Cyclostationary Modeling of Surface Electromyography Signal During Gait Cycles and Its Application for Cerebral Palsy Diagnosis

doi:10.1007/s12204-018-2023-9

摘要/Abstract

摘要： Cerebral palsy (CP) is a group of permanent movement disorders that appear in early childhood. The electromyography (EMG) signal analysis and the gait analysis are two most commonly used methods in the clinic. In this paper, a cyclostationary model of the EMG signal is proposed. The model can combine the aforementioned two methods. The EMG signal acquired during the gait cycles is assumed to be cyclostationary due to the physiological characteristics of the EMG signal production. Then, the spectral correlation density is used to analyze the cyclic frequency (corresponding to the gait cycles) and spectral frequency (the frequency of EMG signal) in a waterfall representation of the two kinds of frequencies. The experiments show that the asymptomatic (normal) subjects and symptomatic subjects (with CP) can be distinguished from the spectral correlation density in a range of cyclic frequencies.

关键词: electromyography (EMG) signal, gait cycle , spectral correlation density, cyclostationary modeling , cerebral palsy (CP) diagnosis

Abstract: Cerebral palsy (CP) is a group of permanent movement disorders that appear in early childhood. The electromyography (EMG) signal analysis and the gait analysis are two most commonly used methods in the clinic. In this paper, a cyclostationary model of the EMG signal is proposed. The model can combine the aforementioned two methods. The EMG signal acquired during the gait cycles is assumed to be cyclostationary due to the physiological characteristics of the EMG signal production. Then, the spectral correlation density is used to analyze the cyclic frequency (corresponding to the gait cycles) and spectral frequency (the frequency of EMG signal) in a waterfall representation of the two kinds of frequencies. The experiments show that the asymptomatic (normal) subjects and symptomatic subjects (with CP) can be distinguished from the spectral correlation density in a range of cyclic frequencies.

Key words: electromyography (EMG) signal, gait cycle , spectral correlation density, cyclostationary modeling , cerebral palsy (CP) diagnosis

中图分类号:

R 318

YU Liang (余亮), YAN Li (严莉), CHEN Mengjie (陈梦婕), DONG Liangchao (董良超). Cyclostationary Modeling of Surface Electromyography Signal During Gait Cycles and Its Application for Cerebral Palsy Diagnosis[J]. Journal of Shanghai Jiao Tong University (Science), 2018, 23(Sup. 1): 56-61.

参考文献 24

[1]	RICHARDS C L, MALOUIN F. Cerebral palsy: Definition,assessment and rehabilitation [J]. Handbook ofClinical Neurology, 2013, 111: 183-195.
[2]	BINDER H, ENG G D. Rehabilitation managementof children with spastic diplegic cerebral palsy[J]. Archives of Physical Medicine and RehabilitationHome, 1989, 70(6): 482-489.
[3]	BELL K J, ?OUNPUU S, DELUCA P A, et al. Naturalprogression of gait in children with cerebral palsy [J].Journal of Pediatric Orthopaedics, 2002, 22(5): 677-682.
[4]	ROMEO D M M, GUZZETTA A, SCOTO M, et al.Early neurologic assessment in preterm-infants: Integrationof traditional neurologic examination and observationof general movements [J]. European Journalof Paediatric Neurology, 2008, 12: 183-189.
[5]	TREMBLAY F, MALOUIN F, RICHARDS C L, etal. Effects of prolonged muscle stretch on reflex andvoluntary muscle activations in children with spasticcerebral palsy [J]. Scandinavian Journal of RehabilitationMedicine, 1990, 22(4): 171-80.
[6]	DAMIANO D L, KELLY L E, VAUGHN C L. Effectsof quadriceps femoris strengthening on crouch gait inchildren with spastic diplegia [J]. Physical Therapy,1995, 75(8): 658-667.
[7]	MURARY M P. Gait as a total pattern of movement[J]. American Journal of Physical Medicine, 1967,46(1): 290-332.
[8]	DELUCA P A, DAVIS R B, OUNPUU S, et al. Alternationsin surgical decision making in pations withcerebral palsy based on three-dimensional gait analysis[J]. Journal of Pediatric Orthopaedics, 1997, 17(5):608-614.
[9]	KAY R M, DENNIS S, RETHLEFSEN S, et al. Theeffect of preoperative gait analysis on orthopaedic decisionmaking [J]. Clinical Orthopaedics and RelatedResearch, 2000, 372: 217-222.
[10]	GRAUBERT C, SONG K M, MCLAUGHLIN J F, etal. Changes in gait at 1 year post-selective dorsal rhizotomy:Results of a prospective randomized study [J].Journal of Pediatric Orthopaedics, 2000, 20(4): 496-500.
[11]	ZUMSTEG Z S, AHMED R E, SANTHANAM G, etal. Power feasibility of implantable digital spike sortingcircuits for neural prosthetic systems [J]. IEEE Transactionson Neural Systems and Rehabilitation Engineering,2005, 13(3): 237-240.
[12]	WOLD S, ESBENSEN K, GELADI P. Principal componentanalysis [J]. Chemometrics and Intelligent LaboratorySystems, 1987, 2(1): 37-52.
[13]	WANG W, STEFANO A D, ALLEN R. A simulationmodel of the surface EMG signal for analysis of muscleactivity during the gait cycle [J]. Computers in Biologyand Medicine, 2006, 36(6): 601-618.
[14]	DUCH?ENE J, HOGREL J Y. A model of EMG generation[J]. IEEE Transactions on Biomedical Engineering,2000, 47(2): 192-201.
[15]	BROWN B H, SMALLWOOD R H, BARBER D C,et al. Medical physics and biomedical engineering [M].New York, USA: Taylor & Francis Group, 1999.
[16]	DIMITROV G V, DIMITROVA N A. Precise and fastcalculation of the motor unit potentials detected bya point and rectangular plate electrode [J]. MedicalEngineering & Physics, 1998, 20(5): 374-381.
[17]	WILSON F N, MACLEOD A G, BARKER P S. Thedistribution of the action currents produced by heartmuscle and other excitable tissues immersed in extensiveconducting media [J]. Journal of General Physiology,1933, 16(3): 423-456.
[18]	SHEMMELL J, JOHANSSON J, PORTRA V, et al.Control of interjoint coordination during the swingphase of normal gait at different speeds [J]. Journalof Neuro Engineering and Rehabilitation, 2007, 4(1):10-10.
[19]	ANTONI J, RANDALL R B. A stochastic model forsimulation and diagnostics of rolling element bearingswith localized faults [J]. Journal of Vibration andAcoustics, 2003, 125: 282-289.
[20]	HO D, RANDALL R B. Optimization of bearing diagnosticstechniques using simulated and actual bearingfault signals [J]. Mechanical Systems and Signal Processing,2000, 14(5): 763-788.
[21]	ANTONI J, RANDALL R B. Differential diagnosis ofgear and bearing faults [J]. Journal of Vibration andAcoustics, 2002, 124: 165-171.
[22]	FEHSKE A, GAEDDERT J, REED J H. A new approachto signal classification using spectral correlationand neural networks [C]//IEEE International Symposiumon New Frontiers in Dynamic Spectrum AccessNetworks. [s.l.]: IEEE, 2005: 144-150.
[23]	ANTONI J. Cyclostationarity by examples [J]. MechanicalSystems and Signal Processing, 2009, 23: 987-1036.
[24]	ANTONI J, XIN G, HAMZAOUI N. Fast computationof the spectral correlation [J]. Mechanical Systems andSignal Processing, 2017, 92: 248-277.