Docstoc

EMG DIAGNOSIS VIA TIME DOMAIN FEATURES AND BINARY

Document Sample
EMG DIAGNOSIS VIA TIME DOMAIN FEATURES AND BINARY Powered By Docstoc
					                           Gurmanik Kaur et. al. / International Journal of Engineering Science and Technology
                                                                                    Vol. 2(10), 2010, 5192-5196


 EMG DIAGNOSIS VIA TIME DOMAIN
  FEATURES AND BINARY SUPPORT
 VECTOR MACHINE CLASSIFICATION
                                            GURMANIK KAUR†
                                     Research Scholar, EIE Department,
                                 SLIET (Deemed-to-be-University), Longowal
                                    Distt. Sangrur, Punjab, INDIA, 148106
                                       Email-id: mannsliet@gmail.com

                                       DR. AJAT SHATRU ARORA
                                           Professor, EIE Department,
                                 SLIET (Deemed-to-be-University), Longowal
                                    Distt. Sangrur, Punjab, INDIA, 148106
                                        Email-id: ajatsliet@yahoo.com

                                      DR. VIJENDER KUMAR JAIN
                                     Professor & Head, EIE Department,
                                 SLIET (Deemed-to-be-University), Longowal
                                    Distt. Sangrur, Punjab, INDIA, 148106
                                       Email-id: vkjain27@yahoo.com

                                        DR. GURTEJ SINGH SIDHU
                                                RMO & EMO
                                  Silver Oaks Hospital, Phase IX, Sector-63,
                                 SAS Nagar (Mohali), Punjab, INDIA, 160062
                                     Email-id: dr.gurtejsidhu@yahoo.com

Abstract :
There are more than 100 neuromuscular disorders that affect the brain, spinal cord, nerves and muscles. Many of
these diseases are hereditary and life expectancy of many sufferers is considerably reduced. Early detection and
diagnosis of these diseases by clinical examination and laboratory tests is essential for their management as well
as their prevention through prenatal diagnosis and genetic counselling. Such information is also useful in
research which may lead to the understanding of the nature and eventual treatment of these diseases. Laboratory
investigations include neurophysiological tests, nerve and muscle biopsies, biochemical analysis and more
recently deoxyribo nucleic acid (DNA) analysis for the localization and identification of genes.
Electromyographic examination studies the electrical activity of the muscle and forms a valuable
neurophysiological test for the assessment of neuromuscular disorders.

The shapes and firing rates of the motor unit action potentials (MUAPs) in an electromyogram (EMG) signal
provide an important source of information for the diagnosis of neuromuscular disorders. In order to extract this
information from the EMG signals recorded at low to moderate force levels, it is required to identify and
classify the MUAPs composing the EMG signal. The identification of possible MUAPs is done by segmentation
of EMG signal using threshold technique. The identified MUAPs are clustered using a statistical pattern
recognition technique. After extraction of time domain features, MUAPs are classified using binary support
vector machine (SVM) classifier. A total of 12 EMG signals obtained from 3 normal (NOR), 5 myopathic
(MYO) and 4 motor neuron diseased (MND) subjects were analyzed. The classification accuracy of multi-class
SVM with time domain features is 75.06 %. In conclusion, the methodology described in this work make
possible the development of a fully automatic EMG signal analysis which is accurate, simple, fast and reliable
enough to be used in routine clinical environment.
Keywords: Electromyography, motor unit action potentials, support vector machine.




ISSN: 0975-5462                                                                                             5192
                            Gurmanik Kaur et. al. / International Journal of Engineering Science and Technology
                                                                                     Vol. 2(10), 2010, 5192-5196

