Anda di halaman 1dari 51

Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

1
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

Dedicated to my Beloved

Father P.Selvaraj

Mother S.Maheswari

Wife J.Ruby

Children J.Lipika J.Tisa J.Nedumaan

2
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

Chapter 1

3
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

Bi-directional Recurrent Neural Networks in Classifying

Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

P.S.Jagadeesh Kumar, Yanmin Yuan, Yang Yung,


Wenli Hu, Mingmin Pan, Xianpei Li
4
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

Preface

This chapter brings the gestalt of nature and pervasiveness of three


main neurodegenerative disorders; dementia, alzheimer’s disease and
autism spectrum disorder; the practice of living with those disorders;
ubiquity, impact and their diagnosis. The brain is vulnerable to a wide
range of clutters that strike at each phase of life. Brain disorders like
autism spectrum show up in early infantile. The chapter objectives to
provide the necessary classification of various dementias, alzheimer’s
disease and autism spectrum disorder using bidirectional recurrent
neural networks. The classification accuracy is measured using the
criterions; sensitivity and specificity. Again, the fallouts are compared
amid recurrent neural networks and bidirectional recurrent neural
networks. Based on the results, it can be exposed that bidirectional
recurrent neural network provided the needed efficient classification
between various brain disorders. The patients of various age groups,
right from the birth to 85 years of age were experimental. 5
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

1. Introduction

Knowledge is typically not prejudiced, in spite of the fact that there


are usually intellectual changes, for example, issues with memory,
fixation and contemplation. A brain disorder isn't a scholarly
incapacity. It is likewise not a psychological maladjustment, despite
the fact that it can expand the odds of mental issue, for example,
dejection and nervousness. Brain disorder are regularly called the
shrouded inability in light of the fact that there can be issues with
conduct and capacity to think, but there is frequently no obvious
physical change with numerous brain issues, so disputes can be not
entirely obvious or misconstrued [7]. Indeed, even awful brain damage
may leave no obvious scars to show damage occurred. A brain disorder
happens when there is harm or disturbance to the mind even from
birth. Long term impacts will be distinctive for every individual, and
will likewise shift contingent upon the kind of brain disorder.
Alzheimer's sickness causes issues with memory, considering and
conduct [12]. In the beginning time, dementia manifestations might be
insignificant, yet as the illness makes more harm the mind, side effects
exacerbate. While there are presently no medicines to prevent
Alzheimer's disease from advancing, there are pharmaceuticals to treat
dementia manifestations. In the previous three decades, dementia
investigate has given a considerably more profound comprehension of
how Alzheimer's influences the mind.
6
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

Brain infections, like schizophrenia are customarily analysed in


adolescents or early adulthood, in spite of the fact that their inceptions
may lay considerably before throughout everyday life. Because of
maturing one turn out to be gradually defenceless to Alzheimer's
disease, Parkinson's ailment, stroke, and different dementias [10].
Brain disorders are among the most genuine medical issues
confronting our general public, causing untold human enduring and
huge monetary expenses. They are additionally among the most
secretive all things considered, and the obliviousness of the
fundamental ailment instruments is a noteworthy obstruction to the
advancement of better medications. What are required is a legitimate
examination and classification of different brain disorders in view of
present day machine learning strategies that are not just aware, but
rather new remedial methodologies that depend on a major
comprehension of cerebrum systems. Subsequently, the translational
endeavours should along these lines lay on a solid establishment of
essential revelation look into, the method that drives new pragmatic
applications. In this chapter, the classification and their precision in
the midst of three notable neurodegenerative issues, for example,
dementia, alzheimer's disease and autism spectrum disorder exploiting
bidirectional recurrent neural network is customary along different
ages and maturing ideal from the birth to 85 years old.
7
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

2. Dementia

Dementia is a term exploited to portray an accumulation of signs


and indications, say, memory issues, correspondence distresses,
challenges with organizing and scheduling one's everyday life, changes
in state of mind and conduct, and the progressive loss of control of
physical capacities [4]. These manifestations, taken together, are a sign
of physical harm to the mind because of interminable dynamic
degeneration of nerve cells. The harm to the mind might be caused by
a wide range of sicknesses: while dementia is the commonest and best
known reason, there are numerous others, for example, Parkinson's
disease, vascular dementia, Lewy body dementia, front fleeting
dementia, and prion ailments to give some examples. Looking after a
man with memory misfortune or dementia is an excursion, one that can
keep going for a long time, and is loaded with wanders aimlessly. The
trip is certainly not a straight way through stages; every individual with
Alzheimer's disease or other dementia may advance differently and
time permitting [11]. Relatives who are colleagues on this excursion
will require data, support and direction. Dementia is a general term
alluding to lost psychological capacity; recalling, considering, and
thinking; sufficiently extreme to meddle with regular day to day
existence. Dementia is certainly not a particular illness, however a
general term portraying an extensive variety of side effects. It isn't
ordinary maturing. Dementia for the most part influences more
seasoned individuals, despite the fact that there is a developing
consciousness of cases beginning before the age of 65.
8
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

