zhu, M.P. and
Burger, E.L. (2004a). Static load magnitude is a risk factor in the development of
cumulative low back disorder. Muscle and Nerve, 29:300-308.
Sbriccoli, P., Yousuf, K., Kuperstein, 1., Solomonow, M., Zhou, B., Zhu, M.P.
and Lu, Y. (2004b). Static load repetition is a risk factor in the development of
lumbar cumulative musculoskeletal disorder. Spine, 29:2643-2653.
Silverstein, B.A., Fine, L.J. and Armstrong, T.J. (1986). Hand wrist cumulative
trauma disorders in industry. Br J Ind Med, 43:779-784.
Smit, T.H. (2002). The use of a quadruped as an in vivo model for the study of the
spinebiomechanical considerations. Eur Spine Journal, 11(2): 137-144.
Solomonow, D., Davidson, B., Zhou, B., Lu, Y., Patel, V. and Solomonow, M.
(2008). Neuromuscular Neutral Zones Response to Cyclic Lumbar Flexion. J
Biomechanics, 41:2821 -2828.
Solomonow, M. (1984). External Control of the Neuromuscular System. IEEE
Transactions on Biomedical Engineering, Vol. BME-31, No. 12:754.
Solomonow, M. (2003). Clinical spinal instability and low back pain. J
Electromyo and Kines, 13(4):371-379.
Solomonow, M. (2004). Ligaments: a source of work-related musculoskeletal
disorders. J Electromyo and Kines, 14(l):49-60.
Solomonow, M., Baratta, R.V., Banks, A., Freudenberger, C., Zhou, B. (2003).
Flexion-relaxation response to static lumbar flexion in males and females. Clin
Biomech (Bristol, Avon), 18(4):273-279.
Solomonow, M., Baratta, R.V., Zhou, B., Burger, E., Zieske, A. and Gedalia, A.
(2003b). Muscular dysfunction elicited by creep of lumbar viscoelastic tissue. J
Electromyo and Kines, 13:381-396.
71


Solomonow, M. (2003). Flexion-relaxalion response to static lumbar flexion in
males and females. Clin Biomech (Bristol, Avon), 18(4):273-279.
Solomonow, M., Eversull, E., Zhou, B., Baratta, R.V., and Zhu, M.P. (2001).
Neuromuscular neutral zones associated with viscoelastic hysteresis during cyclic
lumbar flexion. Spine, 26:E314-324.
Solomonow, M., Zhou, B., Baratta, R.V., Lu, Y. and Harris, M. (1999).
Biomechanics of increased exposure to lumbar injury caused by cyclic loading:
Part 1. Loss of reflexive muscular stabilization. Spine, 24:2426-2434.
Solomonow, M., Zhou, B., Harris, M., Lu, Y., Baratta, R.V. The ligamento-
muscular stabilizing system of the spine. Spine. 1998; 23:2552-2562.
Stubbs, M., Harris, M., Solomonow, M., Zhou, B., Lu, Y. and Baratta, R. V.
(1998). Ligamento-muscular protective reflex in the lumbar spine of the feline. J
Electromyo and Kines, 8:197-204.
Stulen, F.B. and DeLuca, C.J. (1981). Frequency parameters of the myoelectric
signal as a measure of muscle conduction velocity. IEEE Transactions Biomed
Eng, 28(7):515-523.
van Dieen, J.H., Selen, L.P. and Cholewicki, J. (2003). Trunk muscle activation in
low-back pain patients, an analysis of the literature. J Electromyo and Kines,
13:333-351.
Wilke, H.J. et al. (1995). Stability increase of the lumbar spine with different
muscle groups. A biomechanical in vitro study. Spine, 20(2): 192-198.
Xu, Y., Bach, E. and Orhede, E. (1997). Work environment and low back pain:
the influence of occupational activities. Occup Environ Med, 54:741-745.
Yahia, H., Newman, N., and Richard, C. (1988). Neurohistology of lumbar spine
ligaments. Acta Orthoo Scand, 59(5):508-512.
Youssef, J., Davidson, B., Zhou, B., Lu, Y., Solomonow, M. and Patel, V. (2008).
Neuromuscular Neutral Zones' Response to Static Lumbar Flexion. Clin Biomech,
23:870-880.
72