1. Introduction
The motor unit is the smallest functional unit of the muscle. At slight voluntary muscle contraction the MUAP is
recorded to reflect the electrical activity of a single anatomical motor unit. The MUAP findings are used to
detect and describe different neuromuscular diseases [1]. As the contraction intensity increases, more motor
units are recruited. Different MUAPs will overlap; causing an interference pattern in which the neurologists
cannot detect individual MUAP shapes reliably.
    Traditionally, neurophysiologists assess MUAPs from their shape using an oscilloscope and listening to their
audio characteristics. On this way, an experienced neurophysiologist can detect abnormalities with reasonable
accuracy. Subjective MUAP assessment, although satisfactory for the detection of unequivocal abnormalities,
may not be sufficient to delineate less obvious deviations or mixed patterns of abnormalities [2].These
ambiguous cases call for quantitative MUAP analysis.
    With the development of quantitative EMG techniques, some automated decision making systems of
neuromuscular disorder diagnosis emerged. Coatrieux and associates applied cluster analysis techniques for the
automatic diagnosis of pathology based on MUAP records [3]-[5]. Andreassen and co-workers developed the
MUNIN (Muscle and Nerve Inference Network) which employs a causal probabilistic network for the
interpretation of EMG findings [6]-[8]. Fuglsang-Frederiksen and his group developed a rule-based EMG expert
system named KANDID [9], [10] and Jamieson developed an EMG processing system based on augmented
transition networks [11], [12]. In most of these systems, the generation of the input pattern assumes a
probabilistic model, with the matching score representing the likelihood that the input pattern was generated
from the underlying class [13]. In addition, assumptions are typically made concerning the probability density
function of the input data. Pattichis et al gave a series research yield of classifying MUAPs for differentiation of
motor neuron diseases and myopathies from normal [14]. The classifier they used were mainly neural networks,
e.g. back propagation, the radial basis function and the self organizing feature map network. However, the
aforementioned techniques used to train the neural network classifiers are based on the idea of minimizing the
train error, which is named empirical risk. As a result, limited amounts of training data and over high training
accuracy often lead to over training instead of good classification performance.
    SVMs introduced by Vapnik [15] is founded in the framework of the statistical learning theory, which is
appropriate for approaching classification and regression problems. SVMs represent a new approach to pattern
classification that has attracted a great deal of interest in the machine learning community. They operate on the
induction principle of structural risk minimization, which minimizes an upper bound on the generalization error.
SVMs have shown to be successful in solving many pattern recognition problems [16] and perform much better
than non-linear classifiers such as artificial neural networks in many situations [17].
    In this preliminary study, we investigated the binary SVM classifier for classification of MUAPs recorded
from the biceps brochii muscle. The experimental results proved that the SVM can effectively identify motor
neuron diseased and myopathic subjects from normal.
    This paper is organized as follows: Section 2 presented the material and methodology used for classification
of MUAPs. The experimental results are illustrated in section 3 and section 4 covered the conclusion.



2. Material and Methodology

2.1. Data acquisition and pre-processing
Our data contain real time EMG signal obtained from the Department of Computer Science, University of
Cyprus, Cyprus. All the EMG signals were acquired from the biceps brochii muscle at upto 30% of the
maximum voluntary contraction (MVC) level under isometric conditions. The signals were acquired for 5
seconds, using the standard concentric needle electrode, from NOR, MYO and MND subjects. The EMG signals
were analogue band pass filtered at 3-10 KHz, sampled at 20 KHz with 12-bit resolution and then low pass
filtered at 8 KHz. A typical EMG recording is shown in Fig. 1.




ISSN: 0975-5462                                                                                               5193
                                                    Gurmanik Kaur et. al. / International Journal of Engineering Science and Technology
                                                                                                             Vol. 2(10), 2010, 5192-5196


                                                    300




                                                    200




                                      M gn de )
                                       a itu (µV
                                                    100




                                                     0




                                                   -100




                                                   -200


                                                          0       1000      2000           3000           4000         5000
                                                                                    Sample No.




                                                                         Fig. 1. Raw EMG signal.

2.2. MUAP identification
The EMG signal is segmented into intervals of possible MUAPs. The segmentation algorithm used is data
driven. First, a threshold depending on the maximum value max i xi  and the mean absolute value
       L
1 L  xi    of the whole EMG signal is calculated, where                                        xi the discrete input values and L is the number
      i 1
of samples in the EMG signal. Peaks over the calculated threshold are considered as candidate MUAP’s. Then a
window of 120 sampling points (i.e., 6 ms at 20 kHz) is centered at the identified peak. If a greater peak is found
in the window, the window is centered at the greater peak; otherwise the 120 points are saved as MUAP
waveform [18].
    The segmented EMG signal centered at the maximum peak is shown in Fig. 2.
                                                   400



                                                   300



                                                   200
                         a n d (µ )
                        M g itu e V




                                                   100



                                                      0



                                                   -100



                                                   -200
                                                              0   20       40         60             80          100          120
                                                                                   Sample No.


                     Fig. 2. Segmented EMG signal in segments of 6ms and centered at the maximum peak.