2.1 Ubiquity of Dementia

Dementia is dynamic, which implies the side effects will bit by bit
deteriorate and the condition is presently serious. In any case,
solutions and different mediations can reduce manifestations and
individuals may live with their dementia for two or three decades after
conclusion. In spite of normal discernments, there is an awesome
arrangement that should be possible to encourage individuals with
dementia [5]. If analysed timelily, individuals with dementia and their
carers can get the treatment, care and support; socially and mentally,
and also pharmacologically empowering them to live well with the
condition. Usually about the basic things that are underestimated, such
as influencing some tea without help; somebody with dementia can
discover for all intents and purposes incomprehensible. Dementia is
developing the worldwide test and the elements related with
expanded danger of dementia are appeared in Fig. 1. It is a standout
amongst the most imperative wellbeing and care issues the world faces
as the populace ages. All inclusive there is another instance of
dementia like clockwork, and by 2025 almost 70 million individuals will
live with the condition. Every individual is novel and will encounter
dementia in exceptional way. The distinctive kinds of dementia
likewise tend to influence individuals in an unexpected way,
particularly in the beginning times. Different variables that will
influence how well somebody can live with dementia incorporate how
other individuals react to them and the earth around them [13]. A
patient with dementia will have psychological side effects to do with
9
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

considering or memory. They will regularly have issues with a portion


of the accompanying or every one of them: everyday memory – for
instance, trouble reviewing occasions that happened as of late;
thinking and arranging – for instance, challenges deciding, taking care
of issues or doing a grouping of undertakings; visuospatial abilities –
for instance, issues judging separations and seeing articles in 3D;
introduction – for instance, forgetting about the day or date, getting
to be befuddled whereabouts.

Fig. 1 Dementia and its associated risk

10
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

2.2 Impact of Dementia

Dementia highly affects individuals living with the condition, their


families, carers and society all the more by and large. In more
extravagant nations around the globe, an audit of on-going
investigations led found that the middle analysis rate was 49%, running
from 35% to as high as 60%. The analysis rate for is right now 58%, yet
by 2020 this would absolutely ascend to 66% of individuals with
dementia being analysed and accepting proper help. Over all nations
on the planet an expected 38 million of the 46.5 million individuals
with dementia worldwide don't have a formal analysis. Around 70% of
individuals with dementia have Alzheimer's disease [28]. Roughly, 30%
have vascular dementia and numerous individuals have a blend of the
two. There are different types of dementia. Vascular dementia comes
about because of issues with the blood supply to the brain; without
enough blood, cerebrum cells can bite the dust. The impacts of
vascular dementia can be limited or counteracted by and large
through a solid way of life. Smoking and heftiness, for instance,
influence various sorts of dementia, specifically vascular dementia.
Helping individuals to comprehend the effect of their way of life could
assist them with making better, more educated decisions and lessen
their danger of creating vascular dementia.
11
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

2.3 Diagnosis of Dementia

It is imperative for any individual who has issues with their memory
or thinking to get a legitimate evaluation. These issues might be
caused by a treatable condition, say, sadness or a disease, as opposed
to dementia. Discerning the reason may enable the individual to get
the correct treatment. On the off chance that these issues are a result
of dementia, getting courage has numerous advantages. It gives
somebody a clarification for their manifestations, gives them access to
treatment, exhortation and bolster, and enables them to get ready for
the future and plan ahead. Knowing the sort of dementia for instance,
Alzheimer's disease or vascular dementia is likewise vital, mostly on the
grounds that it might enable the individual to get a suitable
medication treatment. By far most of reasons for dementia can't be
relieved, despite the fact that exploration is proceeding into creating
medications, immunizations and other restorative medicines [30].
There is a considerable measure that should be possible to empower
somebody with dementia to live well with the condition. Care and
support ought to be individual focused. This implies it ought to be
centered on that individual and their individual needs and inclinations.
There are a scope for non-medicate medicines accessible that can
assist somebody with living great with dementia. These incorporate
data, counsel, support, treatments and exercises. Support for the
individual and their carer ought to be accessible after an analysis.
Different sorts of treatment incorporate the accompanying:
12
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

 Talking treatments, example, counselling, can enable somebody


to grapple with their conclusion or talk about their sentiments.
Psychological social treatment might be offered if the individual
creates gloom or nervousness.
 Cognitive recovery can empower a person to hold abilities and
adapt better. There are likewise parts that should be possible at
home to assist somebody with dementia stay free and live well
with memory misfortune.

13
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

2.4 Women and Dementia

Given the more noteworthy predominance of dementia among


ladies, there is a developing acknowledgment of the part of sexual
orientation identifying with dementia, with rising proof proposing a
need to identify and organize dementia as a universal womens'
medical problem. In the writing about dementia, people are analysed,
however seldom with regards to sexual orientation parts and
characters [24]. Various approach improvements featuring the need to
address issues around sex and maturing have been produced,
alongside a more particular spotlight on the difficulties looked by
more established ladies, given the changing socio-economics of this
populace. Joint efforts between nations are creating strategies around
sexual orientation and maturing, to address sex disparity and
specifically the requirements of more seasoned ladies. It is essential to
perceive the requirement for sexual orientation balance for various
zones as the statistic profile changes. Inside wellbeing, it indicates the
need to recognize and bolster the part of providing care, regardless of
whether formal or casual, and the requirement for sexual orientation
suitable social insurance where important.
14
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

2.5 Living with Dementia

Keeping in mind that individuals determined to have dementia are


not totally vulnerable, the variables that add to the well living of
patients with dementia are;

 As dementia advance is genuinely incessant, live the steps and


adapt to the incredibly.
 Involve the individual with the analysis however much as could
reasonably be expected in tentative arrangements for their care.
Plan things ahead of time with regards to future care so later on,
relatives will have the capacity to take after their cherished one's
arrangement.
 Likewise, when vision begins to decline and individuals wear
glasses to peruse; those with dementia utilize memory devices
like cautions and daily agendas to recall things like birthday
celebrations, arrangements and to take their pharmaceutical
[29].
 It is feasible for a patient with direct dementia to keep on
learning and utilizing the memory frameworks like breaking
bigger assignments into little advances, giving visual and verbal
signals to regular exercises, streamlining activities and schedules,
distinguishing and participating in exercises that are vital to the
dementia patients.
 Design the space where the patients with dementia lives to
15

empower him or her to work well.


Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

3. Alzheimer’s Disease

Alzheimer's disease is a type of brain degeneration in which irregular


particles called neurofibrillary tangles and psychotic plaques frame in
the mind which pulverizes solid neuron and brain cell. These variations
from the norm tend to settle in cerebrum zones that control the
capacity to take in another reality and recall them. The two primary
classes of Alzheimer's disease (AD) are familial and sporadic. Familial
Alzheimer's disease alludes to a hereditary type of the infection that is
transmitted starting with one age then onto the next [2]. Just 6% of all
instances of Alzheimer's ailment have been related with a hereditary
part. These people originate from families in which the same number
of as half of the individuals builds up Alzheimer's malady. Fortunately,
this type of the confusion is unusual. The staying 94% of Alzheimer's
ailment cases are sporadic, or haphazardly happening in the populace.
An assortment of hazard factors have been distinguished in people
determined to have AD. While a portion of these components may end
up being helpful, recall that a significant part of the exploration that
has been done around [14]. This implies the exploration is directed by
contrasting a gathering of patients determined and AD to have a
gathering of solid age coordinated grown-ups. These kinds of
investigations give data about the quantity of people determined to
have AD who have a specific trademark contrasted and the solid
people. While these outcomes are valuable in coordinating future
research considers, they don't give data about circumstances and end
16

results.
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

3.1 Prevalence of Alzheimer’s Disease

Alzheimer's disease affronts basic metabolic procedures that keep


neurons solid. These disturbances cause nerve cells in the cerebrum to
stop working, lose associations with other nerve cells, lastly bite the
dust [17]. The decimation and passing of nerve cells causes the
memory disappointment, identity changes, issues in completing day by
day exercises, and different highlights of the ailment. The brains of
individuals with AD have a wealth of two unusual structures; amyloid
plaques and neurofibrillary tangles that are made of misfolded
proteins. This is particularly valid in specific locales of the cerebrum
that are critical in memory. The third primary element of AD is the loss
of associations between cells. This prompts decreased cell capacity
and cell passing. Trail starts somewhere down in the brain, in the
entorhinal cortex, a mind area that is close to the hippocampus and
has guide associations with it. Sound neurons in this region start to
work less productively, lose their capacity to convey, and at last kick
the bucket as appeared in Fig 2. This procedure bit by bit spreads to
the hippocampus, the cerebrum locale that assumes a noteworthy part
in learning and is engaged with changing over here and now
recollections to long term recollections. Influenced areas start to
decay. Ventricles, the liquid filled spaces inside the mind, start to
amplify as the procedure proceeds [24]. Sooner or later, the harm
happening in the mind starts to show itself in early clinical signs and
indications. Much research is being done to recognize these early
17

changes, which might be helpful in foreseeing dementia or AD. An


Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

imperative piece of this examination exertion is the improvement of


progressively refined neuroimaging strategies and the utilization of
biomarkers. Biomarkers are markers, for example, changes in tangible
capacities, or substances that show up in body liquids, for example,
blood, cerebrospinal liquid. Biomarkers can demonstrate introduction
to a substance, the nearness of an ailment, or the movement after
some time of an infection. For example, high blood cholesterol is a
biomarker for danger of coronary illness. Such strategies are basic to
helping researchers identify and comprehend the early signs and side
effects of AD.

As few people get aged, they create memory issues more prominent
than those normal for their age. Be that as it may, they don't encounter
the identity changes or different issues that are normal for AD. These
individuals may have a condition called mild cognitive impairment
(MCI). Individuals with MCI are a basically critical gathering for
investigate on the grounds that a considerably higher level of them go
ahead to grow AD than do individuals without these memory issues.
18
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

Fig. 2 Prevalence of Alzheimer’s disease

19
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

3.2 Influence of Alzheimer’s Disease

In the vast majority with Alzheimer’s disease, manifestations initially


show up in their mid-60s. Alzheimer's illness is right now positioned as
the 6th driving reason for death, however late gauges show that the
turmoil may rank third, simply behind coronary illness and tumour, as a
reason for death for more established individuals. Scientists don't yet
completely comprehend what causes Alzheimer's illness in a great
many people. In individuals with early-beginning Alzheimer's, a
hereditary change is normally the reason. Late-beginning Alzheimer's
emerges from an unpredictable arrangement of cerebrum changes
that happen over decades. The causes most likely incorporate a mix of
hereditary, ecological, and way of life factors. The significance of any
of these components in expanding or diminishing the danger of
building up Alzheimer's may contrast from individual to individual.
One of the immense puzzles of Alzheimer's disease is the reason it to a
great extent strikes more established grown-ups. Research on typical
cerebrum maturing is revealing insight into this inquiry [23]. For
instance, researchers are figuring out how age-related changes in the
mind may hurt neurons and add to Alzheimer's. These age-related
changes incorporate contracting of specific parts of the mind,
aggravation, creation of unsteady particles called free radicals, and
mitochondrial brokenness; a breakdown of vitality generation inside a
phone. A great many people with down disorder create Alzheimer's.
20
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

3.3 Diagnosis of Alzheimer’s Disease

Alzheimer's disease is intricate, and it is improbable that any one


medication or other intercession will effectively treat it [25]. Existing
tactics center around helping individuals keep up mental capacity,
oversee conduct side effects, and moderate or defer the indications of
ailment. Specialists want to create treatments focusing on particular
hereditary, sub-atomic, and cell components with the goal that the
genuine basic reason for the sickness can be ceased or counteracted.

