THE USE OF ISOMETRIC EMG AND FORCE DATA TO STUDY THE
RELATIONSHIPS BETWEEN PATHOLOGICAL INTENTION
TREMOR IN SEGMENTS OF THE UPPER EXTREMITY
by
Chen-An Chen
B.S. in Mechanical Engineering
National Chiao-Tung University, Hsin-Chu, Taiwan
(1986)
SUBMITIED TO THE DEPARTMENT OF
MECHANICAL ENGINEERING IN PARTIAL FULFILLMENT OF
THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
at the
MASSACHUSETIS INSTITUTE OF TECHNOLOGY
May 1994
© Massachusetts Institute of Technology 1994
All rights reserved.
Signature of Author
Department of Mechanical Engineering
I,
n
February 12, 1994
Certified by
.
... A
.
.
.^.1
-__
by
Accepted
Ilnahn:ee
Michael J. Rosen
esis SuPervisor
THE USE OF ISOMETRIC EMG AND FORCE DATA TO STUDY THE
RELATIONSHIPS BETWEEN PATHOLOGICAL INTENTION
TREMOR IN SEGMENTS OF THE UPPER EXTREMITY
by
Chen-An Chen
Submitted to the Department of Mechanical Engineering on May 13, 1994
in partial fulfillment of the requirements for the degree of
Master of Science in Mechanical Engineering
Abstract
The purpose of this study was to investigate the relationships between tremors at the
elbow and shoulder joints using the EMG and force data collected from isometric
experiment. The Adelstein Manipulandum, a 2-dof "virtual environment" joystick capable
of simulating different mechanical loads and obtaining positions, velocities, accelerations
and force at the end point of the joystick were used to conduct the experiment. In the
isometric experiment , the joystick was locked so that the arm posture was fixed and
constrained to the horizontal plane. Static targets were used in pursuit tracking tasks to
induce tremor at certain muscles at the elbow and/or shoulder joints. In addition to the
force data output from the manipulandum, EMG activities were monitored by three EMG
electrodes at different locations.
Ten able-bodied and three tremor-disabled subjects participated in the experiments. The
results showed that whether interactions between tremor at the elbow and the shoulder
joints occurred are difficult to determined without further inhibiting reflex loops between
muscles. The results, however, showed the promise of using the isometric experiment to
provide information for physicians to diagnose tremor and for therapists to adopt an
approach in assisting tremor-disabled patients.
Thesis Supervisor: Dr. Michael J. Rosen
Title: Principal Research Scientist
Acknowledgements
I would like to thank my thesis supervisor, Dr. Michael Rosen, for his guidance and
patience. Without his understanding, this thesis could have never been finished. I would
also like to thank the members of the tremor group -- Allison Arnold, Karen Palmer, Jack
Kotovsky; individuals in the Newman Laboratory, particularly Judy Schreidell, Pat
McCosco, Marie Stuppard, Norman Berube; the staff of the Burke Rehabilitation Center.
Karen Palmer provided many good suggestions for writing this thesis. Allison Arnold has
been a terrific friend, whom I have learned many things from, academicwise and lifewise.
Finally, I would like to thank those participant in these experiments, without whom this
work cannot be accomplished.
I would like to dedicate this volume to my wife and my parents.
-- Chen-An Chen
This work was supported by the Burke Rehabilitation Center.
Table of Contents
Page
Title Page..........................................................................
A bstract .............................................................
...........
Acknowledgements ...............................................................
Table of Contents................................................................
List of Figures...........................................................................
List of Tables ....................................................................
1
2
3
4
6
8
Chapter 1 Introduction
1.1Statement of Thesis..........................................................
9
1.2 Tremor - Definition and Classification.........................
..... 9
1.2.1 Definition..........................................................
9
1.2.2 Classification ........................ ................................ 9
1.3 Abnormal Intention Tremor...........................................
10
1.4 Tremor Mechanism.................................
.............. 10
1.4.1 Biomechanical Resonance.............................. ...... 10
1.4.2 Reflex Loop........................... . .................................. 11
1.4.3 Central Drive .................... .................................... 11
1.4.4 Tremor Mechanism Interaction..........................
. 12
1.5 Rationale for Multi-Joint Tremor Studies.............................. 12
1.5.1 Goals of Current Tremor Studies.............................. . 12
1.5.2 What is lacking in Current Tremor Experiments.................. 12
1.5.3 Significance of a Multi-Joint Tremor Experiment................ 13
1.5.4 Tremor Study at MIT Tremor Group.............................. 13
1.6 Contents of the Remainder of this Thesis............................... 14
Chapter 2 Basic Idea of the Experiment................................
2.1 Limiting the Arm to an Isometric Posture..............................
2.2 Constraining the Arm to the horizontal Plane............................
2.3 Focusing on the Elbow and Shoulder Joints...........................
2.4 Conducting Tracking Tasks to Induce Tremor..........................
2.5 Monitoring EMG Signals and Endpoint Forces.........................
15
15
15
16
16
19
Chapter 3 Apparatus.........................................
............
3.1 Overview of the Experimental Setup..................................
3.2 The Manipulandum...............................
...............
3.2.1 Moment Cancellation Mechanism..................
3.2.2 Human Arm and Manipulandum Interface .........................
3.3 The Test Chair ................................................................
3.3.1 Arm Support..................................
..............
3.3.2 Torso Support..................................
.............
3.4 Video Display and Data collection Unit.................................
3.5 Myoelectric activity Measurement Devices..............................
3.5.1 Surface Electrodes...............................
...........
3.5.2 Myoprocessors.........................
...............
21
21
21
25
29
29
30
30
32
32
32
34
Chapter 4 Experimental protocol...............................
........ 35
4.1 The Fundamentals of the MATLAB Spectrum Function.................. 35
4.1.1 Fourier Tansform, DFT, and FFT................................. 35
4.1.2 Parseval's Theorem, Power Spectrum, and Power Spectrum
Estimation........................... .....
.............. 36
4.2 The MATLAB Spectrum Function.................................
38
Chapter 5 Data Analysis Technique.......................................
39
5.1 Pursuit Tracking Tasks and the Related Experimental Variables......... 39
5.2 Other Experimental Variables............................
........ 40
5.2.1 Day-to-day Variation of Tremor................................
40
5.2.2 Fatigue................................................
40
5.2.3 Learning Effect..................................
........... 40
5.3 Experimental Design...................................
........... 41
5.3.1 Randomization and Replication.............................. 41
5.3.2 Experimental Protocol...............................
....... 42
5.4 Testing Procedure.................................
............... 42
5.4.1 Subject and Equipment Preparation..........................
. 42
5.4.2 A Typical Testing Session ...................................... . 44
5.5 Subjects..................................................
44
5.5.1 Criteria for Selecting Subjects................................
44
5.5.2 Description of Subjects..............................
....... 45
Chapter 6 Presentation of Results and discussion...........................
6.1 Results for Able-bodied Subjects..................................
6.2 Results for Disabled Subjects.............................. .......
6.2.1 Results for Subject A..........................................
6.2.2 Results for Subject B.........................................
6.2.3 Results for Subject C........................................
47
47
48
48
48
49
Chapter 7 Conclusions and recommendations for future work........... 134
7.1 Conclusions................................................................... 134
7.2 Recommendation for Future Work.................................
134
R eferences .................................
..........................................
135
List of Figures
Chapter 2
Page
Figure 2.1: Model of the arm in the experiment......................................
Figure 2.2: Minimum set of muscles required to apply a force in the
positive Y direction.......................................................
Figure 2.3: Minimum set of muscles required to apply a force in the
positive X direction............................
..............
18
18
20
Chapter 3
Figure 3.1: Schematic of isometric experiment setup................................
Figure 3.2: Photograph of the Manipulandum...............................
Figure 3.3: Photograph of the Manipulandum with the locked joystick and
the arm interface...............................
.............
Figure 3.4: Schematic of the original moment cancellation mechanism design......
Figure 3.5: Photograph of the modifications for the moment cancellation
mechanism - the cap, the extension screw element, and the collar......
Figure 3.6: Photograph showing the test chair and the torso support.............
Figure 3.7: Photograph showing the arm support.......................
22
23
24
26
28
31
31
Chapter 6
Figure 6.1:
Figure 6.2:
Figure 6.3:
Figure 6.4:
Figure 6.5:
Figure 6.6:
Figure 6.7:
Results for Subject N1A (target direction "up" and force level 10 N)...
Results for Subject N1A (target direction "up" and force level 20 N)..
Results for Subject N2A (target direction "up" and force level 10 N)..
Results for Subject N2A (target direction "up" and force level 20 N)..
Results for Subject A (target direction "up" and force level 10 N).....
Results for Subject A (target direction "up" and force level 20 N).....
Results for Subject A (target direction "up" and force level 10 N)
(minimum set of muscles were not used).............................
Figure 6.8: Results for Subject A (target direction "up" and force level 20 N)
(minimum set of muscles were not used).............................
Figure 6.9: Results for Subject B (target direction "up" and force level 10 N).....
Figure 6.10: Results for Subject B (target direction "up" and force level 20 N)...
Figure 6.11: Results for Subject B (target direction "down" and force level 10 N)
50
53
56
59
62
65
68
71
74
77
80
Figure 6.12:
Figure 6.13:
Figure 6.14:
Figure 6.15:
Figure 6.16:
Figure 6.17:
Figure 6.18:
Figure 6.19:
Figure 6.20:
Figure 6.21:
Figure 6.22:
Figure 6.23:
Figure 6.24:
Figure 6.25:
Figure 6.26:
Figure 6.27:
Figure 6.28:
Results for Subject B (target direction "down" and force level 20 N)
Results for Subject B (target direction "left" and force level 10 N)...
Results for Subject B (target direction "left" and force level 20 N)..
Results for Subject B (target direction "right" and force level 10 N).
Results for Subject B (target direction "right" and force level 20 N).
Results for Subject C (target direction "up" and force level 10 N)
(EMG data are for biceps, triceps, and anterior deltoid)..........
Results for Subject C(target direction "up" and force level 10 N)
(EMG data are for biceps, triceps, and posterior deltoid).........
Results for Subject C (target direction "up" and force level 20 N)
(EMG data are for biceps, triceps, and anterior deltoid)...........
Results for Subject C(target direction "up" and force level 20 N)
(EMG data are for biceps, triceps, and posterior deltoid).........
Results for Subject C (target direction "down" and force level 10 N)
Results for Subject C(target direction "down" and force level 20 N).
Results for Subject C (target direction "left" and force level 10 N)
(EMG data are for biceps, triceps, and anterior deltoid)...........
Results for Subject C(target direction "up" and force level 10 N)
(EMG data are for biceps, triceps, and posterior deltoid).........
Results for Subject C (target direction "left" and force level 20 N)
(EMG data are for biceps, triceps, and anterior deltoid)..........
Results for Subject C(target direction "up" and force level 20 N)
(EMG data are for biceps, triceps, and posterior deltoid).........
Results for Subject C (target direction "right" and force level 10 N).
Results for Subject C(target direction "right" and force level 20 N)..
83
86
89
92
95
98
101
104
107
110
113
116
119
122
125
128
131
List of Tables
Chapter 2
Page
Table 2.1: Relationships between the force directions and the corresponding
minimum set of muscles required to perform the tracking tasks....... 20
Chapter 5
Table 5.1: Typical experimental session......................................
43
Chapter 1
Introduction
1.1 Statement of Thesis
The objective of this study is to investigate the possible neurological or biomechanical
coupling of tremors at the shoulder and elbow joints of people with abnormal intention tremor
using isometric force and EMG data.
1.2 Tremor - Definition and Classification
1.2.1 Definition
Tremor is defined as involuntary oscillations that are superimposed upon the volitional
force or kinematic output of any limb segment or body part. This definition is chosen to
distinguish pathological tremor from other hyperkinesias (Jankovic and Fahn 1980).
1.2.2 Classification
Existing schemes for the clinical classification of tremor are too simplistic to encompass the
indisputable complexities of all tremors. These classification schemes emphasize one or more
of the following criteria: (1) the behavioral circumstances of the tremor occurrence (i.e. rest,
postural, and intention tremor); (2) "physiological" versus "pathological"; (3) the amplitude and
frequency of the tremor (i.e., kinematic data); (4) topographic distribution (e.g., face, lower
extremities); (5) associated disease states (Brumlik and Yap 1970; Findley and Capildeol984).
These clinical classification schemes are fraut with ambiguities and are little more than an
empirical lexicon for communication among medical professionals.
Tremor is commonly classified according to the behavioral circumstances of the occurrence
of tremor. Based on this classification scheme, tremor is divided into four major categories:
rest tremor, postural tremor, action tremor, and intention tremor. Rest tremor refers to tremor
that is present when the affected body part is in repose and supported externally against gravity
so that no muscle contractions are required. Posturaltremor refers to tremor that is present
9
when the posture of the affected body part must be maintained in a gravity field or against a
force. Common examples of such postural activity include sustaining the arms horizontally in
front of the body and keeping the head and body erect. Action tremor refers to tremor present
during voluntary locomotor activity, such as when moving a limb. Intention tremor refers to
an exaggeration of the action tremor when the limb is reaching for a target, and is most
pronounced on approaching the destination (Findley and Capildeol984). Intention tremor
interferes with intended motor tasks and impedes daily living activities, and is thus the most
disabling form of tremor. Therefore, this experimental investigation was focused on intention
tremor as to extend a history of studies by the tremor group at MIT.
1.3 Abnormal Intention Tremor
Intention tremor is commonly associated with damage to the cerebellum (cerebellar tremor)
or the midbrain (rubral tremor). The damage may be caused by stroke, head trauma, tumor,
demyelination, or neurological degeneration due to chronic alcohol intoxication or metabolic
poisoning. It is estimated that approximately 50 percent of the one million tremor population in
the U.S. has intention tremor and the most common cause of cerebellar intention tremor is
multiple sclerosis (Elble & Koller 1990, Chen 1991), the cause of which is still unknown.