2.3. MUAP clustering
In this step, the MUAP clusters are automatically detected and for each cluster the average or template shape is
determined. We have used statistical pattern recognition technique for clustering of similar MUAPs. In this
technique the euclidian distance is used to identify and group similar MUAP waveforms. The group average is
continuously calculated and is used for the classification of MUAPs using a constant threshold [19]. The steps
are:
Step1: Start with the first waveform x as input (the first member of the class).
Step2: Calculate the vector length of x ( l x ) and the distance between it and the other segmented waveforms y.
Step3: Find the waveform y with the minimum distance d min . The waveform y having minimum distance with
the x has the greatest similarity with x and remove it from the input data.
Step4: If d min / l x  0.3 then group, calculate group average and go to step 1 with group average as input.
else if number of group members > 2, then form a new class.
else waveform is superimposed, go to step 1 with y as input.
    This process continues where it stopped comparing the last encountered waveform with all the remaining until all
waveforms are processed. The threshold values were chosen heuristically after extensive testing. It is noted that again
there are no widely applicable threshold criteria for assigning a MUAP to a class. The threshold used in this work is




ISSN: 0975-5462                                                                                                                             5194
                                       Gurmanik Kaur et. al. / International Journal of Engineering Science and Technology
                                                                                                Vol. 2(10), 2010, 5192-5196

critical because a smaller value may split a MUAP class with high waveform variability in two or more subclasses,
whereas a greater threshold value may merge resembling MUAP classes. The averaged class waveforms are again the
unique MUAP waveforms composing the EMG signal. Fig. 3 illustrates clustered EMG signal.
                                                                                                          140
                   150                                         300

                                                                                                          120
                                                               250
                   100                                                                                    100
                                                               200

                                                                                                           80
                                                               150
                   50
                                                                                                           60
                                                               100

                    0                                                                                      40
                                                                50


                                                                 0                                         20
                   -50
                                                                -50                                        0


                  -100                                         -100                                       -20
                         0   20   40    60   80   100   120           0   20   40   60   80   100   120         0   20   40   60   80   100    120




                                                  Fig. 3. Clustered EMG signal of a NOR subject.


2.4. Computation of MUAP features
The next stage is the computation of MUAP features. To perform this computation an MUAP is expanded to
25ms on the original signal where the position of identified peak was marked during segmentation. The rationale
is that the MUAP duration is in most of the cases longer than 6ms and the signal expansion is therefore,
necessary for a correct measurement of parameters [20]. The following time domain parameters are computed
from the MUAP waveforms:
Spike duration: measured from the first to the last positive peak.
Amplitude: Amplitude difference between minimum positive and maximum negative peak.
Area: Rectified MUAPs integrated over the calculated duration
Phases: Number of baseline crossings where amplitude exceeds ±25µV, plus one.
Turns: Number of positive and negative peaks where the difference from the preceding and following turn
exceeds 25µV.

2.5. MUAP classification
In order to classify the clustered MUAPs into NOR, MYO and MND classes, a binary SVM classifier is
employed [21], [22]. A classification task based on SVM usually involves training and testing data, which
consist of a number of data instances. Each instance in the training set contains one ‘‘target value’’ (class labels)
and several ‘‘attributes’’.
   In this work we have used binary SVM classifier first, to classify the normal and diseased subjects. If the
signal is diseased, then another binary classifier is used to classify MYO and MND signals.


3. Results
EMG data collected from 12 subjects were analyzed using the methodology described in Section II. Data were
recorded from 3 NOR, 5 MYO and 4 MND subjects. Only subjects with no history or signs of neuromuscular
disorders were considered as normal. MATLAB was used for implementing the algorithms. The means and
standard deviations of the aforementioned time domain features of each subject are computed as the input
feature vector of support vector machine. The classification accuracy of binary SVM classifier is 75.06%.


4. Conclusion
An integrated binary classifier based on SVM is adopted in clinical electromyography for differentiating
neuromuscular disorders. The objective of SVM is to find optimal hyperplane for separating MUAP clusters.
Experimental results show that the binary SVM classifier can be effectively trained for MUAPs diagnosis. Still
the diagnostic results could be further investigated in future works with larger data set and other feature sets.

References
[1]   Stalberg, E.; Nandedkar, S. D.; Sanders, D. B.; et al. (1996): Quantitative Motor unit potential analysis, Clin. Neurophysiol, 13(5), pp.
      401-422.
[2]   Richfield, E. K.; Cohen, B. A.; Albers, J. W. (1981): Review of quantitative and automated needle electromyographic analyses, IEEE
      Trans. Biomed. Engg., pp. 506-514.
[3]   Coatrieux, J. L.; Toulouse, P.; Rouvrais, B.; Le Bars, R. (1983): Automatic classification of electromyographic signals, EEG Clin.
      Neurophysiol., 55, pp. 333-341.