 Common signs of Alzheimer's disease include restlessness,


meandering, disturbance, nervousness, and hostility. Researchers
are realizing on these indications and are anticipating new
medicines; medication and nondrug to oversee them [24].
Research has demonstrated that treating conduct manifestations
can make individuals with Alzheimer's more agreeable and
makes things less demanding for guardians.
 Alzheimer’s disease research has shaped to a point where
researchers can look past treating side effects to consider
tending to basic sickness forms. In on-going clinical
preliminaries, researchers are creating and testing a few
conceivable intercessions, including vaccination treatment;
tranquilize treatments, subjective preparing, physical action, and
medicines utilized for cardiovascular sickness and diabetes.
21
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

3.4 Women and Alzheimer’s Disease

Alzheimer's disease can likewise be named a complex polygenetic


illness, for example, hypertension, coronary corridor sickness, or
schizophrenia, on the grounds that there are both hereditary and
ecological components that outcome in the phenotypic articulation of
what is clinically baptized as Alzheimer’s disease. Because of the
statistic shifts and noteworthy therapeutic advances of the most recent
century, maturing populaces are currently the quickest developing bits
of the populace, and ladies constitute a larger part of this populace
because of their expanded future in respect to guys. Accordingly,
women would give off an impression of being at scarce hazard for AD.
Meta-investigation of populace pervasiveness ponders found no sex
contrast in general dementia commonness by sexual orientation, yet
demonstrated that AD rates have a tendency to be higher in females,
and vascular dementia has a tendency to be more typical in men [22].
A meta-examination of occurrence thinks about crosswise over various
nationalities indicated little distinction in general sexual orientation
particular rate. Flyer rate for ladies demonstrated a more extreme age-
occurrence bend, showing a higher frequency at more seasoned ages.
Studies are as of now in progress in geologically or hereditarily
secluded populaces, that may help decide if hereditary or natural
elements, for example, poisons, proficiency and instruction, physical
action, smoking, or different factors are vital in deciding AD
improvement. One methodological issue in these investigations,
22

predominantly in developing countries, is the blaming impact of


Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

contending wellsprings of mortality, which would tend to support AD


articulation in ladies [27]. All the more inconspicuously, the beneficial
encounters of women, their instructive and professional openings, and
their entrance to human services, solutions, and ailment exposures may
affect their dangers of AD in ways that ebb and flow look into is just
start to understand. In view of expanded life span and the exceptional
difficulties of estrogen substitution treatment, ladies are all around put
to both is in danger for and the recipients of advances in AD
treatment. The natural reason for this distinction in dialect capacities
in AD might be because of a few factors, for example;

 Women with AD have a considerably more grounded connection


of infection state with post dexamethasone cortisol levels than
men.
 Regional cerebral blood streams estimated by positron discharge
tomography in AD have been depicted, with ladies having higher
mean non-weighted metabolic rates than men [21]. The
clarifications for these wonders incorporate contrasts in the
provincial example of cell misfortune and β-amyloid affidavit,
which is coupled to territorial cerebral blood stream.
 Dexamethasone obstruction has been related with harm to the
hippocampal procedure, which is personally engaged with short
term memory preparing. Gauge contrasts in neural association,
like, expanded lateralization in men, may influence examples of
dialect breakdown, getting to be amplified with illness
23

movement.
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

3.5 Living with Alzheimer’s Disease

Alzheimer's is exploited to portray the slow crumbling of scholarly


capacities and conduct that in the end meddles with standard day by
day living exercises. Standard day by day living exercises incorporate
adjusting the check book, keeping house, driving the auto,
contribution in social exercises, and working at one's typical
occupation. There may likewise be changes in identity and feelings. In
opposition to mainstream thinking, AD is definitely not a typical result
of maturing, yet is caused by infections that influence the cerebrum
[20]. It impacts all parts of brain and conduct, including memory,
judgment, dialect, fixation, visual recognition, demeanour, and social
cooperation. The side effects are in the end evident to everybody, in
the beginning periods uncommon assessments are important to exhibit
the variations from the norm. The accompanying points in overseeing
and living with Alzheimer's malady;

 Assisted living, a kind of lodging for individuals who require help


with some medicinal and individual care yet can in any case
handle a few sections of day by day living without anyone else.
 Providing drug updates.
 24-hour care if there should be an occurrence of serious AD
 Assisting in communal and entertaining accomplishments.
 To address a considerable measure of necessities, including
mysticism and diet.
24
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

4. Autism Spectrum Disorder

Autism spectrum disorder (ASD) is a neurodevelopmental issue


characterized by relentless deficiencies in social correspondence and
social cooperation as represented in Fig. 3, joined by confined,
repetitive examples of conduct, interests, or exercises. The expression
spectrum alludes to the extensive variety of manifestations, abilities,
and levels of hindrance, or incapacity, that youngsters with ASD can
have. A few kids are somewhat weakened by their indications, yet
others are seriously handicapped [6]. The indications of ASD are
typically obvious in early adolescence. Intense it is as yet thought
about a long lasting determination, with suitable early mediation;
people with ASD can lead profitable, comprehensive, and satisfying
lives. Side effects of extreme introvertedness range issue fluctuate
starting with one tyke then onto the next, yet when all is said in done,
they fall into three zones: social weakness, communication challenges,
repetitive and stereotyped practices. ASD right now characterizes five
issues, here and there called pervasive developmental disorders
(PDDs):

