You are on page 1of 22

Gait Dynamics in Neurodegenerative diseases

ACKNOWLEDGEMENT

I express my heartfelt gratitude to The Almighty, The Supreme Guide, for bestowing his blessing
on me in my entire Endeavour.

I extend my sincere thanks to Dr. Gopakumar. A, Professor and Head of the Department,
Electronics and Communication Engineering, for encouraging and aiding me in successfully
presenting this seminar.

I convey my deepest gratitude to my seminar coordinator Mrs. Mredhula.L, Associate Professor,


ECE for giving me the confidence and opportunity to go ahead with this seminar.

I am highly indebted to my guide for the seminar, Mr. Edet Bijoy. K, Assistant Professor, ECE
for believing in me and showing me the right path for presenting the Seminar.

Finally I convey my gratitude to my dear friends, lecturers and all those people who served as
guiding light in completion of this Seminar.

Department of ECE, MESCE


Gait Dynamics in Neurodegenerative diseases

ABSTRACT

Human brain is the command center of the human body. It receives input from sensory organs
and sends outputs to the muscles. The received muscles at the synovial joints of the muscles
cause the contraction of muscles, which results in the displacement of the legs. The degeneration
of neurons in the human brain causes disturbance or imbalance of the body and walking manner
or Gait. The disturbance of gait dynamics of a an healthy person and an patient suffering from
neurodegenerative disease is being detailed throughout this report.

Neurodegenerative diseases are of many types, two of the familiar ones are Alzheimer’s (AD)
and Parkinson’s diseases (PD). Even though these two exhibit degeneration or death of neuron’s
PD exhibit huge variation in the gait pattern of human. This technique is not a solution for earlier
recognition of the disease as the disturbance in the gait will be seen only after four or five years
of occurrence of the disease. Gait measurement can be helpful for doctor’s to recommend
various measures to improve the body from effects of PD in the body.

Measurement of Gait dynamics make use of Vertical ground reaction force (VGRF) to determine
the force involved during the locomotion of a patient affected with PD. 16 load sensors are being
used to determine the force. The multi sensor data are being analyzed and modeled using Tensor
decomposition. Tensor decomposition include various techniques such as Canonical polyadic
decomposition (CPD), Tucker decomposition.

Department of ECE, MESCE


Gait Dynamics in Neurodegenerative diseases

TABLE OF CONTENTS

CHAPTER
TITLE PAGE NO:
NO:

ACKNOWLEDGEMENT 1
ABSTRACT 2
TABLE OF CONTENTS 3
LIST OF FIGURES 4
1 INTRODUCTION 5
1.1 MOTIVATION 5
1.2 LITREATURE SURVEY 5
1.3 OVERVIEW OF REPORT 6
2 NEURO DEGENERATIVE DISEASES 7
INTRODUCTION TO NEUROLOGICAL
2.1 7
DISORDERS
2.2 CAUSES OF NEUROLOGICAL DISORDERS 8
2.3 PARKINSON DISEASE 9
2.4 CAUSES OF PARKINSON DISEASE 10
2.5 SYMPTOMS OF PARKINSON DISEASE 10
2.6 TREATMENT FOR PARKINSON 11
3 GAIT DYNAMICS 13
3.1 INTRODUCTION TO GAIT DYNAMICS 13
3.2 GAIT ANALYSIS : METHODOLOGY 14
4 TENSOR DECOMPOSITION 18
4.1 SINGULAR VALUE DECOMPOSITION 18
4.2 HIGHER ORDER SVD 20
5 CONCLUSION 21
REFERENCE 22

Department of ECE, MESCE


Gait Dynamics in Neurodegenerative diseases

LIST OF FIGURES

FIGURE TITLE PAGE


NO: NO:

2.1 DIAGRAM OF MOTOR NEURON 7


2.2 DIAGRAM SHOWING HUMAN CELL AND CHROMOSOME 8
3.1 FLOW OF NERVE SIGNAL CAUSING MUSCLE 13
CONTRACTION
3.2 SENSOR PLACING OVER LF AND RF 15
3.3 VGRF FROM 8 SENSORS AT THE RF OF HC 16
3.4 VGRF FROM 8 SENSORS AT THE RF OF PD 17
3.5 PHASE IN A GAIT CYCLE 18
4.1 IMAGE COMPRESSED INTO 3X2 MATRIX 20
4.2 HOSVD 21

