Neuromuscular Disease Classification

Neuromuscular Disease Classification by using
Wavelet Decomposition Technique
Shravanti Kalwa
Department of Instrumentation and Control
Cummins College of Engineering for Women
Pune, India
Email: [e-mail]
Prof.H.T.Patil
Department of Instrumentation and Control
Cummins College of Engineering for Women
Pune, India
Email: [e-mail]
Abstract—Electromyograph (EMG) is a recording of electrical activity of the skeletal muscles. Various muscle related and neuromuscular diseases are diagnosed by analysing EMG signals.
In this work, neuromuscular diseases such as amyotrophic lateral sclerosis (ALS), myopathy and normal subjects are analysed by using discrete wavelet transform (DWT). In time domain analysis
root

To demonstrate classification, an EMG data consisting of 5 ALS, 5 myopathy and 5 normal subjects are considered. Keywords- Amyotrophic lateral sclerosis (ALS), discrete wavelet transform(DWT), electromyograph (EMG), feature extraction, myopathy.
I. I NTRODUCTION
EMG signals are acquired either by inserting a needle electrode deep inside the skin or by placing an electrode on the skin surface. Needle electrodes are mostly preferred to analyse deep seated muscles i.e, to measure single muscle fibre activity.
Surface electrodes are used to measure superficial muscle fibre activity. EMG signals are used to determine muscle contraction. The structural and functional unit of the muscle is known as motor unit. Each motor unit consist of single alpha motor neuron and muscle fibres. EMG signal analysis plays an important role in diagnosis of various disorders such as identification of neuromuscular diseases, assessing low-back pain and motor control related disorders.
This work focuses on identification of neuromuscular dis-eases such as amyotrophic lateral sclerosis and myopathy [1].
Amyotrophic lateral sclerosis which is a neurodegenerative disease caused by rapidly progressive weakness due to

DWT decomposes a signal into an approximate and detail coefficients. This is achieved by passing a signal through a low pass filter and a high pass filter. To analyse high frequency content, high pass filter is used and a low pass filter is used to analyse low frequency content in a signal. Signal resolution is changed by filtering operation. And upsampling and downsampling operations are performed on a signal to change a scale.
The original signal x(n) is filtered by passing a signal through high pass filter h(n) and low pass filter g(n) to produce an output of the first level decomposition, which is expressed as, y high
(k) =  n x(n):h(2k n) (4) y low (k) =  n x(n):g (2k n) (5)
Here y high (k) and y low (k) are obtained after performing a downsampling by 2 operation. DWT coefficient y low (k) are termed as approximate coefficients and y high (k) are termed as detailed coefficients. In this work approximate coefficients are taken into consideration because energy of the signal is present in low frequency region.
D. Dimensionality Reduction Method
In frame by frame wavelet decomposition method