1.4 Tremor Mechanisms
Three hypothesized tremogenic mechanisms have been proposed. They are biomechanical
resonance, reflex loop instability, and central drive. These individual mechanisms and the
possible interaction among them are described in this section.
1.4.1 Biomechanical Resonance
A human limb, by virtue of its bones, muscles, tendon, etc., possesses inertia. Muscles,
even in a relaxed state, produce stiffness about the joint. Therefore, a limb acts as a passive
mass-spring system, which has a natural frequency at which it oscillates. If external loading as
well as the mechanical properties of the limb are considered, the natural frequency WOn is
10
determined approximately by fiim, where k is the equivalent joint stiffness and m is the
combined limb-load inertia. The biomechanical resonance model states that the combined limb
and external load have the characteristics of an underdamped second order system. This
passive system can oscillate when perturbed by an external or internal force such as heart beat,
or noisy muscle force.
1.4.2 Reflex loop
In an intact limb, a number of different types of sensory receptors contribute afferent
activity to the 'functional stretch reflex' (Marsden et al., 1971; Houk et al., 1970), and there
may be a number of different reflex pathways with different time delays (Hammond et al.,
1956; Phillips, 1969; Evarts and Tanji, 1974; Lee and Tatton, 1975). Any active reflex will
have some filtering effect on tremor, though different pathways with their different delays will
enhance different tremor frequencies, and their effects may sometimes cancel each other out
(Stein and Oguztoreli, 1976) A reflex may, if it is powerful enough, actually generate tremor
by its own activity. In these situations, the movement of a joint in one direction evokes a reflex
in the extended muscle which is sufficiently powerful to initiate an opposite shortening
movement; the shortening movement then leads to a pause in the muscle's reflex discharge and
subsequent re-extension, and so the reciprocating movement continues at a frequency
determined by the the conduction and activation delays in the reflex pathways. Tremors at
different joints may also interact with each other through reflex pathways.
1.4.3 Central Drive
The central oscillation hypothesis proposes that oscillatory sources in the central nervous
system - unaffected by short-term peripheral feedback from muscle or other internal
transducers - simply entrain groups of spinal neurons to send synchronous signals. These
central sources include brain areas responsible for the coordination of movement and regions
in the spinal cord. For example, typical cerebellar intention tremor has bee produced by
cerebellectomy (Gilman et al., 1976), by selective stereotatic lesions involving the dentate
nucleus (Growdon et al., 1967), by reversible cooling of the deep cerebellar nuclei (Brooks et
11
al., 1973; Cooke and Thomas, 1976; Villis and Hore, 1977), by lesions of the red nucleus
(Ranaish and Soechting, 1976), or by damage to the superior cerebellar peduncle.
1.4.4 Tremor Mechanism Interaction
The basic mechanisms of tremor are not mutually exclusive and never exist in isolation.
Indeed, it has been convincingly argued that tremors result from the interaction of the
segmental stretch reflex and the muscle mechanical mechanism (Joyce and Rack, 1974; Stein
and Oguztoreli, 1976), and it has even been proposed that a transcortical stretch reflex
contributes to tremor (Milner-Brown et al., 1975; Stein and Bawa, 1976).
1.5 Rationale for Multi-Joint Tremor Studies
1.5.1 Goals of Current Tremor Studies
Although tremor research has grown exponentially over the last three decades, diagnosis
remains difficult because tremor etiologies, underlying mechanisms, and characteristics are not
consistently related. The major methods of treatment for tremor include medication, surgery,
and biofeedback, none of which give uniform results. Medication is the most effective
treatment; however, determining the proper drug and dosage is often a trial-and-error process
(Rondot et al 1978, Shahini and Young 1978). Treatment by this method is time consuming
and requires the patience of both the physician and the patient. Researchers agree that a
complete understanding of the tremogenic mechanisms that drive pathological tremor is
necessary for a more effective treatment. Therefore, the goals of current tremor studies among
researchers are: (1) to understand the mechanisms of tremor to aid diagnosis and treatment and;
(2) to provide useful information for designing a whole-arm tremor suppression orthosis.
1.5.2 What is Lacking in Current Tremor Experiments
To date, only a few researchers have attempted to record tremor in multi-dimensional
space. Sailzer (1972) measured the hand tremor with a three-dimensional accelerometer in
order to find axes in which the tremor amplitudes are smallest or greatest. Frost (1978) also
developed a triaxial vector accelerometry system for quantitatively assessing tremor and ataxia.
12
Elble et al (1990) apllied a commercially-available digitizing tablet and personal computert to
quatify tremor. Arblaster et al (1990) recorded wrist tremor bilaterally and simultaneously .
Will et al (1990) employed the VPL data glove to quntatatively analyzed tremor. However,
simultaneous measurements of tremors at different joints in a limb have not been reported.
There are two reasons why such experiments have not been done. First, systems capable of
recording tremor at multiple joints generally have greater complexity than one-dimensional
systems due to. Second, analyzing tremor motion in multidimensional space is highly complex
when the kinematics of an entire multi-segmented limb is considered.
1.5.3 Significance of a Multi-Joint Tremor Experiment
The value of a multi-joint tremor experiment lies in an enhanced understanding of tremor
such as whether tremors present at different locations interact with each other, and whether the
tremors are coupled, and the nature of the coupling (mechanical or neural). This understanding
will improve our knowledge of tremor mechanisms and will therefore add to the diagnosis and
treatment of tremor. It will also aid the development more accurate tremogenic models, which
will contribute to the design of assistive devices.
1.5.4 Studies of the MIT Tremor Group
Tremor studies at MIT have had a three-fold purpose: 1)to quantify tremor characteristics,
2) to develop models of tremor mechanisms, and 3) to establish guidelines for the design and
development of assistive devices for people with tremor. There are currently five devices
developed by the group which can be used to analyze tremor. The Adelstein Manipulandum, a
2-dof "virtual environment' joystick capable of simulating different mechanical loads and
obtaining signals for positions, velocities, accelerations and forces at the endpoint (Adelstein
1989) was used to conduct the experiment. A detailed description of the Manipulandum will be
presented in Chapter Three.
13
1.6 Contents of the Remainder of this Thesis
Chapter Two discusses how the basic idea of the experiment were formulated and Chapter
Three gives a description of the experimental apparatus. Data analysis methods are stated in
Chapter Four while experimental protocol and subject selection are dealt with in the fifth
chapter. In Chapter Six, the results are presented and their relevance is discussed. The
conclusions and recommendations for future work are treated in the seventh chapter.
14
Chapter 2
Basic Idea of the Experiment
This experiment is the first known to explore the relationship between tremors at different
joints of the upper extremity. Therefore our primary goal was to keep the experiment simple,
yet powerful enough to permit conclusions about possible coupling. In addition to this, the
experimental apparatus, the Manipulandum, also imposes restrictions on the experimental
design. The desire for simplicity and the restriction imposed by the Manipulandum resulted in
an experiment which keeps the arm posture fixed, constrains the arm to the horizontal plane,
and focuses on the elbow and shoulder joints.
2.1 Limiting the Arm to an Isometric Posture
As mentioned in Section 1.5.2, one of the difficulties encountered in a multi-joint tremor
experiment is in motion analysis. To avoid complicated motion, the arm posture isometric
(i.e., fixed) during the experiment (muscles are not truly isometric, so stretch reflexes still play
a part). With the arm posture fixed, the dynamic effect of the arm can be neglected. Instead of
positions, velocities or accelerations, forces are the concern in an isometric experiment.
2.2 Constraining the Arm to the Horizontal Plane
Although free-motion experiments will be included in future work, only isometric tests
were conducted in the work reported here. In the free-motion experiments, the arm will be
constrained to the horizontal plane to eliminate gravity effects. Hence, the arm is positioned in
the horizontal plane in this study in order to permit comparison of the results of the isometric
tests with our future free-motion experiments. Also the force transducer of the Manipulandum
(described in Chapter 3) was built specifically to measure force in the horizontal plane and
therefore placing the arm in the horizontal plane would eliminate the need to measure force in
the vertical degree-of-freedom.
15
2.3 Focusing on the Elbow and Shoulder Joints
The Manipulandum can measure force in only two degrees of freedom in the horizontal
plane. Due to this constraint, tremors can only be investigated at two adjacent joints. We focus
on the elbow and shoulder joints in this study. There are two reasons for choosing these two
joints instead of the wrist and elbow joints. First, an interface which couples the hand to the
joystick is difficult to design, whereas it is relatively easy to design an interface for the
forearm. Second, to separate tremor of one joint from that of the other, it is required that two
adjacent limbs be orthogonal to each other. To meet this criterion, the arm has to be posed so
that the hand is perpendicular to the forearm (if the wrist and elbow joints are chosen ) or the
forearm perpendicular to the upper arm (if the elbow and shoulder joint are chosen). The
former can not be obtained easily because of the nature of the wrist joint; the latter, however,
can be accomplished without difficulty.
2.4 Conducting Tracking Tasks to Induce Tremor
To induce intention tremor, tasks requiring voluntary muscle contractions and target
pursuit must be performed by subjects. Tracking tasks are used to serve this purpose. The
tracking tasks should be designed to induce intention tremor in particular muscles so that
the possible coupling between tremors at selected muscles and other muscles can be
investigated. A posture shown in Figure 1 is adopted to achieve this goal. With this
posture, the minimum set of muscles required to perform the tracking tasks can be
determined by the direction in which the subject has to apply a force, thus eliciting intention
tremor at those muscles.
In this study, the forearm was attached to the endpoint of the Manipulandum's joystick
which acts in the same way as a simple bearing. This endpoint is fixed in position but is
free to rotate in the horizontal plane. If no muscle contractions occur around the elbow and
shoulder joints, their behavior can be considered similar in nature to simple bearings since
the joint themselves and the connective tissues around them can resist force (if the force is
16
not too large) but not moment. With these features, the arm in the experiment can be
modeled as shown in Figure 1.
If a force applied by the subject is seen in the positive Y direction at the endpoint of the
joystick, as shown in Figure 2(a), the minimum set of muscles necessary to produce this
force are pectoralis major and anterior deltoid. The reason is stated as follows: If the
muscles around the elbow joint are relaxed, the force can be transmitted to the forearm only
through the elbow joint and the force transmitted should be in the positive Y direction as
shown in Figure 2(b). Therefore the reaction force to the upper arm at the elbow is in the
negative Y direction. For the upper arm to be in equilibrium, the resultant force produced
by muscles Fm and that by passive connected tissues around the shoulder joint Fs must be
equal to the reaction force Fr , and the resultant moment by all forces must be equal to zero.
The minimum set of muscles needed to generate Fm therefore are pectoralis and anterior
deltoid since the passive connective tissues provide the required force to balance the upper
arm and muscles such as deltoid posterior side are not necessary. Similarly, if a force is
applied in the negative Y direction, the minimum set of muscles needed would be teres
major, teres minor, and posterior deltoid posterior. If a force is sensed in the positive X
direction, the minimum set of muscles are triceps and deltoid. In this case, the triceps have
to contract to generate a force in this direction, as shown in Figure 3(a). Consider the force
equilibrium of the upper arm (see Figure 3(b)). Although the passive tissues around the
shoulder joint contribute the force balance, they are not strong enough to hold the shoulder
joint, it was found from preliminary experiments done by the author that both anterior
deltoid and posterior deltoid were necessary to stabilize the shoulder joint. The minimum
set of muscles then are triceps and deltoid . For the same reason, when a force in the
negative X direction is applied, the minimum set of muscles are triceps and deltoid. The
relationships between the force direction and the minimum sets of muscles are summarized
in Table 2.1.
17
Figure 2.1 Model of the arm in the experiment.
I
Fmy
Anterior
deltoid
Fmx
AI
I
Anterior deltoid
FSX
(b)
(a)
Fsy
Figure 2.2 Minimum set of muscles required to apply a force in the positive Y
direction: (a) simplified muscular and skeletal model for the arm in equilibrium;
(b) the upper arm in equilibrium, where Fr is the reaction force at the elbow, Fmx
and Fmy respectively are the X and Y component of the resultant force generated
by the minimum set of muscles, and Fsx and Fsy respectively are the X and Y
component of the force produced by the connective tissues and articular surfaces
around the shoulder.
18
2.5 Monitoring EMG Signals and Endpont Forces
Preliminary experiments also suggest that force data alone does not provide enough
information to study tremor couplings. Therefore, it was decided to monitor surface EMG
signals in addition to endpoint forces. EMG stands for electromyograph. EMG signals are
the spatially summed electrical activity of the motor units. A surface EMG signal is
detected at the surface of the skin ( J.V. Basmajian and C.J. De Luca, 1985). Because
only three myoprocessors were available at the time of the experiment, EMG signals from
three muscles were selected for every singe test despite although there were four muscles
needed to be monitored. Regardless of this, EMG signals form all four muscles were
available for analysis for every experiment condition even though they were collected from
different runs.
19
*
.Fmax
e
Fey
t
Anterior deltoid
Fmay
F atF_
I
IFe
L
-
ALI
(a) Posterior deltoid
Fex
ex~e
SFmpyj
(b)
Triceps
S
Fpx
Ftx
Fty
Figure 2.3 The minimum set of muscles required to apply a force in the positive X
direction: (a) simplified muscular and skeletal model for the arm in equilibrium; (b) the
upper arm in equilibrium, where Fe, Ft , Fma, and Fmp respectively are the reaction
force at the elbow, the force generated by the triceps, the resultant force produced by
the muscles on the anterior side, the resultant force produced by the muscles on the
posterior side. The subscript x and y represent the x and y components of these forces.
on the anterior side, the resultant force of the mus
Applied Force Direction
Minimum Set of Muscles Required
Positive Y
Pectoralis Major, Deltoid Anterior Side
Negative Y
Teres Major, Teres Minor, Deltoid Posterior Side
Positive X
Triceps, Deltoid
Negative X
Biceps, Deltoid
Table 2.1 The relationships between the force directions and the corresponding minimum
set of muscles required to perform the tacking tasks.