Department of ECE, MESCE


Gait Dynamics in Neurodegenerative diseases

CHAPTER 1

INTRODUCTION
Chapter deals with overall view of the gait dynamics in a healthy person and a patient suffering
from PD. The chapter also dealt with neurodegenerative diseases

1.1 MOTIVATION

Neurodegenerative diseases are common disease still under research having variety of disease
which are still haven’t determined. There are various diseases like Motor neuron disease (MND),
Alzheimer’s (AD), Parkinson disease (PD). Even though the symptoms and side effects are well
known the cause and treatment are still seems to be undetermined. Some of studies says that
these disease are caused due to genetically inherited and due to the environmental factors. The
risk taken by the brain can also cause these types of disease.

Parkinson disease shows the gait disturbance in a vibrant manner as the Parkinson is caused due
to the lack of dopamine neurons in the Substantia Nigra (SN) portion of the brain. dopamine
neuron contribute for the balance of the body. The lack of dopamine neuron may cause the
tremor of the hand and cause the leaning of the patient either backward or forward.

1.2 LITREATURE SURVEY

The paper dealt with neurodegenerative diseases and various types of neurodegenerative disease
involved in the analysis. PD was briefly mentioned and a comparison between the PD and
Healthy person was analyzed using various methodologies. One of the methodology was
explained in it. The method of modeling was also discussed in the paper. Tensor decompostition
was used as we require the processing of multi sensor data.
Contribution of paper as follows
 Paper gives a brief idea about the neurodegenerative disease and various types of disease.
 Paper also gives methodology use to determine the vertical ground reaction force (VGRF).
 Tensor decomposition was briefed and the application of Tensor decomposition in Gait
analysis was also well explained.

Department of ECE, MESCE


Gait Dynamics in Neurodegenerative diseases

1.3 OVERVIEW OF THE REPORT

Chapter 1 deals with the introduction and brief account of gait dynamics and application of gait
analysis in neurodegenerative diseases. Chapter also discussed about the literature survey done
during the research. Chapter 2 discussed about the neuro degenerative disease and gave a brief
account of causes and symptoms of neurodegenerative diseases. It also dealt with Parkinson
disease, causes of Parkinson disease, symptoms of Parkinson disease, and treatment for
Parkinson. Chapter 3 gave an introduction to gait dynamics and discussed the methodology used
for gait analysis. Chapter 4 discussed about singular value decomposition and higher order
singular value decomposition. Chapter 5 concluded the total research done for the seminar.

Department of ECE, MESCE


Gait Dynamics in Neurodegenerative diseases

CHAPTER 2
NEURO DEGENERATIVE DISEASE
Chapter deals with the Neuro degenerative disease and various types of neurodegenerative
disease. The scope of gait analysis in neurodegenerative disease is also being explained well
throughout the chapter.
2.1 INTRODUCTION TO NEURO LOGICAL DISORDERS

Neurological disorders are diseases of the central and peripheral nervous system. In other words,
the brain, spinal cord, cranial nerves, peripheral nerves, nerve roots, autonomic nervous system,
neuromuscular junction, and muscles. The development of the human brain begins during
pregnancy and continues through infancy, childhood and adolescence. Most brain cells are
formed before birth but the trillions of connections between these nerve cells (neurons) are not
developed until infancy.

Fig 2.1 : Diagram of a motor neuron


The brain is composed of gray matter (neurons and interconnections) and white matter (axons
surrounded by a myelin sheath). A motor neuron (above) carries impulses away from the
brain.The brain is self-organizing. It selects information to forward its growth and development.
It also adapts to the environment. Experience of the environment through the senses of touch,
smell, sight, taste and hearing produces connections in the brain.

Department of ECE, MESCE


Gait Dynamics in Neurodegenerative diseases

All neurologic disorders involve the brain, spinal column or nerves. Symptoms depend on where
damage occurs. Areas that control movement, communication, vision, hearing or thinking can be
affected.Neurologic disorders are wide ranging. They have various causes, complications and
outcomes. Many result in additional needs requiring life-long management.Symptoms of
neurologic disorders vary. Physical, cognitive (or thinking), emotional and behavioral symptoms
may be present, with specific disorders having combinations or clusters of these symptoms. For
example, cerebral palsy tends to have more physical symptoms whereas ADHD tends to have
greater effects on behavior.