 Asperger’s disorder (Asperger syndrome)


 Pervasive developmental disorder not otherwise specified (PDD-
NOS)
 Rett’s disorder (Rett syndrome)
 Childhood disintegrative disorder (CDD)
25
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

Fig. 3 Autism and its deficits

26
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

4.1 Children and Autism

In accordance to the inquest, there is no known single reason for


extremely introverted confusion; however it is by and large
acknowledged that it is caused by irregularities in cerebrum structure
or capacity. Cerebrum examines indicate contrasts in the shape and
structure of the mind in extremely introverted versus mentally
unbalanced kids. Specialists are researching various speculations,
including the connection between genetics, hereditary qualities and
therapeutic issues. In numerous families, there have all the earmarks of
being an example of extreme introvertedness or related incapacities,
moreover supporting a hereditary premise to the confusion. While no
quality has been famed as causing risky introvertedness, specialists are
hunting down unpredictable sections of hereditary code that mentally
unstable kids may have acquired. It also creates the impression that a
few youngsters are conceived with a vulnerability to extreme
introvertedness; however analysts have not yet recognized a solitary
trigger that makes a mental imbalance [9]. Childs with ASD don't take
after the regular examples of child advancement. In a few youngsters,
clues of future issues might be obvious from birth. Much of the time,
the issues in correspondence and social aptitudes turn out to be more
discernible as the kid falls encourage behind other kids a similar age.
Some other youngsters begin off all around ok. As a rule in the vicinity
of 12 and three years old, the distinctions in the way they respond to
individuals and other unordinary practices wind up obvious.
27
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

4.2 Occurrence of Autism

It is distinguished that a recurrence rate for those with a mental


imbalance and going with learning incapacity as pretty much 40 in
each 15,000 youngsters. Concentrates found a pervasiveness of
advanced a mental imbalance and Asperger disorder of 65 for each
15,000 and another 56 for every 10,000 with social weaknesses which
fit into the mental imbalance range. Numerous examinations are
joined to appraise recurrence of a mental imbalance range issue on the
planet nations at 201 for each 15,000 head of populace. Late littler
scale considers have affirmed the same through directing an overview
of 2-multi year olds more than 50,000 youngsters [16]. They found a
recurrence rate of 5 out of 500. Roughly, 15% of the uncommon needs
populace had analyses of a mental imbalance range issue and around
60% of standard grade schools have no less than one understudy with
an extreme introvertedness range issue.

28
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

4.3 Diagnosis of Autism

Diagnosing autism spectrum disorder (ASD) can be troublesome,


since there is no salutary test, similar to a blood test, to analyse the
clutters. Specialists take a gander at the tyke's conduct and
advancement to make an analysis [9]. ASD can now and then be
recognized at year and a half or more youthful. By age 2, a conclusion
by an accomplished proficient can be viewed as extremely solid. Be
that as it may, numerous kids don't get a last determination until
significantly much older. This defer implies that youngsters with an
ASD won't not get the assistance they require. Diagnosing an ASD
happens in two stages:

 Developmental Screening
 Comprehensive Diagnostic

Developmental screening is a short test to tell if youngsters are


learning fundamental abilities when they should, or on the off chance
that they may have delays. Amid formative screening the specialist may
ask the parent a few inquiries or talk and play with the kid amid an
exam to perceive how she learns, talks, carries on, and moves. A
deferral in any of these territories could be an indication of ASD. The
second step of finding is a comprehensive diagnostic. This exhaustive
audit may incorporate taking a gander at the kid's conduct and
improvement and meeting the guardians. It might likewise incorporate
a hearing and vision screening, hereditary testing, neurological testing,
29

and other therapeutic testing.


Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

4.4 Living with Autism

To comprehend and live with ASD, one must be aware of the core
deficits that describe extreme introvertedness and hinder the
advancement of the central essential aptitudes central for learning [6].
Some unique care for patients with a mental imbalance may include;

 Improve consideration challenges


 Analyse sound-related handling impedances.
 Increase ability to sum up learning exchange.
 Lessen challenges with learning by acuity and impersonation.
 Minimize issues with assignment/occasion sequencing.
 Stabilize even examples of qualities and shortcomings.
 Subordinate issues with association and arranging.
 Subsidize challenges with time ideas and making advances.

30
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

4.5 Child to Adult Autism

What happens when somebody with autism spectrum disorder (ASD)


leaves school and influences the progress to grown-up to
administrations, school, work, work preparing, or another living
circumstance? What does look into say in regards to the issues that
influence grown-ups with ASD and their families? The street to
adulthood formally starts for some teenagers when they graduate.
Hitherto, for individuals with a mental imbalance, leaving secondary
school is a more fantastic advance, one that will change their
relationship to managements and backings [16]. Enough number of
youngsters with extreme introvertedness battle for vital living
including cleanliness, riding a transport, shopping, and setting up a
supper. Specialists approximate it's essential to center on educating
such aptitudes as a key to autonomy. Numerous battles to acclimate to
the difficulties of school: dormitories, freedom, intense classes and
another social world. Be that as it may, for individuals with extreme
introvertedness, the change can be more emotional. The versatile
aptitudes that adolescents and youthful grown-ups with a mental
imbalance range issue need to obtain to enable them to lead
protected, beneficial, and satisfying grown-up lives. These aptitudes
include security, cleanliness, work, social capability, basic leadership,
self-administration, and communication. The diagram defining a
mental imbalance and hierarchical extreme introvertedness approach
with low and high association based confidence is seemed in Fig. 4.
31
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