20
Chapter 3
Apparatus
3.1 Overview of the Experimental Setup
The entire experimental setup is composed of four major parts - the Manipulandum, the test
chair, the video display and data collection unit, and the myoelectric activity measurement
devices. The Manipulandum and the modifications made to the Manipulandum by the author is
described in Sec 3.2. The test chair is depicted in Section 3.3. The video and data collection
unit is explained in Section 3.4. In Section 3.5, the features of the myoelectric activity
measurement devices are illustrated.The experimental setup is shown schematically in Figure
3.1. During the experiment, the subject sit on the test chair, the forearm was coupled to the
Manipulandum, and three electrodes were applied at appropriate locations. The EMG signals
from the electrodes were received and processed by the myoprocessors. The force data from
the Manipulandum and the processed EMG signals were connected to the data collection unit,
which converted the analog signals to digital data. The converted data were then stored and the
force data was also used for the video display.
3.2 The Manipulandum
As mentioned in Chapter 1, the Manipulandum, shown in Figure 3.2, is a 2-dof "virtual
environment" joystick. The joystick is coupled to two brush-type DC motors through a gimbal
mechanism. Angular displacements and velocities of the motor shafts are measured by
encoders and tachometers mounted on each shaft. Angular accelerations are sensed by two
translational accelerometers attached to two horizontal pulleys driven by the two motor shafts
respectively. The combination of a miniature force transducer and a moment cancellation
mechanism allow the measurement of the force perpendicular to the joystick applied at the
endpoint of the joystick (Adelstein, 1989). A hard-wired electronic feedback controller
21
Display and A/D conversion
Myoprocessors
Figure 3.1 The isometric experiment setup.
22
Figure 3.2 Manipulandum.
23
- r
Figure 3.3 Manipulandum with the locked joystick and the arm interface.
24
(Control Interface Unit) was also developed by Adelstein to implement mechanical loads such
as simple springs, dampers, and inertias. The simulated loads can be tuned through a
computer. However, the load simulation capability was not used in this study. The joystick
was mechanically locked by a fixture as shown in Figure 3.3 to ensure the isometric condition.
Thus, the Manipulandum was used merely as a force measurement device.
Two modification of the Manipulandum were necessary to meet our experimental
requirements. Preliminary experiments indicated that the original moment cancellation design
allowed too much play to maintain isometric conditions. As a result, the moment cancellation
mechanism was modified to ensure isometric conditions. An interface coupling the forearm the
joystick was also made since the Manipulandum system did not include such an interface.
3.2.1 Moment Cancellation Mechanism
The moment cancellation mechanism was developed mainly because the force transducer is
incapable of distinguishing between forces and moments. The moment cancellation mechanism
counteracts the force transducer's moment sensitivity and allows the force transducer to
measure forces only.
The original moment cancellation design (Adelstein, 1989) is shown in Figure 3.4. The
mechanism is based on a combined spherical bearing and thrust bearing arrangement. An
adapter sleeve is pressed into the ball of the spherical bearing and then fitted onto the force
transducer's beam element. The spherical bearing race is then joined through a collar to a flat
thrust washer which is held between two retainers. The retainers are held in turn between two
flat thrust washers. The handgrip slips over the top end of the spherical bearing race collar and
is kept in place by circumferential set screws. Vertical preload of the stacked thrust bearing is
provided by the top cover of the decoupler housing. The moment cancellation mechanism
works because the spherical bearing prevents bending moments about a horizontal axis and
twist of the handgrip about the vertical axis from being transmitted to the transducer beam.
25
SPHERICAL
HANDGRIP
BEARING RACE COLLAR
•E'OUPLER HOUSING
SzHERICAL
BEARING
THRUST WASHER
THRUST WASHER
THRUST WASHER
SPHERICAL BEARING
1 THRUST
BEARING
STACK
FORCE JOYSTICK BEAM
DECOUPLER
)RCE JOYSTICK TOP COVER
ANDLE SHAFT
Figure 3.4 The original moment cancellation design.
26
Moments about any axes perpendicular to the shaft direction tl as defined in Figure 3.4 are
opposed by reaction forces in the thrust bearing stack, thus allowing them to be absorbed by
the mechanism housing.
It was noticed that the original design allowed too much play in the handgrip of the joystick
and could not sustain the isometric condition when large forces were applied. Investigations
revealed that the play primarily came from two sources - the spherical bearing and the
combination of the middle thrust washer and the collar. The use of the spherical bearing
introduces both friction and backlash. The backlash in turn caused the play in the joystick.
Although the backlash can be reduced by increasing the bearing preload, this increases the
friction. Since the original mechanism was primarily designed for free-motion experiments,
the friction had been more of a concern than the play and the mechanism had been adjusted so
that no significant friction was perceivable, however, at the expense of a significant play.
Moreover, the adjustment was permanent; no further adjustment was allowed without changing
the original design. The middle thrust washer was made of stainless steel and was press-fitted
onto the the aluminum collar. It was found that the steel thrust washer deformed the aluminum
collar when the applied forces were too large The deformation resulted in the tilting of the
collar with respect to the thrust washer, which caused the play.
Several modifications were completed to reduce the play. First, the middle thrust washer
and the collar were made into one piece of stainless steel as shown in Figure 3.5. This
modification prevents the collar from tilting with respect to the thrust washer. Second, an
extension screw element, which allows the adjustment of the preload to the spherical bearing,
was designed to reduce the backlash in the spherical bearing. The element, shown in Figure 8,
is actually a screw with a surface which matches the top of the spherical bearing race at one end
and a slot at the other. The threaded part of this element was screwed into the threaded hole of
the collar. The preload to the spherical bearing can be adjusted by turning the screw element
with a screw driver from the top of the collar. Extending the element to the spherical bearing
increases the preload, while withdrawing the element decreases the preload. Third, a threaded
27
Figure 3.5 Modifications made for the moment cancellation mechanism - the cap, the
extention screw element, and the collar.
28
cap, shown in Figure 3.5, was used to stop the thrust bearing, i.e., to resist the moment
applied to the moment cancellation mechanism. Turning the cap downward loads the thrust
bearing. This design allowed the adjustment of the load to the thrust bearing.
3.2.2 Human Arm and Manipulandum Interface
An interface between the human arm and the Manipulandum joystick was built to conduct
the isometric experiment. The interface is shown in Figure 3.3. The interface design issues
were: 1)the interface should be capable of transmitting the applied force to the force transducer
and, 2) it should shield the human against any possible damage by the metal parts of the
joystick and the interface. This is of particular concern for people with tremor, since they have
little control over their tremor.
This interface is composed of three parts - the support component, the force-transmission
component, and the cuff. The support component is fastened with screws to the force
transmission component and is coupled to the joystick by screwing the collar of the moment
cancellation mechanism into the hole of the support componenet. The force-transmission
component is an U-shaped aluminum plate. A commercial hand cuff, which is made of
thermoplast on the outside and of foam on the inside was attached to the bottom of the Ushaped component. The cuff is used to protect the human arm against metal parts. Since the
cuff is not stiff, two screws are needed to joined the cuff to each side of the force transmission
component in order to transmit forces to the joystick and to maintain isometric conditions (see
Figure 3.3).
3.3 The Test Chair
A test chair made of Dexion was built for the isometric experiment because standard chairs
are not high enough for the subject to reach the joystick with the desired posture. An arm
support and a torso support were also mounted on the test chair. A torso support was also
included as part of the test chair. The arm support and the torso support are described in the
following two sections. The test chair is shown in Figure 3.6.
29
3.3.1 Arm Support
It was realized from the preliminary experiments that an arm support was needed so that
the subject did not have to use muscles which were not necessary to perform the tracking tasks
in order to maintain the arm in the horizontal plane against gravity. The arm support, as shown
in Figure 3.7, is attached to the upper arm by an adjustable cuff with loop-and-hook strips.
The cuff sits on a aluminum plate which is free to rotate on a simple bearing joint. Two other
horizontal links and ball bearings provide the required flexibility to allow for supporting the
arm in any place in the horizontal plane. The height of the arm support is adjustable to allow
for different subject heights.
3.3.2 Torso Support
A torso support, as shown in Figure 3.6, was designed to limit the lateral motion of the
trunk. It consists of two 6-inch L-shaped pieces of moldable plastic sheets (Multiform), which
snugly fit the torso and medial torso. The medial surface is padded with 1/4-inch thick foam.
The Multiform is attached to the side rails by four screws with wing nuts to adjust the width of
the torso support. The torso support reduces the lateral motion of the trunk significantly within
the range of a centimeter.
30
Figure 3.6 Manipulandum in use (showing the test chair and the torso support).
Figure 3.7 Manipulandum in use (showing the arm support).
31
3.4 Video Display and Data Collection Unit
A computer was used for data collection and video display. The computer includes a
MetroByte DAS-8 board, which is a 12-bit analog-to-digital conversion board. The board was
set to convert voltages between -5 and +5 to 12-bit binary numbers. Two force signals from
the Manipulandum and three EMG signals from the myoprocessors were sampled at the rate of
60 Hz and were converted to digital numbers. The sampling rate was chosen based on the
speed of the computer and the tremor frequencies, which will be explained more in Chapter 4.
The digital data were stored on the hard drive of the computer for data analysis. The force data
were also used for controlling the response cursor in the tracking tasks. The response cursor
was programmed so that the movement of the cursor was proportional to the force applied by
the subject; a force in the positive X and Y directions corresponded to a rightward and a
upward displacements of the cursor respectively.
3.5 Myoelectric Activity Measurement Devices
3.5.1 Surface Electrodes
Active bipolar surface electrodes (Boston Elbow MY0111 Electrode) were used to
detect myoelectric activities in the experiment. Active surface electrodes were preferable
not only because they provide EMG signal of greater fidelity, but also because they are
convenient to use (John Basmajian and Carlo Deluca, 1985). The bipolar configuration
eliminates unwanted electric signals from sources other than the muscle being investigated.
3.5.2 Myoprocessors
Three myoprocessors designed by Prof. Neville Hogan at MIT and modified by Ted
Clancy were used in this experiment, all of which accept as input a single electrode
connected via a Lemo connector. The myoprocessor is shown in Figure 12. They supply
the electrode with +/- 9v for power. Each myoprocessor has three outputs: 1) RAW - the
electrode EMG amplified via gain controlled by a pot on the myoprocessor housing; 2)
32
FILTER - the electrode EMG amplified high pass filtered at approximately 4-7 Hz, and 3)
SMOOTH - the electrode EMG amplified, high pass filtered at approximately 4-7 Hz, full
wave rectified (all values greater than zero), and low pass filter smoothed at a frequency of
approximately 2 Hz. The SMOOTH outputs were used in this study because the high pass
filter eliminated low frequency noise and the low pass filter avoided aliasing which might
occur due to sampling(more detail in Chapter 4). In general, the tremor frequencies are
between 2 and 6 HZ. The cut off frequencies of the high pass filter and the low pass filter
are within the range of tremor; however, those filter are only first-order, which means
tremor signals would not be significantly diminished. Moreover, the main concern for the
study was to find whether certain muscles contracted during the experiment, and the
relative magnitude of the EMG signals were not important.
During the experiments, two of the three myoprocessors were placed in a basket
attached to the test chair and the other was taped to the test chair.
33
Chapter 4
Data Analysis Techniques
Since we are interested in the tremor frequencies of sampled data and the related power
densities, the data were transformed into the frequency domain using the MATLAB Spectrum
function. The fundamentals of the MATLAB Spectrum function are reviewed in Section 4.1.
The MATLAB Spectrum function itself is discussed in Section 4.2.
4.1 The Fundamentals of the MATLAB Spectrum Function
The MATLAB Spectrum function.performs FFT (Fast Fourier Transform) analysis using
the Welch method of power spectrum estimation. FFT is, in fact, an algorithm for computing
DFT (Discrete Fourier Transform) of a sampled signal . DFT is, in turn, the discrete version
of the Fourier transform. The Fourier transform, DFT, and FFT are reviewed in Section 4.1.1
to provide the basics to understand the MATLAB Spectrum function. The definition of the
power spectrum is obtained from the Parseval's theorem. They are introduced along with the
methods of power spectrum estimation in Section 4.1.2.
4.1.1 FourierTransform, DFT,and FFT
The Fourier transform of a signal f(t) is defined as
F(jo) = f(t)e-jO dt
(4.1)
The Fourier transform is considered an operation that creates from f(t) a function F(jo) in the
frequency domain, where the frequency content of f(t) appears explicitly. In this investigation,
we are concerned with sampled data; therefore, the discrete version of the Fourier transform,
DFIT, should be applied and is defined as:
F(jO
fncei
=
n=-o
(4.2)
34
Any actual computation of the discrete Fourier transform must of course involve a finite
summation of terms rather than infinite sum. In practice, signals are truncated in various ways
to obtain a finite sample set. Suppose that there are altogether N samples of f(t). Eq.4.2 then
becomes
N-I
F(jo) = C
fneiT
n=o
(4.3)
Although the frequency o in Eq.4.3 is a continuous variable, only N values (real and
imaginary) of F(jO) can actually be independent. Furthermore, since F(jO) is a periodic
function, the independent values of f(jo) should be computed within one of its periods, say
from oT = 0 to 2n. Therfore, letting
S2m.