Many neurologic disorders emerge during the early years of development and may be diagnosed
at birth. Some are diagnosed later because symptoms only appear when:

 A child misses developmental milestones or has developmental difficulties (e.g. autism).


 A damaging infection occurs (e.g. meningitis).
 An accident causes brain injury (stroke, trauma, hypoxia).

Fig 2.2: Diagram showing human cell and chromosomes

2.2 CAUSES OF NEUROLOGICAL DISORDER


Many neurologic disorders are “congenital,” meaning they were present at birth. But some of the
disorders are “acquired,” which signifies that they developed after birth. Those with an unknown
cause are termed “idiopathic.”

Department of ECE, MESCE


Gait Dynamics in Neurodegenerative diseases

Congenital Causes (present at birth): Genetic factors can influence the development of a
variety of neurologic disorders that are typically inherited from parents through genes and
chromosomes.Chromosomes are long strands of DNA supported by protein that are found in the
nuclei of human cells.Sections of DNA called genes carry the chemical code which makes us
who we are. Chromosomes are composed of thousands of genes.

A human body cell normally contains 46 (23 pairs) of chromosomes, half inherited from the
mother and half from the father.

2.3 PARKINSON DISEASE

Parkinson's disease is a neurodegenerative disorder, which leads to progressive deterioration of


motor function due to loss of dopamine producing brain cells.The cause of Parkinson’s Disease
is unknown, but researchers speculate that both genetic and environmental factors are involved.
Some genes have been linked to the disease. Primary symptoms include tremor, stiffness,
slowness, impaired balance, and later on a shuffling gait.

Some secondary symptoms include anxiety, depression and dementia. Parkinson's disease is the
second most common neurodegenerative disorder and the most common movement disorder.
Characteristics of Parkinson’s disease are progressive loss of muscle control, which leads to
trembling of the limbs and head while at rest, stiffness, slowness, and impaired balance. As
symptoms worsen, it may become difficult to walk, talk, and complete simple tasks.

The progression of Parkinson's disease and the degree of impairment vary from person to person.
Many people with Parkinson's disease live long productive lives, whereas others become
disabled much more quickly. Complications of Parkinson’s such as falling-related injuries
or pneumonia can cause premature death. However, studies of patent populations with and
without Parkinson’s disease suggest the life expectancy for people with the disease is about the
same as the general population.

Department of ECE, MESCE


Gait Dynamics in Neurodegenerative diseases

Most individuals who develop Parkinson's disease are 60 years of age or older. Since overall life
expectancy is rising, the number of individuals with Parkinson's disease will increase in the
future.

2.4 CAUSES OF PARKINSON DISEASE

A substance called dopamine acts as a messenger between two brain areas - the substantia nigra
and the corpus striatum - to produce smooth, controlled movements. Most of the movement-
related symptoms of Parkinson's disease are caused by a lack of dopamine due to the loss of
dopamine-producing cells in the substantia nigra. When the amount of dopamine is too low,
communication between the substantia nigra and corpus striatum becomes ineffective, and
movement becomes impaired; the greater the loss of dopamine, the worse the movement-related
symptoms. Other cells in the brain also degenerate to some degree and may contribute to non-
movement related symptoms of Parkinson's disease.

Although it is well known that lack of dopamine causes the motor symptoms of Parkinson's
disease, it is not clear why the dopamine-producing brain cells deteriorate. Genetic and
pathological studies have revealed that various dysfunctional cellular processes, inflammation,
and stress can all contribute to cell damage. In addition, abnormal clumps called Lewy bodies,
which contain the protein alpha-synuclein, are found in many brain cells of individuals with
Parkinson's disease.