Fig. 4 Autism and Organizational Autism Policy

32
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

5. Bidirectional Recurrent Neural Networks

Bidirectional Recurrent Neural Networks (BRNNs) was created in


1997 by Schuster and Paliwal. BRNNs were acquainted with the
interest of increasing the measure of information accessible to the
system [1]. For instance, multilayer layer perceptrons (MLPs) and time
delay neural networks (TDNNs) have restrictions on the information
adaptability, as they require their information to be settled. Standard
recurrent neural networks (RNNs) likewise have confinements as the
future information can't come from the present state. Unexpectedly,
BRNNs don't require their information to be settled. In addition, their
future information is accessible from the present state [3]. The
necessary thought of BRNNs is to subordinate two hidden layers of
inverse headings to an indistinguishable yield as appeared in Fig. 5.
Through this organization, the output layer can acquire data from past
and future states. BRNNs are particularly valuable when the setting of
the info is required.

33
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

Fig. 5 (a) Recurrent Neural Networks


(b) Bidirectional Recurrent Neural Networks

34
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

5.1 Structure of Recurrent Neural Networks

The term "recurrent neural network" is utilized to elude two wide


classes of systems with a comparative general structure, where one is
limited motivation and the other is interminable drive. The two classes
of systems demonstrate transient unique conduct [8]. A limited drive
repetitive system is a coordinated non-cyclic diagram that can be
unrolled and supplanted with an entirely feed forward neural system,
while a vast motivation intermittent system is a coordinated cyclic
chart that can't be unrolled. Fundamental RNNs are a system of
neuron-like nodes, each with a guided one-route association with each
other node. Every node has a period differing genuine esteemed
actuation. Every association like neurotransmitter has a modifiable
genuine esteemed weight. Nodes are either input layers getting
information from outside the arrangement, output layers yielding
outcomes, or hidden layers that alter the information on course from
contribution to yield as outlined in Fig. 6. Recurrent neural network are
like feed-forward systems, aside from that each layer has an
intermittent association with a tap delay related with it. This enables
the system to have a vast unique reaction to time arrangement input
information [15]. This system is like the time delay and distributed
delay neural networks, which have limited information reactions. This
enables it to display dynamic transient conduct for a period succession
and the Matlab execution of fundamental RNNs is exhibited in Fig. 7.
Dissimilar to feed forward neural systems, RNNs can utilize their inner
35

state memory to process arrangements of data sources. RNNs are


Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

famous models that have indicated extraordinary guarantee in


numerous NLP undertakings. In any case, in spite of their on-going
fame just a predetermined number of assets used RNNs because of
their info information adaptability methodology.

Fig. 6 Basic Recurrent Neural Networks

Fig. 7 Matlab Implementation of Basic Recurrent Neural Networks


36
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

5.2 Limitations of Recurrent Neural Networks

The idea behind RNNs is to make utilization of successive data. In a


conventional neural system, it is expected that all information
sources and yields are free of each other. In any case, for some
errands that is not a good idea [18]. RNNs are called repetitive on
the grounds that they play out a similar errand for each component
of an arrangement, with the yield being relied upon the past
calculations. Another approach to consider RNNs is that they have a
memory which catches data about what has been computed up until
now. In principle RNNs can make use of data in discretionarily long
successions, nevertheless, practically speaking they are constrained
to thinking back just a couple of steps. Here are the facts that basic
RNNs linger behind;

 Restrictions on the input data flexibility, as they involve their


input data to be stable.
 The future input information cannot be reached from the current
state.
 Different networks trained on the same data are no longer
regarded as independent.
37
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

5.3 Structure of Bidirectional Recurrent Neural Networks

The principle of BRNNs is to part the neurons of standard RNNs into


two bearings, one for positive time course in forward states, and
another for negative time heading in reverse states. Those two states'
yields are not associated with contributions of the other way states as
shown in Fig. 8. The general structure of RNNs and BRNNs can be
portrayed properly by utilizing two time bearings, input data from the
past and eventual state of the present time allotment can be utilized
not at all like standard RNNs which require the postponements for
including future data [19]. BRNNs can be prepared utilizing
comparable calculations to RNNs, in light of the fact that the two
directional neurons don't have any communications. By the by, when
back-propagation is connected, extra procedures are required on the
grounds that refreshing input and output layers is impossible without a
moment's delay. General methodology for preparing is as per the
following; for forward pass, forward states and in reverse states is
passed to begin with and afterward yield neurons are passed. For in
reverse pass, yield neurons are passed to start with, and after that
forward states and in reverse states are passed directly. After forward
and in reverse passes are done, the weights are refreshed. To conquer
the constraints of general RNNs, a bidirectional recurrent neural
network that can be prepared utilizing all accessible data before and
eventual state of a particular time period is ordinary in grouping
applications. BRNNs can give, in contrast with other RNN's, a quicker
38

progression of genuine applications with better outcomes.


Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

Fig. 8 Structure of Bidirectional Recurrent Neural Networks

39
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

The mathematical representation of the BRNNs structure is given by;

( )

(∑ ∑ )

( )

(∑ ∑ )

( )

(∑ ∑ )
40
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

5.4 Training Bidirectional Recurrent Neural Networks

1. To train forward pass, RUN every input data for one time period
over the BRNNs and define the expected outputs.
(a) Forward pass for forward states (n = 1,.., N) and backward
states (n = N,.., 1).
(b)Forward pass for output.
2. To train backward pass, CALCULATE the functional derivatives of
the time period recycled in the forward pass.
(a)Backward pass for output.
(b) Backward pass for forward states (n = N,.., 1) and backward
states (n = 1,.., N).
3. UPDATE the weights.