NT '
m = 0, 1..., N-1
(4.4)
The usual formal for computing the independent values of DFT becomes
N-i
fne j (
)=
Fm = F
mnN )
n=O
(4.5)
The fast Fourier transform (FFT) is not a new type of transformation. It is instead an
algorithm for computing DFIT. It eliminates most of the repetition in the DFT formula and
allows a much more rapid computation of the DFT. The FFT also generally allows a more
accurate computation of the DFT by reducing round-off errors. People who are interested in
the algorithm of FFT should refer to references in signal analysis
4.1.2 Parseval's Theorem, Power Spectrum, and Power Spectrum Estimation
In signal analysis, the total power, P, in f(t), is usually defined as
P=
(
f2(t)dt
(4.6)
The use of "power" is a generalization of the physical meaning of the term. If f(t) is an
electrical potential or current measured in volts or amperes, then f2(t) is the instantaneous
power in watts dissipated in a 1_5 resistor "fed" by f(t). The Parseval's theorem states that
35
If2(t)dt = 12 IF(jc)
1do
2H
2
(4.7)
which means that the total power in a signal is the same whether it is computed in the time
domain or in the frequency domain. From Parseval's theorem, it is very clear that IF(j) 12 is
the power spectrum. The discrete form of Parseval's theorem states that
N-1
Total Power =-1N 11 1-J =
m=o
Again, from the above statement, we know that IF
12
N-1
N2 n=O
if~1
(4.8)
represents the power spectrum of the
discrete signal.
Estimation of the power spectrum is the central objective of signal processing of which the
purpose is to obtain unbiased consistent estimation. There are two basic approaches to estimate
the power spectrum. One approach is referred as periodogram analysis and is based on direct
Fourier transformation of finite-length segments of the signal. the second approach is to first
estimate the autocovariance sequence and then compute the Fourier transform of this estimate.
Only the periodogram method is discussed here since it is the method adopted by the MATLAB
Spectrum Function.
For an N-point data sequence [fn], the perodogram is defined as
INO 1
N
N(j0)12
(4.9)
where
XN(jO)
N-1
= C fne-jncT
n=O
(4.10)
The periodogram is equivalent to the Fourier transform of a truncated autocorrelation estimate.
Although the periodogram can sometimes serve as a useful estimate of the power spectrum, it
is not a consistent estimate. Two common methods are used to reduce the variance. The first
method is to average the estimate over several independent trials. The second method is
achieved by windowing the signal with proper window sequence. The method employed in
the MATLAB Spectrum function is called the Welch method, which combines the concepts of
36
averaging and windowing into a single approach. The data is partitioned into K blocks of
length M. Then, the window function is applied directly to each data block.
4.2 The MATLAB Spectrum Function
As stated before, the Spectrum function performs the FFT analysis of two sequences using
the the Welch method of power spectrum estimation. The two sequences of N points are
divided into k sections of M points each (M must be a power of two). Using an M-point FFT,
successive sections are Hanning windowed, FFT'd and accumulated. The format of the
function is as follows:
Function P = SPECTRUM(X,Y,M,NOVERLAP)
where M and NOVERLAP specifies that the M-point sections should overlap NOVERLAP
points. SPECTRUM returns the M/2 by 8 array
P = [Pxx Pyy Pxy Txy Cxy Pxxc Pyyc Pxyc]
where
Pxx = X-vector power spectral density
Pyy = Y-vector power spectral density
Pxy = Cross spectral density
Txy = Complex transfer function from X to Y
Cxy = Coherence function between X and Y
Pxxc,Pyyc,Pxyc = Confidence range (95 percent).
In this investigation, 4096-points data were input to the fuction and they are divided into 16
256-point sections. No overlap was used here because it did not show significat improvement
of the power spectrum estimation
37
Chapter 5
Subjects and Experimental Protocol
This chapter deals with the experimental protocol of this study. The targets used in the
tracking tasks and the related experimental variables are discussed in Section 5.1. Section 5.2
illustrates other variables such as day-to-day variation of tremor, fatigue, and learning effect.
Control of experimental variables through experimental design and the resulting experimental
protocol are presented in Section 5.3. In Section 5.4, we describe in detail the subject and
equipment preparation and testing procedure. Specifications for selecting subjects and a
description of them are given in Section 5.5.
5.1 Pursuit Tracking Tasks and the Related Experimental Variables
Pursuit tracking tasks were used to induce intention tremor in this study. Experimental
variables such as the muscles and the force levels required to perform the tracking tasks were
controlled by managing the targets employed in the tracking tasks. The key factors include the
target type, directions of movement, and the distance the target travels. The effects these
variable have on the experiment are discussed in this section.
Selection of the Target Type
The choice of the target type was made between a moving target and a static target. The
static target was chosen in this experiment for two reasons. First, the static target is simpler
than a moving target. Factors such as the target speed and the target frequency do not have to
be taken into account. Second, a static target requires the subject to apply a constant force,
which means that during a trial the firings of the same motor units occur. Activations of the
same motor units not only make the collection of the EMG signals easier but also provide more
statistically reliable data.
Target Directions
By changing the target direction, the muscles the subject uses during the tracking task may
be controlled. In this experiment, the target either moves up, down, left, or right on the screen
38
during a trial, which correspond respectively to a force in the Y, negative Y, negative X, or X
direction. The particular muscles used to apply a force in each of those directions are
mentioned in Chapter 2.
Target Travelling Distance
The force level the subject must exert during tracking is indicated by the distance the
target travels from its home position. Although the force level is an important variable in
any tremor experiment, it was not the main goal of this experiment to explore how force
levels influence tremors. Therefore, only two force levels were used in this experiment.
The force levels were chosen based on the results of the preliminary experiment, and
correspond to 10 N and 20 N.
5.2 Other Experimental Variables
5.2.1 Day-to-day Variation of Tremor
Tremors vary from day to day with factors such as patient's physiological or psychological
state, the patient's drug, caffeine, or tobacco intake, and the degree of general motor activities
that has proceeded the testing of tremor (Potvin et al 1975), some of which are beyond the
control of the experimenter and the subject. To eliminate this effect, the experiment is designed
so that no comparisons between data obtained on different days was necessary. A complete set
of data was collected in one session.
5.2.2 Fatigue
The effects of fatigue on tremors is not fully understood. Fatigue can be either
psychological or physiological. To avoid the physiological fatigue of the muscle, the subject
was given about 30 seconds of rest between each trial. To minimize the mental fatigue, only
about 30 trials were done per session; this number was chosen empirically.
5.2.3 Learning Effect
One of the most important concerns for this experiment was for the subject to use the
appropriate muscles to perform tracking tasks. Therefore, practice trials were given before
39
each set of trials to let the subject learn how to use particular muscles and to become familiar
with the experiment.
5.3 Experimental Design
5.3.1 Randomization and Repfication
Randomization and replication are the basic principles of experimental design to increase
the reliability of the experimental results. Randomization "average out" the effects of
extraneous factors that may be present By randomization we mean that the order of the
individual trials of the experiment were randomized. In this study, the order of the individual
trials were not totally randomly determined. Trials with the same target direction were gathered
into the same set, the order of which was then randomized. The reason for not randomizing all
trials is that changing the EMG electrode position could be reduced if trials with the same target
direction were put together since EMG signals are monitored at the same muscles for these
trials. Changing electrode position during the experiment caused the experiment to last longer,
which would. The selective randomization, however, raises the question as to whether or not
grouping trials would affect tremors. Both experiments with totally and partially random
orders of trials were conducted on one subject. The results were found to be of no
considerable difference, therefore it was assumed that grouping of trials would not affect
tremor significantly, which justifies the partial randomization of the experiments in this study.
By replication we mean repetition of trials with the same testing condition. Replication
allows the experimenter to not only check the repeatability of the results, but also to obtain an
estimate of the experimental error. In this study, a minimum of three trials with the same
experimental condition were conducted to meet this requirement.
40
5.3.2 Experimental Protocol
The experimental protocol was primarily based on the considerations stated above and the
experience obtained from preliminary experiments. It was decided that at least two
experiments should be conducted for each subject, each on different days on the arm with
worse tremor. If time allowed and if the tremor in the contralateral arm was significant, a
third-day experiment on the other arm was performed. The objective of the experiments on
Day 1 and Day 2 was to investigate possible coupling between tremors at the elbow and the
shoulder joint of the worse arm and on Day 3 that of the opposite arm. Trials with the
same target direction were grouped together and called a set. Trials within a set were
randomly ordered. The order of the experimental sets were also randomized for
experiments on different days. A typical test protocol on one day is shown in Table 5.2.
5.4 Testing Procedure
5.4.1 Subject and Equipment Preparation
There were a few things which needed to be done for every subject prior to his/her first
experiment. First, the subject was informed to wear a short-sleeve shirt for every experiment
to facilitate applying EMG electrodes. Second, the subject was asked to sign the informed
consent form after the experimenter explained to him/her the purpose of the experiment, the
testing equipment, and the testing procedure. Third, the testing equipment was adjusted to the
subject's size. The arm support was raised to the height of the shoulder joint. The height was
marked so that the arm support could be adjusted before any other experiments to save time and
for repeatability of posture. The Manipulandum was then lifted to make the whole arm lie in
the same horizontal plane as the shoulder joint. Again, the height of the Manipulandum was
marked.
Before any experiments, the myoprocessors, the Manipulandum and the computer system
were tested to ensure good working conditions.
41
Force Level Electrode 1 Electrode 2 Electrode 3
No. of Trials
Target
1
up
10 N
biceps
triceps
ant. deltoid
2
up
20 N
biceps
triceps
ant. deltoid
3
up
20 N
biceps
triceps
ant. deltoid
4
u
10 N
biceps
triceps
ant. deltoid
5
up
20 N
biceps
triceps
ant. deltoid
6
up
10 N
biceps
triceps
ant. deltoid
7
left
20 N
biceps
triceps
pos. deltoid
8
left
10 N
biceps
triceps
pos. deltoid
9
left
10 N
biceps
triceps
pos. deltoid
10
left
20 N
biceps
triceps
pos. deltoid
11
left
10 N
biceps
triceps
pos. deltoid
12
left
20 N
biceps
triceps
pos. deltoid
13
down
20 N
biceps
triceps
pos. deltoid
14
down
20 N
biceps
triceps
pos. deltoid
15
down
10 N
biceps
triceps
pos. deltoid
16
down
20 N
biceps
triceps
pos. deltoid
17
down
10 N
b
triceps
pos. deltoid
18
down
10N
biceps
triceps
pos. deltoid
19
ri
10 N
biceps
triceps
ant. deltoid
20
right
20 N
biceps
triceps
ant. deltoid
21
right
10 N
biceps
triceps
ant. deltoid
22
right
20 N
biceps
triceps
ant. deltoid
23
right
20 N
biceps
triceps
ant. deltoid
42
24
1 right
I
ON
Ibiceps
I triceps
Iant. deltoid
Table 5.1 Typical experimental session (ant. stands for anterior and pos. for posterior).
43
I
5.4.2 A Typical Testing Session
In a typical testing session the subject was seated on the test chair, resting the upper
arm on the arm support. The electrodes monitoring the biceps and the triceps were put at
proper locations and held by a elastic 3-inch wide adjustable strap with hook and loop
fastener. The forearm was coupled to the joystick of the Manipulandum through the
interface. Practice trials were performed until the subject were able to execute the tasks.
The testing protocol was then conducted and data were recorded. Through the test session
the third electrode was switched between the deltoid anterior side and the deltoid posterior
side depending on the tracking tasks. This electrode was held by the experimenter since the
strap could not reach the spot since the strap could not wrap around the shoulder and keep
the electrode in position. The subject were given approximately 30 seconds of rest
between each trial and a 5-minute break after two sets of trials. An assistant helped by
typing and keeping the track of the experimental progress.
5.5 Subjects
5.5.1 Criteriafor Selecting Subjects
The disabled individuals who participated in this study were selected from inpatient and
outpatient populations at the Burke Rehabilitation Center in White Plains, NY. All participants
were diagnosed as having moderate to severe pathological cerebellar intention tremor due to
multiple sclerosis, cerebellar injury. Participants had to be free of cognitive or visual
impairments that would hinder their performance on pursuit tracking tasks, and they had to
demonstrate volitional movement. Further, participants had to have the necessary physical
strength and mental stamina to complete a 30 to 60 minute test session, the mobility to be
transferred to the test chair, and the willingness to cooperate for two to four sessions of testing.
44
5.5.2 Descriptionof Subjects
Three participants met the criteria outlined in Section 3.3.1 and completed at least two
sessions of the experimental protocol. These participants, in order of increasing tremor
severity, are:
Age
Sex
Description
Subject A
43
F
Subject A has a profound unilateral intention tremor and
mild cognitive impairment due to chronic progressive MS.
She is ambulatory with the use of a cane. Subject A is
extremely cooperative but finds the tracking tasks
frustrating because they "show off' her disability.
Subject B
23
M
Subject C has severe intention tremor and ataxia in both
arms as a result of a traumatic brain injury which occurred
seven years ago. He ambulates and dresses himself with
some assistance and successfully uses a computer with a
trackball interface at home.
Subject C
40
M
Subject E has low-frequency, moderate intention tremor
and ataxia affecting his hands and arms due to chronic
progressive MS. He is confined to a wheelchair and
transfers with help. Subject E does not mind participating
in experiments because they make him "hopeful."
45
In addition to the three tremor-disabled individuals, seven able-bodied individuals were
recruited from the student and staff population at MIT. It will be shown in Chapter six that the
results can be identified into two typical kinds. Therefore, the able-bodied subjects are
classified into two groups according to their results. Group one represents the subjects whose
physical strengths are under average. Group two, on the other hand, consists of those who
appears physically strong. The able-bodied "normal" participants are:
Group one
Age
Sex
Subject NIA
28
M
Subject NIB
38
F
Subject Nic
37
M
Subject N1D
53
M
Subject NJE
25
F
Age
Sex
Subject N2A
25
F
Subject N2B
30
F
Group two:
46
Chapter 6
Presentation and Discussion of Results
Results of the experiments are presented in this chapter. Force and EMG data in time and
frequency domain for normal subjects are shown in section 6.1. The results for disabled
subjects are discussed in Section 6.2. As mentioned earlier in Chapter 2, EMG data from all
four muscles for the same condition are available despite the fact that they were collected from
different tests. However, only the results which are considered meaningful are show here.