2.5 SYMPTOMS OF PARKINSON DISEASE

The primary symptoms of Parkinson's disease are all related to voluntary and involuntary motor
function and usually start on one side of the body. Symptoms are mild at first and will progress
over time. Some individuals are more affected than others are. Studies have shown that by the
time that primary symptoms appear, individuals with Parkinson's disease will have lost 60% to
80% or more of the dopamine-producing cells in the brain. Characteristic motor symptoms
include the following:

Department of ECE, MESCE


Gait Dynamics in Neurodegenerative diseases

 Tremors: Trembling in fingers, hands, arms,feet, legs, jaw, or head. Tremor most often occur
while the individual is resting, but not while involved in a task. Tremor may worsen when an
individual is excited, tired, or stressed.
 Rigidity: Stiffness of the limbs and trunk, which may increase during movement. Rigidity
may produce muscle aches and pain. Loss of fine hand movements can lead to cramped
handwriting (micrographia) and may make eating difficult.
 Bradykinesia: Slowness of voluntary movement. Over time, it may become difficult to
initiate movement and to complete movement. Bradykinesia together with stiffness can also
affect the facial muscles and result in an expressionless, "mask-like" appearance.
 Parkinsonian gait: Individuals with more progressive Parkinson's disease develop a
distinctive shuffling walk with a stooped position and a diminished or absent arm swing. It
may become difficult to start walking and to make turns. Individuals may freeze in mid-stride
and appear to fall forward while walking.

While the main symptoms of Parkinson's disease are movement-related, progressive loss of
muscle control and continued damage to the brain can lead to secondary symptoms. These vary
in severity, and not every individual will experience all of them. Some of the secondary
symptoms include:

 Anxiety, insecurity, and stress


 Confusion, memory loss and dementia (more common in elderly individuals)
 Constipation
 Depression
 Excessive Salivation
 Diminished Sense Of Smell
 Increased sweating

2.6 TREATMENT FOR PARKINSON

There is currently no treatment to cure Parkinson's disease. Several therapies are available to
delay the onset of motor symptoms and to ameliorate motor symptoms. All of these therapies are
designed to increase the amount of dopamine in the brain either by replacing dopamine,

Department of ECE, MESCE


Gait Dynamics in Neurodegenerative diseases

mimicking dopamine, or prolonging the effect of dopamine by inhibiting its breakdown. Studies
have shown that early therapy in the non-motor stage can delay the onset of motor symptoms,
thereby extending quality of life.

The most effective therapy for Parkinson's disease is levodopa ,which is converted to dopamine
in the brain. However, because long-term treatment with levodopa can lead to unpleasant side
effects (a shortened response to each dose, painful cramps, and involuntary movements), its use
is often delayed until motor impairment is more severe. Levodopa is frequently prescribed
together with carbidopa (Sinemet), which prevents levodopa from being broken down before it
reaches the brain.

In earlier stages of Parkinson's disease, substances that mimic the action of dopamine (dopamine
agonists), and substances that reduce the breakdown of dopamine (monoamine oxidase type B
(MAO-B) inhibitors) can be very efficacious in relieving motor symptoms. Unpleasant side
effects of these preparations are quite common, including swelling caused by fluid accumulation
in body tissues, drowsiness, constipation dizziness, hallucination and nausea.

For some individuals with advanced, virtually unmanageable motor symptoms, surgery may be
an option. In Deep Brain Stimulation (DBS), the surgeon implants electrodes to stimulate areas
of the brain involved in movement. In another type of surgery, specific areas in the brain that
cause Parkinson's symptoms are destroyed.

An alternative approach that has been explored is the use of dopamine-producing cells derived
from stem cells. While stem cell therapy has great potential, more research is required before
such cells can become of therapeutic value in the treatment of Parkinson's disease.

In addition to medication and surgery, general lifestyle changes (rest and exercise), physical
therapy, occupational therapy, and speech therapy may be beneficial.

Department of ECE, MESCE


Gait Dynamics in Neurodegenerative diseases

CHAPTER 3
GAIT DYNAMICS

Chapter deals with discussion on Gait manner of a healthy person and a person affected with
Parkinson disease. VGRF is also being discussed throughout the chapter.

3.1 INTRODUCTION TO GAIT DYNAMICS

Human gait may be defined as a “series of rhythmical, alternating movements of the trunk and
limbs which result in the forward progression of the center of gravity of the body A gait cycle
refers to a single sequence of functions by one limb. Gait cycle begins when the reference foot
contacts the ground and ends with subsequent floor contact of the same foot. Study of gait
patterns of an Healthy person and person suffering from neurological problems may help to
determine the stage and status of the disease. It can also help to improve the treatment even
though neurological disorders are incurable.