ISSN: 0975-5462                                                                                                                               5195
                                   Gurmanik Kaur et. al. / International Journal of Engineering Science and Technology
                                                                                            Vol. 2(10), 2010, 5192-5196

[4]    Rouvrais, B.; Toulouse, P.; Coatrieux, J. L.; Le Bars, R. (1983): A possible method of automatic electromyographic analysis and
       diagnosis on line, Electmyogr. Clin. Neurophysiol, 23, pp. 457-470.
[5]    Toulouse, P.; Coatrieux, J. L.; Le Marec, B. (1985): An attempt to differentiate female relatives of Duchenne type dystrophy from
       healthy subjects using an automatic EMG analysis, J. Neurolog. Sci., 67, pp. 45-55.
[6]    Andersen, S. K.; Andreassen, S.; Woldbye, M. (1986): Knowledge representation for diagnosis and test planning in the domain of
       EMG, Proc. 7th Eur. Conf. on Artificial Intell., Brighton, U.K., pp. 357-368.
[7]    Andreassen, S.; Andersen, S. K.; Jensen, F. V.; Woldbye, M.; Rosenfalck, A.; Falck B.; Kjaerluff, U.; Sorensen, A. R. (1987):
       MUNIN- An expert system for EMG, Electroenceph. Clin. Neurophvsiol., 66.
[8]    Andreassen, S.; Jensen, F.; Andersen, S. K.; Falck, B.; Kjaerulff, U,; Woldbye M.; Sorensen, A. R.; Rosenfalck, A.; Frank, J. (1989):
       MUNIN- An expert EMG assistant, In Computer Aided Electromygraphy and Expert Systems, J. E. Desmedt, Ed. New York,
       Elsevier Science Publisher B.V., pp. 255-277.
[9]    FugIsang-Frederiksen, A.; Jeppesen, S. M. (1989): A rule-based EMG expert system for diagnosing neuromuscular disorders, In
       Computer Aided Electromyography and Expert Systems, J. E. Desmedt, Ed. New York, Elsevier Science Publishers B.V., pp. 289-
       296.
[10]   Fuglsang-Frederiksen, A.; Ronager, J,; Vingtoft, S. (1990): A plan-test diagnose expert system for EMG: KANDID, J. Neurolog. Sci.,
       98 pp. 150.
[11]   Jamieson, P. W. (1990): Computerized interpretation of electromyographic data, Electroencephalogr. Clin. Neurophysiol., 75, pp. 392.
[12]   Jamieson, P. W. (1991): A model for diagnosing and explaining multiple disorders, Comput. Biomed. Res., 24, pp. 307-320.
[13]   Lippmann, R. P. (1987): An introduction to computing with neural nets, IEEE ASSP Mag.
[14]   Contantinos Pattichis S.; Andread Elia C. (1999): Autoregressive and ceptral analyses of motor unit action potentials, Med. Engg.
       Physic., 21, pp. 405-419.
[15]   Vapnik, V. P. (1995): The nature of statistical learning theory, Springer,Verlag, New York.
[16]   Gurmanik, K.; Dr. Ajat Shatru, A.; Dr. Vijender Kumar, J. (2009) Multi-Class Support Vector Machine Classifier in EMG Diagnosis,
       WSEAS Transactions on Signal Processing, 5(12), pp. 379-389.
[17]   Yu-Dong, C.; Xiao-Jun, L.; Xue-Biao, X.; et. Al. (2002): Prediction of protein structural classes by support vector machines, Comput.
       Chemi, 26, pp. 293-296.
[18]   Gurmanik, K.; Dr. Ajat Shatru, A.; Dr. Vijender Kumar J. (2009): Comparison of the techniques used for Segmentation of EMG
       Signals, Proc. 11th WSEAS Inter. Conf. on Mathematical and Computational Methods in Science and Engineering, Baltimore, USA,
       pp. 124-129.
[19]   Gurmanik, K.; Vijender Kumar, J.; Ajat Shatru, A. (2010): Automatic Discovery of the MUAP Clusters in EMG Signals, Proc.
       National Conference on Recent Advances in Computational Techniques in Electrical Engineering, SLIET, Longowal, Sangrur, Punjab,
       INDIA.
[20]   Katsis, C. D.; Fotiadis, D. I.; Likas, A.; Sarmas, I. (2003): Automatic discovery of the number of MUAP clusters and superimposed
       MUAP decomposition in electromyograms, Proc. 4th Annual IEEE Conf. on Information Technology Applications in Biomedicine,
       UK, pp. 177-180.
[21]   Boser, B.; Guyon, I.; Vapnik,V. (1992): A training algorithm for optimal margin classifiers, Proc. 5th Annual Workshop on
       Computational Learning Theory, ACM Press.
[22]   Cortes, C.; Vapnik, V. (1995): Support-vector network, Machine Learn, 32, pp. 273-297.




ISSN: 0975-5462                                                                                                                      5196

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:1
posted:1/7/2012
language:
pages:5
jianghongl jianghongl http://
About