6.1 Results for Able-bodied Subjects
Able-bodied subjects can be classified into two groups. The first group represents those
subjects who can use the minimum sets of muscles to perform the task. Among a total of
eleven able-bodied subjects tested, nine of them belong to this group. The results within this
group are not very different from each other and they mainly reveal that only the minimum
muscle sets were used. Results from Subject N1A, for the target in the "up" direction are
presented as typical of this group. Figure 6.1 are the results for a force level of 10 N. Figure
6.2 shows the results for 20 N. Subjects N2A and N2B are in the second group. They
appeared to be weaker and had to use more muscles than the minimum sets to accomplish the
experiment. Figure 6.3 and Figure 6.4 show the results from subject N2A for the target in the
"up" direction "up" for force levels of 10 and 20 N, respectively. The results indicate that
subject N2A used biceps and triceps in addition to deltoid anterior side to perform the task for
both force levels. The results for other target directions are not shown here. They also suggest
that subject N2A used muscles other than the minimum set to execute the task for both force
levels. Subject N2B had similar results for all experimental conditions; as a result, they are not
shown here.
47
6.2 Results for Disabled Subjects
The results for disabled subjects are presented in this section. It should be noticed that the
plots for those subjects are not necessarily scaled to the same degree in order to best show the
characteristics of the tremor. The readers should also bear in mind that the results for the
disabled subjects are not very consistent and some results are presented to show the difference.
6.2.1 Results for Subject A
The results for subject A suggest that this subject was able to use the minimum set of
muscles to do the experiment for both force levels for all target directions. The results also
suggest that the tremors at the minimum set of muscles did not induce tremor at the other
muscles. Results for "up" target and for force levels, 10 and 20 N are shown Figure 5 and
Figure 6, respectively. However, there were occasions in which muscles other than the
minimum set were used. Results from two of those trials are shown in Figure 7 and Figure 8
and the conditions are for "up" target and for force levels, 10 and 20 N, respectively. Note that
the EMG pattern is very different from that of an able-bodied subject
6.2.2 Results for Subject B
Figure 6.9 to Figure 6.16 show the results for subject B. For "up" target at a force level of 10
N, subject B demonstrated the ability to use the minimum set of muscles to pursue the target.
However, at force level 20 N, subject B could not accomplish the task without using his
triceps. The EMG spectrum of the triceps show peaks around 10 Hz although tremor at this
frequency was not observed. For "down" and "left" targets, subject B appeared to be able to
use the minimum set to for both force levels. No significant tremor was identified in those
cases. For "right " target at both force levels, subject B was able to perform the task using the
minimum set, however tremor at 4 Hz was found in the force spectrum. In the EMG
spectrum, peaks appeared at 4 Hz and around 10 Hz for both force levels. When the force
level was increased from 10 to 20 N, the amplitude of both peaks increased. The 4 Hz peak
increase x% while the 10 Hz peak increased y%.
48
6.2.3 Results for Subject C
Figure 6.17 to Figure 6.28 show the results for subject C. For "up" target, EMG activities
are found in biceps, triceps, anterior deltoid, and posterior deltoid for both force levels.
Tremor at a frequency of 3 Hz can be seen in both EMG spectra in the force spectrum. For
"down" target, subject C used only triceps and posterior deltoid, the minimum set, to perform
the task for both force levels. Tremor at 3 Hz again can be identified in both EMG spectra and
in the force spectrum. For "left" target, only the minimum set, biceps and posterior deltoid
was activated. The tremor frequency is 3Hz. For "right" target, posterior deltoid was activated
in addition to the minimum set, biceps and anterior deltoid. It can be known from the results
that subject C did not have control over his posterior deltoid. Posterior deltoid was active in all
cases even when it was not required.
49
Time record- x force
0
-I
-10
-10
-20
-20
-
I
10
20
3
Time (sec)
(Spectrum - x force)
I
(Spectrum - y force)
I-
1111
0.8
0.8
E
0.6
0.6
-
0.4
0.4
-
0.2
0.2
I
0
2
30
20
10
Time (sec)
i
4
6
Frequency (Hz)
I
\-I
J
8
0
2
I
I
4
6
Frequency (Hz)
Figure 6.1(a) Subject N1A: Force data in time and frequency domain for target
direction "up" and for a force level of 10 N.
50
·
8
10
5
EMG2 record
EMG1 record
4
3
2
1
ur -- · -uL-·-L----·--
0
10
20
Time (sec)
----- Y-----·-
30
Time (sec)
EMG3 record
I
I
Target-up
4
Force level-10 N
EMG1-biceps
3
EMG2-triceps
2
EMG3-deltoid(anterier)
.j..Li. LI.
ad
~
-
-q~P
I
Time (sec)
Figure 6.1(b) Subject N1A: EMG data in time domain for target direction "up"
and for a force level of 10 N.
·- ·-
----
0.05
0.05
0.04
0.04
0.03
0.03
0.02
0.02
0.01
0.01
0
2
4
6
8
10
(Spectrum - EMG3) -- -
j
0
2
Frequency (Hz)
0.05
I
I
I
I
6
4
Frequency (Hz)
Target-up
0.04
0.03
0.02
-
Force level- 10 N
EMG1-biceps
-
EMG2-triceps
-
EMG3-deltoid(anterier)
0.01
-
2
4
6
8
10
Frequency (Hz)
Figure 6.1(c) Subject N1A: EMG data in frequency domain for target direction
"up" and for a force level of 10 N.
8
10
Time record- y force
Time record- x force
2010
0i
K
-10
-10
-20
-20)-
20
10
Time (sec)
Time (sec)
0.8
0.6
0.4
0.2
0
2
4
6
8
0
10
Frequency (Hz)
2
4
6
Frequency (Hz)
Figure 6.2(a) Subject N1A: Force data in time and frequency domain for target
direction "up" and for a force level of 20 N.
53
8
10
EMG2 record
EMG1 record
3
-
YIC·r~Cr'
I--"~1
~-r.rC.
m_.
'~I"-"Y'
-
Time (sec)
Time (sec)
EMG3 record
Target-up
Force level-20 N
EMG 1-biceps
EMG2-triceps
EMG3-deltoid(anterier)
I
Time (sec)
Figure 6.2(b) Subject NIA: EMG data in time domain for target direction "up"
and for a force level of 20 N.
54
(Spectrum - EMG1)
0.05
0.04
0.04
-
0.03
-
0.02
-
(Spectrum - EMG2)
I-
0.05
0.03
I
-
-
0.02
E
0.01
0.01
E
I
0
2
I
4
I
6
I
8
I
10
I
0
2
Frequency (Hz)
,
i
6
4
Frequency (Hz)
0.05
Target-up
0.04
Force level-20 N
EMG1-biceps
0.03
EMG2-triceps
0.02
EMG3-deltoid(anterier)
0.01
0
2
4
6
8
10
Frequency (Hz)
Figure 6.2(c) Subject N1A: EMG data in frequency domain for target direction
"up" and for a force level of 20 N.
.
8
10
Time
0~
record-
x
Timerecor- I oc
Time rcord-x forc
20
20
10
10
A
0
O)
-10
-20
-20
1
(Spectrum- x force)
0.8-
-
0.8
0.6-
0.6
0.4
9- 0.4
o0.2-
0
30
20
Time (sec)
Time (sec)
1
AI
force
10
0
30
20
10
y
0
-10
0
record-
Time
force
0.2
2
4
6
Frequency (Hz)
8
10
0
2
4
6
Frequency (Hz)
Figure 6.3(a) Subject N2A: Force data in time and frequency domain for target
direction "up" and for a force level of 10 N.
8
10
EMG2 record
EMG1 record
I
I
4
3
2
,.~ ~ml~l.ll._..
UL.·
k.r
.............
.....
1 LL LLI.-~I~. 1 L. 111..
-
I .- rL~
r....
I- 7"'-
|oI·.....L_
WPM-
.
•!
L--L~u •..
....
*1
-·-..·
.,
W"--
Time (sec)
Time (sec)
EMG3 record
I
I
Target-up
4
Force level-10 N
EMG1-biceps
EMG2-triceps
EMG3-deltoid(anterier)
I_
_
_
Time (sec)
Figure 6.3(b) Subject N2A: EMG data in time domain for target direction "up"
and for a force level of 10 N.
· ·
m1
•
a
(Spectrum - EMG 1)
0.05
0.04
0.04
0.03
0.03
0.02
0.02
0.01
0.01
2
0.05
4
6
Frequency (Hz)
(Spectrum - EMG2)
0.05
I-
0
8
2
4
6
Frequency (Hz)
(Spectrum - EMG3)
I-
-
Target-up
0.04
Force level-10 N
EMG1-biceps
0.03
EMG2-triceps
0.02
EMG3-deltoid(anterier)
0.01
II
2
-I
4
6
Frequency (Hz)
-
8
Figure 6.3(c) Subject N2A: EMG data in frequency domain for target direction
"up" and for a force level of 10 N.
58
8
10
Time record- x force
-10
-20
Time (sec)
1
8
2
30
20
10
Time (sec)
0
30
20
10
(Spectrum - y I force)
I
-
1-
I
2
4
6
8
2
10
4
6
Frequency (Hz)
Frequency_ (Hz)
Figure 6.4(a) Subject N2A: Force data in time and frequency domain for target
direcion "up" and for a force level of 20 N.
59
8
c
EMG1 record
EMG2 record
5
4
4
E
3
2
2
E
1
C]
0
10
20
iLMi~
30
Time (sec)
Tiune (sec)
EMG3 record
Target-up
Force level-20 N
EMG1-biceps
EMG2-triceps
EMG3-deltoid(anterier)
10
20
Time (sec)
Figure 6.4(b) Subject N2A: EMG data in time domain for target direction "up"
and for a force level of 20 N.
0.05
0.05
0.04
0.04
0.03
0.03
(Spectrum
- EMG2)
1L I
I
-
-
E
0.02
0.02
E
0.01
0.01
I
----- ------
0
2
4
6
Frequency (Hz)
8
4
10
I
6
Frequency (Hz)
(Spectrum - EMG3)
0.05
Target-up
0.04
Force level-20 N
I-I
EMG 1-biceps
0.03
EMG2-triceps
0.02
EMG3-deltoid(anterier)
0.01
0
2
4
6
8
10
Frequency (Hz)
Figure 6.4(c) Subject N2A: EMG data in frequency domain for target direction
"up" and for a force level of 10 N.
8
10
Time record- x force
-10
-20
10
20
30
20
10
0
30
Time (sec)
Time (sec)
(Spectrum - x force)I
I
I
I
2
E
i
0
I
i
2
4
i
6
Frequency (Hz)
i
8
0
10
2
4
6
Frequency (Hz)
Figure 6.5(a) Subject A: Force data in time and frequency domain for target
direction "up" and for a force level of 10 N.
62
8
10
EMG2 record
EMG1 record
4
3
-
3
E
-r
· YIC'
Time (sec)
Time (sec)
EMG3 record
I
I
Target-up
Force level-10 N
EMG -biceps
3
EMG2-triceps
2
Ji
~
hLJhB
EMG3-deltoid(anterier)
IA1lWI~I~~L
Time (sec)
Figure 6.5(b) Subject A: EMG data in time domain for target direction "up"
and for a force level of 10 N.
63
II
0.05
0.05
0.04
0.04
0.03
0.03
0.02
0.02
0.01
0.01
(Spectrum - EMG2)
I
0
2
0.05
0.04
0.03
0.02
I
4
6
Frequency (Hz)
8
10
0
2
I
I
6
4
Frequency (Hz)
(Spectrum - EMG3) °I
I
I
Target-up
-
Force level- 10 N
EMG1-biceps
-
EMG2-triceps
W
EMG3-deltoid(anterier)
0.01
n
2
4
6
Frequency (Hz)
8
10
Figure 6.5(c) Subject A: EMG data in frequency domain for target direction
"up" and for a force level of 10 N.
64
8
10
Time record- x force
20
10
-10
-20
0
10
20
0
30
10
Time (sec)
20
Time (sec)
30
(Spectrum - x force)
I
I
I
I
N
z
6
4
a,:
CLi
01
"0
2
n)
C)
4
6
Frequency (Hz)
8
2
10
4
6
Frequency (Hz)
Figure 6.6(a) Subject A: Force data in time and frequency domain for target
direction "up" and for a force level of 20 N.
65
8
10
EMG2 record
EMG1 record
4
4
3
3
-
2
1
IJ I_
n
0
20
10
-J
I
I r'
Time (sec)
Time (sec)
5C
'
30
EMG3 record
Target-up
Force level-20 N
EMG 1-biceps
EMG2-triceps
EMG3-deltoid(anterier)
Time (sec)
Figure 6.6(b) Subject A: EMG data in time domain for target direction "up"
and for a force level of 20 N.
66
'
-
0.05
0.05
0.04
0.04
> 0.03
0.03
0.02
0.02
0.01
0.01
0
0
0
2
4
6
Frequency (Hz)
8
10
(Spectrum - EMG2)
-II
0
2
4
6
Frequency (Hz)
(Spectrum - EMG3)
0.05
Target-up
0.04
-
0.03
-
0.02
-
Force level-20 N
EMG1-biceps
EMG2-triceps
EMG3-deltoid(anterier)
.01 r
2
4
6
8
10
Frequency (Hz)
Figure 6.6(c) Subject A: EMG data in frequency domain for target direction
"up" and for a force level of 20 N.
8
10
Time record- xx force
force
Time recordI
I
20
20
10
10
a
0
0
-10
-10
-20
-20
Time (sec)
Time (sec)
10
(Spectrum - x force)-
10
30
20
10
0
30
20
10
0
8-
8
4 -
4
2-
o
2-
0
0
2
,
0
2
4
6
8
0
10
Frequency (Hz)
2
4
6
8
Frequency (Hz)
Figure 6.7(a) Subject A: Force data in time and frequency domain for target
direction "up" and for a force level of 10 N (minimum set of muscles
were not used in this run).