Fig 3.1 : Flow of nerve signals which cause muscle contraction

Department of ECE, MESCE


Gait Dynamics in Neurodegenerative diseases

Flow of the nerve signals from the brain to the muscle which cause the muscle contraction and
thus displacement of the human body is as follows:

• Activation of signals in Central nervous system.


• Transmission of nerve signals to peripheral nervous system.
• Contraction of muscle that develop tension.
• Generation of forces and moments across, synovial joints.
• Displacement (i.e., movement) of the segments in a manner that is recognized as functional
gait.
• Generation of ground reaction forces.

3.2 GAIT ANALYSIS: METHODOLOGY

VGRF is being measured in order to find the gait patterns of the HC person and a PD patient.
VGRF is the force exerted by the ground on the person who gives a force over to the ground. It
follows Newton’s III law. The magnitude of VGRF will be equal to the weight of the person. up
to 20N can be neglected or set to zero as it can be considered as noise.

Gait is being analyzed using load sensors placed on both of the legs. Total of 16 load sensors are
equipped inside the sole of a shoe. Then the PD patient and a HC person is allowed to walk for 2
minutes with the sensor equipped shoe. The vertical force is being measured in Newton using
load sensors under each foot. If both of the legs are kept parallel to each other it is considered as
(x,y) coordinate system with values 0,0 for x and y respectively. That implies the person is in rest
and there is no variation in the VGRF.

When there occurs any displacement in the position of the legs soon the VGRF is being
measured. Multi sensor data’s are digitalized and stored in a recording unit placed over at the
waist. The recording unit is of size (19x14x14.5) and 1.5Kg carried on waist. The output of the
sensor is digitalized for 100 samples per second. There occurs variation in the VGRF of a person
suffering from PD while compared to a healthy person. The sensor at the heel will not have much
readings as compared to the sensor placed over at the toe as a PD patient keep leaning forward.
This could be helpful to determine the stage of the disease the patient is presently holding over.

Department of ECE, MESCE


Gait Dynamics in Neurodegenerative diseases

Fig 3.2: Sensor placing over both Left and Right foot

The multi sensor data obtained from the 16 load sensors are being modeled analyzed using
Tensor decomposition.

Fig 3.3: VGRF obtained from 8 sensors placed at the RF of HC

Department of ECE, MESCE


Gait Dynamics in Neurodegenerative diseases

Fig 3.4: VGRF from 8 sensors placed at the RF of PD patient

Figures 3.3 and 3.4 show short segments of the VGRF time series recorded with 8 sensors on the
right foot of an HC subject and a PD patient, respectively (the database consists of only records
of time and forces without annotations of phases of the gait cycle such as heel stride, foot flat,
mid-stance, heel-off, toe-off, and mid-swing). In a normal gait, the VGRF signal has two peaks:
the first peak is when the foot strikes the ground and the second peak is caused by the force
pushing off from the ground. The shape of the VGRF signal is abnormal in PD ,as can be seen in
Figure 3.4 (h), showing the VGRF measured with sensor #8 located under the right foot of a PD
subject, where narrow peaks in the VGRF signal can be observed. The magnitudes of the VGRF
signals of the PD subject are generally lower than those of the HC subject due to reduced forces
for the heel contact and the pushing-off phase.

Department of ECE, MESCE


Gait Dynamics in Neurodegenerative diseases

Fig 3.5: Phases in a Gait cycle

There are two types of phases included in a gait mainly stance phase and swing phase. It is the
natural process taking place during walking. First the stance phase occur and later the swing
phase. Stance phase happens when we strike the foot over the floor. Starts with heel strike
followed by foot flat, midstance, Heel off and toe off. Swing phase is the next phase happening
soon after the stance phase. There occurs the change of foot striking over the floor. The placing
of next leg over the floor by releasing the first contacted leg is called Swing phase.

Department of ECE, MESCE


Gait Dynamics in Neurodegenerative diseases

CHAPTER 4
TENSOR DECOMPOSITION

Chapter deals with the tensor decomposition which is used for modeling and analysis of multi
sensor data. Chapter will also gives a brief account of Singular Value Decomposition (SVD) and
HOSVD.