41
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

5.5 Advantages of Bidirectional Recurrent Neural Networks

BRNNs have numerous advantages over other common neural


networks in classification applications [26]. Due to its neuron-like
structure, BRNNs provides ease of categorizing neurodegenerative
syndromes and classifying brain abnormality by its inherent ability to
learn from both past and future. Few of BRNNs advantages are listed
below;

 BRNNs prompts quick and stable convergence contrasted with


other neural networks including RNNs.
 Increased computational and memory requests can be mostly
alleviated by BRNNs because of its adaptability and bidirectional
nature.
 BRNNs with synchronous and limited memory properties permit
future conditions with all past and last input data.
 BRNNs accomplish better execution at learning highlights along
the most limited reliance way.
 BRNNs give larger amount of resilience to distortions.
 Most clearly, if the information arrangements are transient and
not spatial there exists a good fortuitous to recognize past and
future contributions as on account of brain disease
classifications.
 The BRNNs can focus all the learning exertion on forward
expectation, and re-utilize the scholarly conditions in reverse
42

surmising.
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

6. Implementation and Results

A dataset of 300 patients of various ageing from newly born infant to


85 years were exemplified. Several brain imaging modalities such as
Computed Tomography, Magnetic Resonance Imaging, and Positron
Emission Tomography were used to perceive the brain disorders.
Initially, 100 patients of each dementia, alzheimer’s disease and autism
spectrum tomography accounted for the examination of bidirectional
recurrent neural networks in classification process. The results
obtained were compared with the recurrent neural networks to
measure the efficiency of bidirectional recurrent neural networks.
Sensitivity and specificity are the two criteria used in evaluating the
accuracy of the classification process. The consequences obtained
were stated in Table 1.

Sensitivity and specificity are defined accordingly and their


mathematical correlations are given by;

Sensitivity also termed as the “True Positive Rate” i.e. total number
of true positive patients with respect to total number of disorder
positive patients. True positive patients are the total number of
patients correctly classified for a definite disorder. Disorder positive
patients are the total number of patients rightly classified as correct
for that particular disorder.



43
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

Specificity also termed as the “True Negative Rate” i.e. total number
of true negative patients with respect to total number of disorder
negative patients. True negative patients are the total number of
patients wrongly classified for a definite disorder. Disorder positive
patients are the total number of patients rightly classified as wrong for
that particular disorder.


Accuracy is measured as the percentage of True Positive Rate


(Sensitivity) with respect to True Negative Rate (Specificity).

44
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

The comparative results demonstrate that the bidirectional


recurrent neural networks showed much better classification of
dementia, alzheimer’s disease and autism spectrum disorder with
respect to recurrent neural networks. It is observed that BRNNs
provided at-least 15% better performance than RNNs in classifying the
brain disorders. The false positive rate and false negative rates are not
included in determining the accuracy of the classification process with
the account that it is more sufficient to measure the classification
accuracy with sensitivity and specificity in medical analysis unlike
engineering applications. BRNNs showed better results due to the
synchronous and limited memory properties that permit them to
access the future conditions with all past and last input data.

45
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

Table 1. Evaluation of Accuracy in Percentage

Sensitivity Specificity
Number of Type of Type of
(True (True Accuracy
S.No. Patients Neural Brain
Positive Negative (Percentage)
Examined Network Disorder
Rate) Rate)

1.20 1.45 82.75%


Dementia

Recurrent
Alzheimer’s 1.40 1.94 72.16%
Neural
1. 300 Disease
Networks
(RRNs)
Autism
1.15 1.47 78.23%
Spectrum
Disorder

1.02 1.07 95.32%


Dementia
Bidirectional
Recurrent Alzheimer’s 1.04 1.16 89.65%
2. 300 Neural Disease
Networks
(BRRNs) Autism
1.03 1.12 91.96%
Spectrum
Disorder
46
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

7. Conclusion

This chapter recommended a new direction in the classification of


brain disorders using bidirectional recurrent neural networks. BRNNs
though proved to be efficient in many engineering applications; its
utility in medical analysis was not attempted. Classifying brain
disorders is very crucial to diagnose and treat the patients efficiently
for their well-being. The three major brain disorders; dementia,
alzheimer’s disease and autism spectrum disorder were accounted in
this proposed strategy. The results provided here are raw facts and it
can be further dealt in depth based on the different categories and
types of brain disorders. This proposed method can be extended in
evaluating the other types of brain disorders as well in performing a
complete analysis of bidirectional recurrent neural networks in
classifying the brain disorders. This chapter provides the first step in
this trend and a complete analysis can be done as a future direction.
Nonetheless, bidirectional recurrent neural networks established 15%
better performance than recurrent neural networks. Until performing
a complete brain disorder classification using bidirectional recurrent
neural networks, discussing the limits of bidirectional recurrent neural
networks in this stratagem will not be appropriate.
47
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

References

[1] Mike Schuster and Kuldip K. Paliwal. (1997) Bidirectional Recurrent


Neural Networks. IEEE Transactions on Signal Processing, Vol. 45, No.
11.

[2] Gilman S. (2010) Oxford American handbook of neurology. Oxford


University Press: Oxford, UK.

[3] Bayer, J. and Osendorfer, C. (2014) Learning stochastic recurrent


networks. arXiv preprint arXiv: 1411.7610.