68
10
EMG1 record
EMG2 record
4
E
33
1.
n
...-
0
10
..
J
" i
)1.
.
.I n
I
lu
..
.Ll
I
~MI~LYIPIYI- -- ~"LY--Y·YU~I-'
20
Time (sec)
EMG3
record
II
Target-up
Force level-10 N
3K
EMG 1-biceps
EMG2-triceps
EMG3-deltoid(anterier)
~.L.
A~JII a..
"L
1 1L1 ~..
1J
..-LA
La•
,
.
.
.
20
Time (sec)
Figure 6.7(b) Subject A: EMG data in time domain for target direction "up"
and for a force level of 10 N (minimum set of muscles
were not used in this run).
L
YI111-
20
10
Time (sec)
lmI-Ld
ILd•B
(Spectrum - EMG1)
0.05
I
I
04
I
0.05
I
0.04
-
0.03
0.03
0.02
0.02
0.01
0.01
I
0
2
4
6
0
8
2
Frequency (Hz)
6
4
Frequency (Hz)
0.05
Target-up
0.04
Force level-10 N
EMG 1-biceps
0.03
EMG2-triceps
0.02
EMG3-deltoid(anterier)
0.01
0
2
4
6
Frequency (Hz)
8
10
Figure 6.7(c) Subject A: EMG data in frequency domain for target direction
"up" and for a force level of 10 N (minimum set of muscles
were not used in this run).
8
10
Time record- x force
-10
-20
10
20
30
Time (sec)
Time (sec)
0
2
4
6
Frequency (Hz)
30
20
10
0
8
10
0
2
4
6
Frequency (Hz)
8
Figure 6.8(a). Subject A: Force data in time and frequency domain for target
direction "up" and for a force level of 20 N (minimum set of muscles
were not used in this run).
10
EMG2 record
3. dLL.*IAdIdL
10
20
Time (sec)
11iY
JL~LLM~J
woo;
20
30
Time (sec)
Target-up
Force level-20 N
EMG1-biceps
EMG2-triceps
EMG3-deltoid(anterier)
10
20
30
Time (sec)
Figure 6.8(b) Subject A: EMG data in time domain for target direction "up"
and for a force level of 20 N (minimum set of muscles
were not used in this run).
0.05
- EMG1)
(Spectrum
I
I
I
0.04
-
0.03
-
- EMG2)
(Spectrum
__
0.05
0.04
I
I
4
6
-
0.03
E
0.02
0.01
-
0.02
-
0.01
--
-'
2
--
4
-
'
8
6
10
2
Frequency (Hz)
Frequency (Hz)
(Spectrum - EMG3)
0.05
Target-up
0.04
-
-
Force level-20 N
I
EMG1-biceps
0.03
0.02
0.01
-
~
2
4
6
EMG2-triceps
EMG3-deltoid(anterier)
_ _I
8
10
Frequency (Hz)
Figure 6.8(c) Subject A: EMG data in frequency domain for target direction
"up" and for a force level of 20 N (minimum set of muscles
were not used in this run).
8
10
Time record- x force
20
20
10
10
Z0
)
O
0
0
-10
-10
-20
-20
Time (sec)
Time (sec)
1
(Spectrum- x force)-
1
S0.8 -
0.8
0.6-
0.6
0.4 -
0.4
0.2
0.2
0
o
30
20
10
0
30
20
10
0
U-
0
2
4
6
8
0
10
2
4
6
Frequency (Hz)
Frequency (Hz)
Figure 6.9(a) Subject B: Force data in time and frequency domain for target
direction "up" and for a force level of 10 N.
74
8
10
EMG2 record
EMG1 record
4
-
2
. .I . I. .I I
li
II
I I
I
Time (sec)
Time (sec)
EMG3 record
Target-up
Force level-10 N
EMG1-biceps
EMG2-triceps
EMG3-deltoid(anterier)
10
20
30
Time (sec)
Figure 6.9(b) Subject B EMG data in time domain for target direction "up"
and for a force level of 10 N.
75
(Spectrum - EMG1)
0.05
0.04
(Spectrum- EMG2)
0.05
0.04
-
0.03
0.03
E
02
-
01
-
0.02
-
0.01
Frequency (Hz)
Frequency (Hz)
(Spectrum - EMG3)-
0.05
Target-up
0.04
Force level-10 N
0.03
-
0.02
-
EMG 1-biceps
EMG2-triceps
EMG3-deltoid(anterier)
0.01
-- I
Frequency (Hz)
Figure 6.9(c) Subject B: EMG data in frequency domain for target direction
"up" and for a force level of 10 N.
Time record- x force
10
-
0 1~c _...
--. -
. -,
.~
10
-10
-20
-20
10
Time (sec)
0.8
0.8
0.6
0.6
0.4
0.4
0.2
0.2
2
4
6
Frequency (Hz)
8
10
0
2
20
Time (sec)
4
6
Frequency (Hz)
Figure 6.10(a) Subject B: Force data in time and frequency domain for target
direction "up" and for a force level of 20 N.
30
8
10
EMG1 record
EMG2 record
4
0
10
20
Time (sec)
Time (sec)
EMG3 record
Target-up
Force level-20 N
EMG1-biceps
EMG2-triceps
EMG3-deltoid(anterier)
10
20
30
Time (sec)
Figure 6.10(b) Subject B: EMG data in time domain
for target direction "up"
and for a force level of 20 N.
(Spectrum - EMG 1)
0.05
0.05
0.04
0.04
-
0.03
0.03
-
0.02
0.02
0.01
0.01
5
10
Frequency (Hz)
Frequency (Hz)
0.05
Target-up
0.04
Force level-20 N
0.03
EMG 1-biceps
EMG2-triceps
0.02
EMG3-deltoid(anterier)
0.01
5
10
Frequency (Hz)
15
Figure 6.10(c) Subject B: EMG data in frequency domain for target direction
"up" and for a force level of 20 N.
79
Time record- x force
Time record- y force
10
0
7~
10
-20
.--
-I~,
·--
-_~ .. .
10
-
-20
-
10
20
-
S
10
2-
1
20
Time (sec)
Time (sec)
(Spectrum - x force):
(Spectrum- y force)
1.5
1.5
1
1
0.5
0.5
--
-I
I
0
2
4
6
Frequency (Hz)
8
10
0
2
4
6
Frequency (Hz)
Figure 6.11(a) Subject B: Force data in time and frequency domain for target
direction "down" and for a force level of 10 N.
80
8
5
EMG2 record
EMG1 record
4
3
2
1
L - L.
0
0
10
20
11
30
Time (sec)
Time (sec)
Target-down
Force level-10 N
EMG 1-biceps
EMG2-triceps
EMG3-deltoid(posterier)
10
20
30
Time (sec)
Figure 6.11(b) Subject B EMG data in time domain for target direction "down"
and for a force level of 10 N.
81
(Setu-MD
V.U
-II
(Spectrum - EMG2)~
0.05
0.04
0.04
-
0.03
0.03
-
0.02
-
0.01
E
(Spectrum
- EMG2)
1
-
0.02
0.01
E
Frequency (Hz)
(Spectrum - EMG3)
0.05
0.04
Frequency (Hz)
I
Target-down
-
Force level- 10 N
0.03
EMG 1-biceps
-
EMG2-triceps
0.02
E
EMG3-deltoid(posterier)
0.01
~AM~JVA
Frequency (Hz)
Figure 6.11(c) Subject B: EMG data in frequency domain for target direction
"down" and for a force level of 10 N.
82
-10
-10
-20
-20
0
10
20
Time (sec)
30
10
0
20
30
Time (sec)
(Spectrum - y force)
1111
1.5
.5
E
1
0.5
0.5
0
B
I
2
4
6
8
10
0
Frequency (Hz)
2
4
I
I
6
8
Frequency (Hz)
Figure 6.12(a) Subject B: Force data in time and frequency domain for target
direction "down" and for a force level of 20 N.
83
10
EMG1
record
EMG1 record
J
I
EMG2 record
I
4
3
2
-
-
IIj L1Al
-aLLI
_L
'r~~~,..,
J
I
LJ IYh-.
--- -1---~_~·--~~----
·----
II·
1162-a~ITI~-----~--
-L
·---- ·--
· _·-·-I-
-- ~·- -
I
- -------
~-----
Tunime (sec)
Time (sec)
EMG3 record
Target-down
Force level-20 N
EMG1-biceps
EMG2-triceps
EMG3-deltoid(posterier)
I
1
0
10
20
Time (sec)
Figure 6.12(b) Subject B: EMG data in time domain for target direction "down"
and for a force level of 20 N.
84
·· ·
-----
U.
00
-
J
EMG1)
•
.--
(Spectrum - EMG2)
0.05
-
(Spectrum
N
0.04
0.03 -
0.04
-
0.03
-
<I-i
0.02
-
oa,,= 0.02
0.01
pt=
--
|
0.01
1-
1-
I
|
Frequency (Hz)
I
Frequency (Hz)
0.05
t
Target-down
0.04
Force level-20 N
> 0.03
EMG1-biceps
EMG2-triceps
= 0.02
EMG3-deltoid(posterier)
~0.01
5
10
Frequency (Hz)
Figure 6.12(c) Subject B: EMG data in frequency domain for target direction
"down" and for a force level of 20 N.
Time record- x force
10
Time record- y force
10
-
0
10
I--L-_-,
prEC~4~1C·~C~k~""."~~
-10
E
-20
-20
Time (sec)
Time (sec)
(Spectrum - y force)
I
1.
1.5
I
I
i-
0.5
E
0
2
4
6
Freqiuency (Hz)
8
10
4
6
Frequency (Hz)
Figure 6.13(a) Subject B: Force data in time and frequency domain for target
direction "left" and for a force level of 10 N.
8
EMG1 record
EMG2 record
I
4
3
)-
-
-I
JILUd·I~~.1.LUIL
tl
IpI
i
"
II
"
~
*r
-- ·--
· ·sr------l·u·-
Time (sec)
----
~
Time (sec)
EMG3 record
I
,
Target-left
Force level-10 N
EMG1-biceps
EMG2-triceps
2
E
EMG3-deltoid(anterier)
I
--
·-
------.
I_~ _~_ ___, __,____
.____~_
Time (sec)
Figure 6.13(b) Subject B EMG data in time domain for target direction "left"
and for a force level of 10 N.
-ul-·
-- ---- ·-----
0.01
0.01
=0.008
0.008
Cl
00.006
00.006
=0.004
•0.004
>0.002
>0.002
00
0
0
5
10
(Spectrum - EMG3J)
lI
U.UI
o>0.006
10
Frequency (Hz)
Frequency (Hz)
o0.008
5
0
15
Target-left
-
Force level-10 N
EMG 1-biceps
-
EMG2-triceps
=0.004
EMG3-deltoid(anterier)
:0.002
C00
V
Frequency (Hz)
Figure 6.13(c) Subject B: EMG data in frequency domain for target direction
"left" and for a force level of 10 N.
88
Time record- y force
W
20
20
10
10
Z
0
0)
I-.
0
·
-10
-10
-20
-20
0
10
20
k~y--""7-·Y--
0
30
10
Time (sec)
20
30
Time (sec)
2
(Spectrum - y force)
2
1.5
",
a
1.5
0.5
g,
o
0n
2;
-
0.5-
fl
0
0
2
4
6
Frequency (Hz)
8
10
0
2
4
6
Frequency (Hz)
Figure 6.14(a) Subject B: Force data in time and frequency domain for target
direction "left" and for a force level of 20 N.
89
8
10
EMG2 record
2
E
I--
10
20
30
Time (sec)
Time (sec)
Target-left
Force level-20 N
EMG1-biceps
EMG2-triceps
EMG3-deltoid(anterier)
0
10
20
30
Time (sec)
Figure 6.14(b) Subject B: EMG data in time domain for target direction "left"
and for a force level of 20 N.
0.01
0.01,
0.008
C-4
(Spectrum - EMG2)
I
0.008
O0.006
O0.006
0.004
0.004
0*0.002
0.002
91
0(
I
V
5
0
Frequency (Hz)
Frequency (Hz)
(Spectrum - EMG3)-
0.01
Target-left
ý0.008
Force level-20 N
EMG 1-biceps
>0.006
EMG2-triceps
=0.004
EMG3-deltoid(anterier)
30.002
0
n
1
0i
I
·
10
- ----
Frequency (Hz)
Figure 6.14(c) Subject B: EMG data in frequency domain for target direction
"left" and for a force level of 20 N.
Time record- x force
I
Time rrcnrd- v forcr
I
20
20
10
10
/liTlcul-cw
Z
I
I
I
0
-10
~G~-----
-10
-20
0
I
I
|
|
10
20
-20
I
30
r
i
0
10
20
Time (sec)
Timune (sec)
2
2
1.5
1.5
30
(Spectrum - y force)
N
a
0.5
B 0.5
0
0
0
2
4
6
Frequency (Hz)
8
10
0
2
4
6
Frequency (Hz)
Figure 6.15(a) Subject B: Force data in time and frequency domain for target
direction "right" and for a force level of 10 N.
92
8
10
EMGl record
EMG2 record
5
5
T i e (see)
Time (sec)
EMG3
record
I
5
I
-= -
.-. .
1
Target-right
Force level- 10 N
3t
i
EMG 1-biceps
Figure 6.15(b) Subject B EMG data in time domain for target direction "right"
and for a force level of 10 N.
J. 1J
\. .
L .
(Spectrum
I
0
.
..
-
. . ..
.
EMG1)
.I
0.15
I
0.1
0.1
E
0.05
0.05
E
I
5
Frequency (Hz)
10
Frequency (Hz)
(Spectrum - EMG3)-
0.15
Target-right
Force level- 10 N
0.1
E
EMG -biceps
EMG2-triceps
5
05
EMG3-deltoid(posterier)
-
Frequency (Hz)
Figure 6.15(c) Subject B: EMG data in frequency domain for target direction
"right" and for a force level of 10 N.