4.1 SINGULAR VALUE DECOMPOSITION (SVD)

Consider a mxn matrix which could be decomposed or factorized to three matrices, two
orthogonal matrix and a diagonal matrix. The diagonal matrix will be of the same order of the
original matrix. The diagonal elements of this matrix are called the singular values of the matrix.
These singular values are used to reconstruct the original matrix, the other values in the matrix
are being discarded.

 2 2 2 
 
 2 0   14 6 21  7 0  1 1 

   3 1  2  
 1 2    0 3  2 2
0 1  14 6 21   1 1 
  1

4  0 0  
 1  2 2 (1)

 14 6 21 
Equation (1) represent the matrix decomposition into three matrix, two orthogonal matrix and a
diagonal matrix. The diagonal values of diagonal matrix are called singular values. Solution to
obtain the matrix is as follows:

A  U  V T  u1 1v1  ......  ur r vr


T T
(2)

 U = Orthogonal matrix.
   Diagonal matrix.
 V T = Orthogonal matrix.

Department of ECE, MESCE


Gait Dynamics in Neurodegenerative diseases

Major application of decomposition is image compression , suppose that a satellite in space is


taking photographs of Jupiter to be sent back to the earth. The satellite digitalises the picture by
subdividing it into tiny squares called pixels or picture elements. Each pixel is represented by a
single number that records the average light intensity in the square. If each photograph was
divided into 400_400 pixels, it would have to send 160,000 numbers to the earth for each picture.
This would take great deal of time and would limit the number of photographs that could be
transmitted. It is much better, If we can approximate this matrix with a matrix which
requires less storage.

Fig 4.1 : Image compressed to a 3x2 matrix.

The image given above is being compressed over to an 3x2 matrix for easy transformation of
data from one place to another. Equation (3) represent the 3x2 matrix to which the image is being
compressed.

Department of ECE, MESCE


Gait Dynamics in Neurodegenerative diseases

 0.9572 0.1419 
 
 0.4854 0.4218 
 0.8003 0.9157 
  (3)

4.2 HIGHER ORDER SVD (HOSVD)

Higher order singular value decomposition is the same as SVD, beside using the singular values
for the reconstruction of the original matrix core tensors are being used. The equation of matrix
decomposition is given below.

(4)

The core tensors are being used to reconstruct the original matrix and other elements of the
matrix are discarded.

Fig 4.2: HOSVD

Department of ECE, MESCE


Gait Dynamics in Neurodegenerative diseases

CHAPTER 5
CONCLUSION

Neuro degenerative diseases are caused due to the degeneration of neurons in the human brain.
The degeneration of neurons may cause imbalance in the the walking manner or gait. Gait
disturbance id vibrantly seen in PD patient compared to Alzheimer’s patient. Gait dynamics of a
PD patient and an HC was clearly analyzed. Gait dynamics are analyzed for the feature
determination of the stage of the disease and to improve treatment methods to reduce the side
effects of this disease. VGRF of PD patients and HC person showed variation in magnitude of
force in case of PD patient. Tensor decomposition is being effectively used to analyze the multi
sensor data. HOSVD is being used to compress the data while the PARAFAC or CPD is being
used to analyze the data.

Department of ECE, MESCE


Gait Dynamics in Neurodegenerative diseases

REFERENCES

[1]. Pham, Tuan D., and Hong Yan. "Tensor Decomposition of Gait Dynamics in Parkinson's
Disease." IEEETransactionsonBiomedicalEngineering (2017).
[2]. Ren, Peng, et al. "Gait Influence Diagrams in Parkinson’s Disease." IEEE Transactions on
Neural Systems and Rehabilitation Engineering 25.8 (2017): 1257-1267.
[3]. Vuthy, Teav, and Angel R. Pineda. "Using the Singular Value Decomposition (SVD) for
ImageCompression."(2010).
[4]. Cichocki, Andrzej, et al. "Tensor decompositions for signal processing applications: From
two-way to multiway component analysis." IEEE Signal Processing Magazine 32.2 (2015): 145-
163.

Department of ECE, MESCE

You might also like