[4] Rodda J, Carter J. Cholinesterase inhibitors and memantine for


symptomatic treatment of dementia. (2012), 344:e2986.

[5] Suh GH, Shah A. (2001) A review of the epidemiological transition


in dementia--cross-national comparisons of the indices related to
Alzheimer disease and vascular dementia. Acta Psychiatrica
Scandinavica, 104(1), pp.4-11.

[6] Lotter V. Epidemiology of autistic conditions in young children.


(1966) Soc Psychiatry. 1, pp.124-135.

[7] Bahdanau, D., Cho, K., and Bengio, Y. (2015) Neural machine
translation by jointly learning to align and translate. Proceedings of the
International Conference on Learning Representations (ICLR 2015).

[8] J. O. Berger. (1985) Statistical Decision Theory and Bayesian


Analysis. Berlin, Germany: Springer-Verlag.
48
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

[9] Fombonne E. (2006) Past and future perspectives on autism


epidemiology. In: Moldin SO, Rubenstein JLR, editors. Understanding
autism: from basic neuroscience to treatment. Boca Raton: CRC Press;
pp.25-45.

[10] Health, D. O. (2006) A Vision for Change: Report of the expert


group on mental health policy.

[11] P.S.Jagadeesh Kumar, Yang Yung. (2018) Medical Image


Processing and Health Care Services, First Edition 2018, InTech.

[12] Fombonne E. (2009) Epidemiology of pervasive developmental


disorders. Pediatr Res., 65:591-8.

[13] Stevens T et al. (2002) Islington study of dementia subtypes in the


community. British Journal of Psychiatry, 180, pp.270-276.

[14] Newens AJ et al. (1993) Clinically diagnosed presenile dementia


of the Alzheimer type in the Northern Health Region: ascertainment,
prevalence, incidence and survival. Psychological Medicine, 23(3),
pp.631644.

[15] C. M. Bishop. (1995) Neural Networks for Pattern Recognition.


Oxford, United Kingdom, Clarendon.

[16] Elsabbagh M, Divan G, Koh YJ, Kim YS, Kauchali S, Marcin C, et al.
(2012) Global prevalence of autism and other pervasive
developmental disorders. Autism Res., 5:160-79.
49

[17] Alzheimer A. (1987) About a peculiar disease of the cerebral


Page

cortex. Alzheimer Dis Assoc Disord, 1: 38.

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

[18] H. Bourlard and C. Wellekens. (1990) Links between Markov


models and multilayer perceptrons. IEEE Trans. Pattern Anal. Machine
Intell., Vol. 12, pp.1167–1178.

[19] Baldi, P., Brunak, S., Frasconi, P., Soda, G., and Pollastri, G. (1999).
Exploiting the past and the future in protein secondary structure
prediction. Bioinformatics, 15(11), pp.937–946.

[20] Maurer K, Volk S, Gerbaldo H. Auguste D and Alzheimer’s disease.


(1997) Lancet, 349 (9064): 15469.

[21] P.S.Jagadeesh Kumar. (2018) Computer Aided Therapeutic of


Alzheimer’s Disease Eulogizing Pattern Classification and Deep
Learning Protruded on Tree-based Learning Method. Progress in
Advanced Computing and Intelligent Engineering. Advances in
Intelligent Systems and Computing AISC Book Series of Springer,
Volume 564, pp.103-113.

[22] Cahill S,Diaz-Ponce AM, Coen RF,Walsh C. (2010) The


underdetection of cognitive impairment in nursing homes in the
Dublin area. The need for on-going cognitive assessment. Age Ageing.
39(1), pp.128-31.

[23] Rodda J, Walker Z. (2009) Ten years of cholinesterase inhibitors.


Int J Geriatr Psychiatry. 24(5), pp.437-42.

[24] Clarke M et al. (1991) The prevalence of dementia in a total


population: a comparison of two screening instruments. Age and
50

Ageing, 20(6), pp.396403.


Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar


Bi-directional Recurrent Neural Networks in Classifying Dementia, Alzheimer’s Disease and Autism Spectrum Disorder

[25] P.S.Jagadeesh Kumar, Yang Yung, Mingmin Pan, Wenli Hu. (2018)
Classification and Evaluation of Macular Edema, Glaucoma and
Alzheimer’s Disease Using Optical Coherence Tomography. I J of
Biomedical Engineering and Technology, Report, 25(2/3/4), pp.370-
388.

[26] Bastien, F., Lamblin, P., Pascanu, R., Bergstra, J., Goodfellow, I. J.,
Bergeron, A., Bouchard, N., and Bengio, Y. (2012) Theano: new
features and speed improvements. Deep Learning and Unsupervised
Feature Learning, NIPS 2012.

[27] Magaziner J,German P, Zimmerman S, et al. (2000) The


prevalence of dementia in a statewide sample of new nursing home
admissions aged 65 and older: diagnosis by expert panel.
Gerontologist 40, pp.663-672.

[28] Tyas SL et al. (2003) Mid-life smoking and late-life dementia: the
Honolulu-Asia Aging Study. Neurobiology of Aging, 24(4), pp.589-
596.

[29] Iliffe S, Koch T, Jain P, Lefford F,Wong G,Warner A, et al. (2012)


Developing an educational intervention on dementia diagnosis and
management in primary care for the EVIDEM-ED trial. Trails, 13:142.

[30] Molloy W. (2011) Let Me Decide. Irish Edition Ed: Newgrange


Press Ltd.
51
Page

The Art of Fixing Alzheimer’s Disease P.S.Jagadeesh Kumar

Anda mungkin juga menyukai