Early Detection of Parkinson's Disease: An Intelligent Diagnostic Approach

Early Detection of Parkinson's Disease: An Intelligent Diagnostic Approach

Debashree Devi, Saroj K. Biswas, Biswajit Purkayastha
DOI: 10.4018/978-1-5225-8567-1.ch005
OnDemand:
(Individual Chapters)
Available
$37.50
No Current Special Offers
TOTAL SAVINGS: $37.50

Abstract

Parkinson's disease (PD) is a neurodegenerative disorder that occurs due to corrosion of the substantia nigra, located in the thalamic region of the human brain, and is responsible for transmission of neural signals throughout the human body by means of a brain chemical, termed as “dopamine.” Diagnosis of PD is difficult, as it is often affected by the characteristics of the medical data of the patients, which include presence of various indicators, imbalance cases of patients' data records, similar cases of healthy/affected persons, etc. Through this chapter, an intelligent diagnostic system is proposed by integrating one-class SVM, extreme learning machine, and data preprocessing technique. The proposed diagnostic model is validated with six existing techniques and four learning models. The experimental results prove the combination of proposed method with ELM learning model to be highly effective in case of early detection of Parkinson's disease, even in presence of underlying data issues.
Chapter Preview
Top

Introduction

Parkinson’s Disease (PD) is a progressive, neurodegenerative disorder that basically effects the motor system of a human being which causes physical problems such as shaking, stiffness, and difficulty in walking, balancing and coordinating movements (Parkinson’s disease) (Mohamed, 2016). PD is basically central nervous system oriented disease, occurred due to disintegration of a region called “substantia nigra” in the thalamic region of human brain (Figure 1).

Figure 1.

Location of Substantia nigra in human brain (Parkinson’s disease)

978-1-5225-8567-1.ch005.f01

This substantia nigra secretes a neurochemical named “dopamine” which is responsible for transmitting of neural signals to the different organs and parts of the human body (Figure 2).

Figure 2.

Comparison of neurons of normal and PD-affected persons (Parkinson’s disease)

978-1-5225-8567-1.ch005.f02

With increase of dopamine loss in the course of time, the PD progresses gradually which ultimately leads to mental disorder such as thinking and behavioral disability, dementia, depression and anxiety, sleep disorder, lack of emotions etc. (Olanrewaju, Sahari, Musa, Hakiem, 2014).The various movement-difficulties experienced by PD-suffered individuals are collectively addressed by the term “Parkinsonism”. It has been revealed in a recent survey in 2015 that approximately there are 7-10 million people are suffering from PD in the worldwide. The occurrence of PD is much higher in case of elderly people, typically over the age of 60 years; with a male-to-female ratio of 3:2 (Parkinson’s disease) (Mohamed, 2016).

Parkinsonism is defined by “Bradykinesia”, a spectrum of movement disorder, commonly known as hypokinesia. Bradykinesia refers to a situation when a person’s movement or initiation of voluntary body-movement has slowed down, due to disruption in basal ganglia of human brain, with an increasing reduction of speed and range of repetitive actions (Parkinson’s disease). Bradykinesia is considered to be one of the four key symptoms of PD, along with rigidity, tremor, and postural instability (Pahwa and Lyons, 2013).

There is no remedy or prevention available for PD. However, PD can be controlled if diagnosed at an early stage. Basically, in clinical diagnosis, PD is diagnosed with the presence of two or three of the four key motor-sign symptoms, namely tremor, rigidity, Bradykinesia and postural disability. The risk factor of PD tends to increase with the age of an individual. Most of the motor-sign symptoms are remained unnoticed as they are similar to the aging-sign of an elderly individual. By early detection of PD, its effect can be minimized on the affected person and proper treatment can be offered. One of the advanced and recent method of early detection of PD is using of biomarkers. Biomarker namely DaTscan, which is approved by FDA, to be used in diagnosis test of PD. The DaTscan (Ioflupane 123I injection, also known as phenyltropane) is a radiopharmaceutical agent which is injected into a patient’s veins in a procedure referred to as SPECT imaging (DAT-scanning) (DaTSCAN, 2011). It is used to detect the loss of nerve cells in the striatum region of human brain, specifically the cells that release dopamine. However, the use of DaTscan also carries risk of allergic reaction, especially in case of pregnant women (DaTSCAN, 2011). Hence, there is a demand for development of intelligent data mining techniques which can provide efficient and early detection and diagnosis of PD, without risking the side-effects of treatments. A number of research works have been proposed in recent years based on this idea (Olanrewaju, Sahari, Musa, Hakiem, 2014) (Alemami and Almazaydeh, 2014) (Mohamed, 2016) (Caliskan, Badem, Baştürk, Yüksel, 2017) (Anita and Aruna Priya, 2016).

Key Terms in this Chapter

Tomek-Link: Tomek-link defines the concept boundary pairs in a data distribution. When two instances, a and b are nearest of each-other, with class (a) class (b ); then the pair (a,b) is termed as Tomek-link pair. These pairs are noise-promoting in the data distribution.

Data Imbalance: Data imbalance defines a classification problem where the distribution of data is not even and leads to yield biased results with higher misclassification rate of the minority class data.

Overlapping Cases: The cases which have equal probability of belonging to the proximity of two regions are termed as overlapping cases. It is very difficult to distinguish the exact class of these cases.

Bradykinesia: Bradykinesia is a symptom of human motor-system that describes slow/difficulty moving of body parts (limbs/legs). It is one of the basic symptoms of Parkinsonism.

Dopamine: Dopamine is defined as a neurotransmitter, which is responsible for transmitting neural signals from the brain to the various body parts and organs.

Extreme Learning Machine: Extreme learning machine is a form of regularized feed-forward neural network without the need of hidden-layer parameter tuning. ELM is popular due to its high regularization ability and low execution time.

Central Nervous System: The central nervous system consists of the brain and spinal cord of the human body and is responsible for incorporating and coordinating activities among all the parts of human body.

One-Class SVM: One-class SVM is a classification model which provides to detect the novel instances, within a region of definite range. It is also known as SVM, as its performance is influenced by the parameter, .

Complete Chapter List

Search this Book:
Reset