94
Time record- x force
-10
-10
-20
-20
0
10
20
30
20
10
0
30
Time (sec)
Time (sec)
2
1.5
(Spectrum - y force)
I
I
i
2
4
6
1.5
1
0.5
0.5
I
0
2
4
6
Frequency (Hz)
8
10
Frequency (Hz)
Figure 6.16(a) Subject B: Force data in time and frequency domain for target
direction "right" and for a force level of 20 N.
95
8
10
EMG1 record
II
,
Ill~J~Ii
h~mMLALUL.L-,kr~hi
L.I.L
1.,L..J.A
L
·.
i
10
Time (sec)
20
Time (sec)
EMG3 record
I
Target-right
Force level-20 N
3 -_
EMG1-biceps
S
.. I
EMG2-triceps
I . I
EMG3-deltoid(posterier)
I
I
Time (sec)
Figure 6.16(b) Subject B: EMG data in time domain for target direction "right"
and for a force level of 20 N.
96
30
0.15
0.5
(Spectrum - EMG 1)
.
0.1
0.1
0.05
0.0
a
I
5
10
10
5
Frequency (Hz)
0
5
10
Frequency (Hz)
(Spectrum - EMG3)
0.15
Target-right
Force level-20 N
0.1
-
EMG 1-biceps
EMG2-triceps
0.05
EMG3-deltoid(posterier)
Frequency (Hz)
Figure 6.16(c) Subject B: EMG data in frequency domain for target direction
"right" and for a force level of 20 N.
15
Time record- x force
20
101
0
-10
-20
30
20
10
0
30
20
10
Time (sec)
Time (sec)
251
20
20
15
15
I
- v Iforce)
(SDectrum
I
I
10
I-
0
2
6
4
Frequency (Hz)
8
2
10
I
6
4
Frequency (Hz)
Figure 6.17(a) Subject C: Force data in time and frequency domain for target
direction "up" and for a force level of 10 N.
98
I
8
10
EMG2 record
EMG1 record
3
[II
m
a JLIAI
ii,i . I
I
10
30
20
10
20
Time (sec)
Time (sec)
EMG3 record
Target-up
Force level- 10 N
EMG1-biceps
I..
I
I
EMG2-triceps
EMG3-deltoid(anterier)
..-
~n~,,....,,,,-
,,,,,,-,?
.. 1.
-
10
Time (sec)
Figure 6.17(b) Subject C EMG data in time domain for target direction "up"
and for a force level of 10 N. Note: EMG data are for biceps,
triceps, and anterior deltoid.
99
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0.2
0.1
0.1
0
2
4
6
8
0
10
2
Frequency (Hz)
4
6
Frequency (Hz)
8
0.5
Target-up
0.4
Force level-10 N
EMG1-biceps
0.3
EMG2-triceps
0.2
EMG3-deltoid(anterier)
0.1
0
2
4
6
Frequency (Hz)
8
10
Figure 6.17(c) Subject C: EMG data in frequency domain for target direction
"up" and for a force level of 10 N. Note: EMG data are for biceps,
triceps, and anterior deltoid.
100
10
-10
-20
0
10
20
30
0
10
Time (sec)
20
Time (sec)
(Spectrum -xforce)-
(Spectrum - y force)
I
I
II
30
I
I
I
20
I
15
10
0
2
4
6
Frequency (Hz)
~crc/
8
10
2
\,
4
6
Frequency (Hz)
Figure 6.18(a) Subject C: Force data in time and frequency domain for target
direction "up" and for a force level of 10 N.
101
8
10
EMG1 record
I
1,1
Ill
II
O
I
I
11
EMG2 record
I
i.
II
I,.
~(CLY(UL~L~U~(Cei~e~CQrmYI~YII'~I-~'~UU'
I
20
0
10
Time (sec)
Time (sec)
5-
20
EMG3 record
Target-up
4
Force level-10 N
EMG1-biceps
EMG2-triceps
EMG3-deltoid(posterier)
0
10
20
30
Time (sec)
Figure 6.18(b) Subject C EMG data in time domain for target direction "up"
and for a force level of 10 N. Note: EMG data are for biceps,
triceps, and posterior deltoid.
102
(Spectrum - EMG1)
0.5
I
SI
(Spectrum - EMG2)
0.5
I
-e
0.4
-
0.3
-
0.4
0.3
0.2
-
0.1
-
0.2
0.1
p
2
~
I
4
6
Frequency (Hz)
8
10
0
2
4
6
Frequency (Hz)
0.5
Target-up
0.4
Force level- 10 N
0.3
EMG 1-biceps
EMG2-triceps
0.2
EMG3-deltoid(posterier)
0.1
0
2
4
6
8
10
Frequency (Hz)
Figure 6.18(c) Subject C: EMG data in frequency domain for target direction
"up" and for a force level of 20 N. Note: EMG data are for biceps,
triceps, and posterior deltoid.
103
8
10
Time record- x force
-10
-20
0
10
20
30
Time (sec)
30
20
10
0
Time (sec)
Ill
(Spectrum - y force)
I
20
15
10
I|
0
2
4
6
Frequency (Hz)
8
10
2
6
4
Frequency (Hz)
Figure 6.19(a) Subject C: Force data in time and frequency domain for target
direction "up" and for a force level of 20 N.
104
8
10
EMG1 record
EMG2 record
I
10
20
30
10
Time (sec)
CAI(I?
20
Time (sec)
~~~~
Target-up
Force level-20 N
EMG1-biceps
EMG2-triceps
EMG3-deltoid(anterier)
10
20
30
Time (sec)
Figure 6.19(b) Subject C EMG data in time domain for target direction "up"
and for a force level of 20 N. Note: EMG data are for biceps,
triceps, and anterior deltoid.
105
iI
30
0.5
0.5
(Spectrum - EMG1)
0.4
-
0.4
-
0.3
0.3
-
0.2
A
0.2
0.1
0.1
m
I
1
0
2
0.3
0.2
8
0
2
6
4
Frequency (Hz)
(Spectrum - EMG3)
0.5 I
0.4
4
6
Frequency (Hz)
-
ml
Target-up
rAII
Force level-20 N
EMG1-biceps
EMG2-triceps
EMG3-deltoid(anterier)
0.1
C..--_,
2
4
6
Frequency (Hz)
8
10
Figure 6.19(c) Subject C: EMG data in frequency domain for target direction
"up" and for a force level of 20 N. Note: EMG data are for biceps,
triceps, and anterior deltoid.
106
8
10
Time record- x force
20
-10
-2(
30
20
10
30
20
10
0
Time (sec)
Time (sec)
(Spectrum - y force)
20
E
15
-
10
-
II
II
2
4
6
Frequency (Hz)
8
--
2
10
6
4
Frequency (Hz)
Figure 6.20(a) Subject C: Force data in time and frequency domain for target
direction "up" and for a force level of 20 N.
107
8
10
l:Mr.9 rpinrd
EMG1 record
2
1
Si I.I,
I ,su
I
I UII
E
Al
I
IUU.
ILL.
" .'..__..
-LgU"J
20
10
0
30
20
Time (sec)
Tune (sec)
EMG3 record
Target-up
Force level-20 N
EMG1-biceps
EMG2-triceps
EMG3-deltoid(posterier)
i
10
Time (sec)
Figure 6.20(b) Subject C EMG data in time domain for target direction "up"
and for a force level of 20 N. Note: EMG data are for biceps,
triceps, and posterior deltoid.
108
30
(Soectrum - EMG1)
0.5
(Spectrum - EMG2)
0.5
-
0.4
0.4
0.3
0.3
0.2
E--
0.2
0.1
0.1
-
~
2
4
6
8
0
2
Frequency (Hz)
4
6
Frequency (Hz)
0.5
Target-up
0.4
Force level-20 N
0.3
EMG 1-biceps
EMG2-triceps
0.2
EMG3-deltoid(posterier)
0.1
0
2
4
6
Frequency (Hz)
8
10
Figure 6.20(c) Subject C: EMG data in frequency domain for target direction
"up" and for a force level of 20 N. Note: EMG data are for biceps,
triceps, and posterior deltoid.
109
8
Time record- x force
-10
-20
10
20
30
0
10
Time (sec)
(Spectrum - x force)I
I
I
I
I
20
20
15
15
10
10 6
I
2
4
6
Frequency (Hz)
I
8
II
(Spectrum - y force)
I
I
1-
2
--
I
i
4
6
Frequency (Hz)
Figure 6.21(a) Subject C: Force data in time and frequency domain for target
direction "down" and for a force level of 10 N.
110
30
Time (sec)
"
10
20
i
8
10
5
-
5
EMGl record
4
4
a
w
w
u
E
W
EMG2 record
.3
3>
3
u
G
2
2
W
1
2
1
0
0
10
30
20
Time (sec)
5,
10
20
Time (sec)
EMG3 record
I
I
0
0
= -.=
Target-down
Force level- 10 N
EMG 1-biceps
EMG2-triceps
EMG3-deltoid(posterier)
Time (sec)
Figure 6.2 1(b) Subject C EMG data in time domain for target direction "down"
and for a force level of 10 N.
30
(Spectrum - EMG1)
0.5
0.4
0.3
(Spectrum - EMG2)
0.5
E
0.4
-
0.3
-
0.2
0.2
-
0.1
0.1
-
-I
ZI\
I
I
4
6
Frequency (Hz)
8
0
d
---
4
2
i
i
6
8
Frequency (Hz)
0.5
Target-down
0.4
Force level-10 N
0.3
EMG 1-biceps
EMG2-triceps
0.2
EMG3-deltoid(posterier)
0.1
0
2
4
6
8
10
Frequency (Hz)
Figure 6.21(c) Subject C: EMG data in frequency domain for target direction
"down" and for a force level of 10 N.
112
10
Time record- y force
Time record- x force
-
-
10
0
-10
-10
-20
-20
-i
30
20
10
0
30
20
10
Time (sec)
Time (sec)
te
*rV"Iirv
\jIYAILullL -
3
f ^
N
Lr~JL/
I
I
I--
n
5N
0
2
6
4
Frequency (Hz)
8
2
10
--..
•
_,
_
n
6
4
Frequency (Hz)
Figure 6.22(a) Subject C: Force data in time and frequency domain for target
direction "down" and for a force level of 20 N.
113
8
EMG1 record
U
10
20
30
10
Tune (sec)
20
Time (sec)
EMG3 recnrd
Target-down
Force level-20 N
EMG1-biceps
EMG2-triceps
EMG3-deltoid(posterier)
0
0
10
20
Time (sec)
30
Figure 6.22(b) Subject C EMG data in time domain for target direction "down"
and for a force level of 20 N.
114
30
0.5
0.5
0.4
0.4
(Spectrum - EMG2)
I
I
I
I
E
0.3
0.3
0.2
0.2
0.1
0.1
-
-
0
2
4
6
8
10
0
Frequency (Hz)
·
-L
2
4
6
Frequency (Hz)
0.5
Target-down
0.4
Force level-20 N
0.3
EMG 1-biceps
EMG2-triceps
0.2
EMG3-deltoid(posterier)
0.1
0
2
4
6
-
8
10
Frequency (Hz)
Figure 6.22(c) Subject C: EMG data in frequency domain for target direction
"down" and for a force level of 20 N.
115
I
8
10
Time record- y force
Time record- x force
20
10
10
-10
-10
-20
-20
E
0
10
20
Time (sec)
Time (sec)
(Spectrum - x force)
I
I
I
3C
20
10
30
I
IlI
(Spectrum - y force)
I
n
6
4;h
C,,,,
I
0I:::
c,,.
n
iin
()
I
i i
4
6
Frequency (Hz)
L //
I
8
0
10
2
I-
I
6
4
Frequency (Hz)
Figure 6.23(a) Subject C: Force data in time and frequency domain for target
direction "left" and for a force level of 10 N.
116
I
I
8
10
EMG1 record
I
"I- -II
'
EMG2 record
'I
I-
2
II1
I-
I
I
I
4
E
-I
I I I ~I II . ri~i·
CUI , 1 111 11,11,1IIIUIIIIIIII
i
r
0
...
1
---.-
0
10
20
3(
Time (sec)
E~~?MG3
J
t
I
I
eVI
U or
Time (sec)
d
!
'
I-
C
-1
"
-.
i
I
-1
I
-j
Target-left
I
Force level-10 N
EMG 1-biceps
EMG2-triceps
I
EMG3-deltoid(anterier)
I
__
-
Time (sec)
Figure 6.23(b) Subject C EMG data in time domain for target direction "left"
and for a force level of 10 N. Note: EMG data are for biceps,
triceps, and anterior deltoid.
117
(Spectrum - EMG1)
0.5
0.4
0.3
0.5
I
(Spectrum
- EMG2)
I
0.4
-
0.3
-
I
I
-
0.2
0.2
-
0.1
0.1
E
\,
4
6
Frequency (Hz)
_,
_1
8
I
10
0
I
I
I
I
4
6
Frequency (Hz)
i
Target-left
S
Force level-10 N
0.3
0.2
2
(Spectrum - EMG3)
0.5
0.4
II
2
EMG1-biceps
EMG2-triceps
r
EMG3-deltoid(anterier)
0.1
2
4
6
Frequency (Hz)
8
10
Figure 6.23(c) Subject C: EMG data in frequency domain for target direction
"left" and for a force level of 10 N. Note: EMG data are for biceps,
triceps, and anterior deltoid.
118
l
8
10
Time record- x force
Time record- v force
20
20
10
10
Z
0
-10
-10
-20
-20
0
10
20
30
0
10
Time (sec)
30
Time (sec)
10
(Spectrum - y force)
10
8
88-
6
S
6-
0
4
4-
o
2
o 2
3:
20
0
0
2
4
6
8
10
0
Frequency (Hz)
2
4
6
Frequency (Hz)
Figure 6.24(a) Subject C: Force data in time and frequency domain for target
direction "left" and for a force level of 10 N.
119
8
10
EMG2 record
EMG1 record
I
I
4
.
32
2
E
1Y-
·-
-
-- ~-'
-----
-~L~-
· ~___
Time (sec)
Time (sec)
Target-left
Force level-10 N
EMG1-biceps
EMG2-deltoid(anterier)
EMG3-deltoid(posterier)
20
10
30
Time (sec)
Figure 6.24(b) Subject C EMG data in time domain for target direction "left"
and for a force level of 10 N. Note: EMG data are for biceps,
triceps, and posterior deltoid.
120
( 0 -- ^ -- *.-- . -U.J
0
(
11•%R/'•_ 1
--
U.,
5
urtcepS
m
-
EMGl)
0.4
0.4
0.3
0.3
= 0.2
= 0.2
o
0.1
I0
0.1
AL
n
I
I/I
0
- I
....
2
4
6
Frequency (Hz)
8
10
0
2
4
6
Frequency (Hz)
8
0.5
Target-left
0.4
o
Force level-10 N
0.3
EMG1-biceps
EMG2-deltoid(anterier)
0.2
EMG3-deltoid(posterier)
S0.1
0
0
2
4
6
Frequency (Hz)
8
10
Figure 6.24(c) Subject C: EMG data in frequency domain for target direction
"left" and for a force level of 10 N. Note: EMG data are for biceps,
triceps, and posterior deltoid.
121
10
-10
-20
10
0
30
20
10
20
Timune (sec)
Time (sec)
1
(Spectrum - v force) I
I
-1
4
2
2
|m
0
2
6
4
Frequency (Hz)
8
10
-- A-
0
2
6
4
Frequency (Hz)
Figure 6.25(a) Subject C: Force data in time and frequency domain for target
direction "left" and for a force level of 10 N.
122
8
10
EMG1 record
EMG2 record
3
2
2
-
E
I,
C-·-urS~UcllWwLkur
10
')Y'
UL4·~CILY-
20
Time (sec)
Time (sec)
Target-left
Force level-20 N
EMG1-biceps
EMG2-triceps
I
EMG3-deltoid(anterier)
10
20
30
Time (sec)
Figure 6.25(b) Subject C EMG data in time domain for target direction "left"
and for a force level of 10 N. Note: EMG data are for biceps,
triceps, and anterior deltoid.
123
Y.L--l
(Spectrum - EMG2)
0.5
0.5
0.4
0.4
-
0.3
0.3
-
0.2
0.2
-
0.1
0.1
-
I
0
2
4
6
Frequency (Hz)
8
0
10
I
I
I
I
2
4
6
8
Frequency (Hz)
(Spectrum - EMG3)·-
0.5
I
Target-left
0.4
E
Force level-20 N
0.3
0.2
EMG 1-biceps
EMG2-triceps
r
EMG3-deltoid(anterier)
0.1
I
I
,
,I
I
4
6
8
10
Frequency (Hz)
Figure 6.25(c) Subject C: EMG data in frequency domain for target direction
"left" and for a force level of 10 N. Note: EMG data are for biceps,
triceps, and anterior deltoid.
124
Time record- y force
Time record- x force
10
-10
-10
-20
-20
0
10
20
-
-
i-
30
10
20
Time (sec)
Time (sec)
(Q~nrtnim -v
I
\
0
2
4
6
3U
8
l-~--~NJ
il\
2
10
Frequency (Hz)
U
II
fnrc.
IrL
IV
\
I
I
I
I
I
4
6
8
10
Frequency (Hz)
Figure 6.26(a) Subject C: Force data in time and frequency domain for target
direction "left" and for a force level of 10 N.
125
EMG2 record
EMG1 record
3
2
1
II
10
-
-
Y1
·~·
·--
EMG3 record
3
~·I-YL--··
Time (sec)
Tune (sec)
4
~r·
30
20
-
-
Target-left
Force level-20 N
EMG1-biceps
EMG2-deltoid(anterier)
EMG3-deltoid(posterier)
10
20
30
Time (sec)
Figure 6.26(b) Subject C EMG data in time domain for target direction "left"
and for a force level of 10 N. Note: EMG data are for biceps,
triceps, and posterior deltoid.
126
-+L.UY ·-
-
L ·
(Spectrum - EMG2)
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0.2
-
0.1
0.1
-
0
2
4
6
8
10
4
2
Frequency (Hz)
6
8
Frequency (Hz)
(Spectrum - EMG3)
0.5
Target-left
0.4
Force level-20 N
0.3
0.2
EMG 1-biceps
EMG2-deltoid(anterier)
E
EMG3-deltoid(posterier)
0.1
7 ~ ŽJ
0
2
m
4
6
Frequency (Hz)
I
8
10
Figure 6.26(c) Subject C: EMG data in frequency domain for target direction
"left" and for a force level of 10 N. Note: EMG data are for biceps,
triceps, and posterior deltoid.
127
10
Time record- y force
Time record- x force
H
10
-10
-10
E
-20
-20
'
I
20
10
Time (sec)
Time (sec)
(Spectrum - y force)
(Spectrum - x force)
I
I
SII
I
I
I
H
6
6
4
2
\\
I I
2
6
4
Frequency (Hz)
a
8
\I
Nn
2
10
-
I
6
4
Frequency (Hz)
Figure 6.27(a) Subject C: Force data in time and frequency domain for target
direction "right" and for a force level of 10 N.
128
8
10
10
10
20
20
Time (sec)
Time (sec)
Target-right
Force level-10 N
EMG1-biceps
EMG2-triceps
EMG3-deltoid(posterier)
10
20
30
Time (sec)
Figure 6.27(b) Subject C EMG data in time domain for target direction "right"
and for a force level of 10 N.
129
30
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0.2
0.1
0.1
0
2
4
6
8
10
0
Frequency (Hz)
4
2
6
Frequency (Hz)
(Spectrum - EMG3)---
0C 5.-
-
I
I
I
Target-down
0.4
Force level-10 N
0.3
EMG 1-biceps
EMG2-triceps
0.2
EMG3-deltoid(posterier)
0.1
0
2
4
6
Frequency (Hz)
8
10
Figure 6.27(c) Subject C: EMG data in frequency domain for target direction
"right" and for a force level of 10 N.
130
8
10
Time record- y force
20
20
10
10
1
0
_____
)
CO
-10
-10
-20
-20
0
10
20
30
0
10
Time (sec)
30
Time (sec)
10
10
(Spectrum - y force)
N
8
8-
z 6
)
20
64-
4
o 2
0O
0
S 2
2
4
6
Frequency (Hz)
0
8
10
0
2
4
6
Frequency (Hz)
Figure 6.28(a) Subject C: Force data in time and frequency domain for target
direction "right" and for a force level of 20 N.
131
8
10
EMG1 record
I
EMG2 record
I
2
I
L
-h4LLL
0
10
-±
20
10
20
Timune (sec)
Time (sec)
,-,.,IIJEGd•LVJAJ
cU "4L
MlrAP
-o
I111.
Target-right
Force level-20 N
EMG1-biceps
EMG2-triceps
EMG3-deltoid(posterier)
10
20
30
Time (sec)
Figure 6.28 (b) Subject C EMG data in time domain for target direction "right"
and for a force level of 20 N.
132
30
(.neetnmm- F.al')
1
(Spectrum - EMG2)
0.8
I
I
4
6
0.8
E
0.6
0.6
0.4
0.4
0.2
0.2
2
4
6
Frequency (Hz)
8
-
|
....
10
Frequency (Hz)
Target-right
0.8
Force level-20 N
0.6
EMG1-biceps
0.4
EMG2-triceps
EMG3-deltoid(posterier)
0.2
2
4
6
Frequency (Hz)
8
10
Figure 6.28(c) Subject C: EMG data in frequency domain for target direction
"right" and for a force level of 20 N.
133
8
10
Chapter 7
Conclusions and Recommendations for Future Work
7.1 Conclusions
Several conclusions can be drawn from the experimental resutls. First, the experiment allow
us to determine those cases in which couplings between tremors at the elbow and the shoulder
do not exist. Nevertheless, the experiment cannot disclose if tremors at muscles other than the
minimum set are due to couplings or simply due to the subject's inability of using the minimum
set of muscles. Second, the experiment can be used to evaluate tremors, which a functional
test can never accomplish. The experiment can reveal whether tremor exists at all muscles or
merely at some certain muscles and in what condition the tremor at thoses muscle will be
initiated. The results from a functional test may show the degree of severity of the tremor and
patients' ability to perform daily activities. It cannot, however, locate the problem, which is
important in order for the therapist to adopt a particular scheme to assist the patient. Third, the
experiment may asssist physician to diagonize tremor since the experiment result might enable
the physician to trace the source of the tremor.
7.2 Recommendations for future work
In order to investigate the coupling between tremors at the elbow and the shoulder joints,
methods of selectively inhibiting muscle or reflex loop while keeping other muscles active
must be employed. In addition to that, more subjects with different etiologies should be
recruited so that a more general concusion can be deduced. In this experiment, the posture
was fixed; however, an similar pursuit tracking experiment which allows motion of the arm
may provide other useful information and add new findings to this experiment.
134
References
Adelstein, B. A Virtual Environment System for the Study of Human Arm Tremor. Ph.D.
Thesis, Department of Mechanical Engineering, MIT, June 1989.
Physiology, 420: 6P, 1990.
Basmajian and DeLuca. Muscles Alive. Baltimore: William and Wilkins, 1979.
Brumlik, J. and Yap, C-B. Normal Tremor: A Comparativestudy. Springfield, Il1: Charles
C Thomas, pp3-15, 1970.
Chen, A. Analysis of the Rehabilitationof Tremor Patients. M.S. Thesis, Sloan School of
Management, MIT, May, 1991.
Cooke, J. and Thomas, J. Fore-arm oscillation during cooling of the dentate nucleus in the
monkey. Can. J. Physiology. Pharmacol., 54:430-436, 1976.
Elble, R. and Koller, W. Tremor. Baltimore: The John Hopkins University Press, 1990.
Elble, R., Sinha, R., and Higgins, C. Quantification of tremor with a digitizing tablet.
JournalofNeuroscience Methods, 32:193-198, 1990.
Evarts, E. and Tanji, J. Gating of motor cortex reflexes by prior instruction. Brain Res.,
71:479-494, 1974.
Findley, L. Tremors: Differential diagnosis and pharmacology. Parkinson'sDisease and
Movement Disorders,Jankovic and Tolosa, eds. Baltimore: Urban & Schwarzenberg,
1988, pp. 243-262.
Findley, L. and Capildeo, R., eds. Movement Disorders: Tremor. New York: Oxford
University Press, 1984.
Frost, J. Triaxial vector accelerometry: A method for quantifying tremor and ataxia. IEEE
Transactionson Biomedical Engineering,BME-25:17-27, 1978.
Gilman, S., Carr, D. and Hollenberg, J. Kinematic effects of deafferentation and cerebellar
ablation. Brain, 99:311-330, 1976
Growdon, H., Chambers, W., and Liu, C. An experimental dyskinesia in the Rhesus
monkey. Brain, 90:603-632, 1967.
Hammond, p., Merton, P., Sutton, C. Nervous gradation of muscular contraction. Br. meed
Bull., 12:214-218, 1956.
Houk, J., Singer, J., Goldman, M. An evaluation of length and force feedback to soleus
muscles of decerebrate cats. Journalof Neurophysiology, 33:784-811,1970.
Jankovic, J. and Fahn, S. Physiologic and pathologic tremors. Annals ofInternal Medicine,
93:460-465, 1980.
135
Joyce, G. and Rack, P. The effects of load and force on tremor at the normal human elbow
joint. Journalof Physiology, 240:375-396, 1974.
Lee, R., Tatton, G. Motor responses to sudden limb displacements in primates with specific
CNS lesions and in human patients with motor system disorders. J.Can.Sci.Neurol,2:285293, 19075.
Marsden, C., Metrons, P., and Morton, H. Servo acton and stretch reflex in human muscle
and its apparent dependence on peripheral sensation. Journalof Phisiology,216:2122,1971.
Milner-Brown, H., Stein, R., and Lee, R. Synchronization of human motor units: possible
roles of exercise and supraspinal reflexes. Electoenceph. Cliln. Neurophysiol., 38: 245Ranish, N. and Soechting, J. Studies on the control of some simple motor tasks. Effects of
thalamic and red nuclei lesions. Brain Res., 102:334-345, 1976.
Rondot, P., Jedynack, C., and Ferrey, G. Pathologicaltremors: nosologicalcorrelates;in
Physiologycal Tremor, PathologicalTremor and Clonus, J.E. Desmedt, ed. Prog. Clin.
Neurophysiol., 5:95-113, Karger, Basel 1978.
Salzer, M. Three-dimensional tremor measurements of the hand. Journalofbiomechanics,
5:217-221, 1972.
Shahani, B. and Young, R. Action tremors: a clinical neurophysiological review; in
PhysiologicalTremor, Pathologicaltremor and Clonus, J.E. Desmedt, ed Prog. Clin.
Neurophysiol., 5:95-113, Karger, Basel 1978.
Stiles, R. and Bawa, P. The reflex responses of human soleus muscles to small perturbations.
Journal of Neurophysiology, 39:1105-1116, 1976.
Stein, R. and Oguztoreli, M. Tremor and Other Oscillations in neuromuscular systems. Blio.
Cybernetics, 22: 147-157, 1976.
Vilis, T. and Hore, J. Effects of changes in mechanical state of limb on cerebellar intention
tremor. Journalof Neurophysiology, 40(5):1214-1224, 1977.
Will, A. et al. Quantitative analysis of tremor and chorea using the VPL Data Glove.
ProgramandAbstracts, FourthAnnual Symposium, PSG/MDS. Annals of Neurology,
28(2):299, 1990
136