Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Wheelchair vibration, whole body vibration and spasticity : a study of the influence of wheel design… Messenberg, Allon 2010

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
24-ubc_2010_spring_messenberg_allon.pdf [ 35.83MB ]
Metadata
JSON: 24-1.0070931.json
JSON-LD: 24-1.0070931-ld.json
RDF/XML (Pretty): 24-1.0070931-rdf.xml
RDF/JSON: 24-1.0070931-rdf.json
Turtle: 24-1.0070931-turtle.txt
N-Triples: 24-1.0070931-rdf-ntriples.txt
Original Record: 24-1.0070931-source.json
Full Text
24-1.0070931-fulltext.txt
Citation
24-1.0070931.ris

Full Text

Wheelchair Vibration, Whole BodyVibration and SpasticityA study of the influence of wheel design on wheelchairvibration and whole body vibration as a trigger ofmuscle spasms in populations with spinal cord injurybyAllon MessenbergBSc, Tel-Aviv University, 2002A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF APPLIED SCIENCEinThe Faculty of Graduate Studies(Mechanical Engineering)THE UNIVERSITY OF BRITISH COLUMBIA(Vancouver)March 2010c© Allon Messenberg 2010AbstractINTRODUCTION: The majority of individuals with chronic spinal cordinjury (SCI) experience spasticity, which can often impair function and de-grade quality of life. Reports by individuals with SCI suggest that wholebody vibration (WBV), as can occur while riding wheelchairs, may triggerspasticity.OBJECTIVES: 1. Examine the influence of wheel design on wheelchairvibration. 2. Develop a system allowing exposure of individuals with SCIto WBV and analysis of muscle activity to identify spasticity.METHODS: 1. A wheelchair wheel comparison study: Vibration ac-celeration and frequency content produced by wheelchairs equipped with 2different wheel designs (steel spoked and composite material spoked) werecompared as:1a. 13 subjects with SCI wheeled through an obstacle course simulating awheelchair user’s daily activities1b. 22 non-SCI subjects wheeled down a ramp and over a vibration inducingobstacle.Vibration acceleration was recorded using 2 accelerometers mounted on thewheelchairs’ main axle and footrest. The influence of wheelchair vibration onspasticity was assessed using questionnaires, completed by the SCI subjects.2. A controlled whole body vibration (CWBV) pilot study: 2 SCI subjectswere exposed to 10 WBV sessions. Each exposure consisted of a singlefrequency, lasted 20 seconds, and was repeated on 2 separate days. TheWBV was applied using an electrodynamic shaker and the subjects’ legmuscles’ activity was recorded using an electromyography (EMG) system.Muscle spasms were identified by calculating the ratio between periods ofincreased muscle activity and the period before exposure to vibration.RESULTS: No statistically significant differences (p=0.05) were found inwheelchair vibration acceleration or frequency content between the 2 testediiAbstractwheel designs and no clear correlation between wheelchair vibration andspasticity was apparent. The CWBV system was able to apply vibration(±0.5Hz,±0.001g) and record muscle activity (±7 mV). The CWBV expo-sures produced several muscle responses that were considered to be spasms.CONCLUSIONS: The tested composite material spoked wheels do notdiffer, in vibration performance, from steel spoked wheels. The developedCWBV apparatus appears suitable for studying muscle activity in responseto WBV.iiiTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiiList of Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . xList of Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiAcknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . xiiDedication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiiiCo-Authorship Statement . . . . . . . . . . . . . . . . . . . . . . xiv1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1 Spasticity in Spinal Cord Injury . . . . . . . . . . . . . . . . 11.2 Wheelchair Vibration & Spasticity . . . . . . . . . . . . . . . 41.3 Whole Body Vibration & Spasticity . . . . . . . . . . . . . . 81.4 The ISO 2631-1 Standard . . . . . . . . . . . . . . . . . . . . 101.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131.6 Research Objectives . . . . . . . . . . . . . . . . . . . . . . . 151.7 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . 162 Wheelchair Wheel Comparison Study . . . . . . . . . . . . . 222.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 222.1.1 Hypothesis . . . . . . . . . . . . . . . . . . . . . . . . 252.1.2 Objective . . . . . . . . . . . . . . . . . . . . . . . . . 252.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262.2.1 Subject Recruitment . . . . . . . . . . . . . . . . . . 26ivTable of Contents2.2.2 Daily Activities Simulation Course Design . . . . . . 272.2.3 Instrumentation . . . . . . . . . . . . . . . . . . . . . 292.2.4 Test Preparation . . . . . . . . . . . . . . . . . . . . . 332.2.5 Test Procedure . . . . . . . . . . . . . . . . . . . . . . 352.3 Controlled Speed Wheel Comparison Studies . . . . . . . . . 362.3.1 Subject Recruitment . . . . . . . . . . . . . . . . . . 372.3.2 Study Design . . . . . . . . . . . . . . . . . . . . . . . 372.3.3 Instrumentation . . . . . . . . . . . . . . . . . . . . . 382.3.4 Test Preparation . . . . . . . . . . . . . . . . . . . . . 392.3.5 Test Procedure . . . . . . . . . . . . . . . . . . . . . . 392.4 Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 402.4.1 Acceleration Analysis . . . . . . . . . . . . . . . . . . 402.4.2 Frequency Power Spectrum Analysis . . . . . . . . . . 432.4.3 Statistical Analysis . . . . . . . . . . . . . . . . . . . 472.5 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 482.5.1 Daily Activities Simulation Course Results . . . . . . 482.5.2 Multiple Subject Controlled Speed Study Results . . 512.5.3 Single Subject Controlled Speed Study Results . . . . 512.6 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 562.6.1 Limitations . . . . . . . . . . . . . . . . . . . . . . . . 592.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 602.7.1 Recommendations for Future Work . . . . . . . . . . 612.8 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . 623 Controlled Whole Body Vibration Study . . . . . . . . . . . 653.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 653.1.1 Hypothesis . . . . . . . . . . . . . . . . . . . . . . . . 723.1.2 Objective . . . . . . . . . . . . . . . . . . . . . . . . . 723.1.3 Ethics Approval . . . . . . . . . . . . . . . . . . . . . 723.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 723.2.1 Specification of System Requirements . . . . . . . . . 733.2.2 Preliminary WBV System Concepts . . . . . . . . . . 833.2.3 Final Vibration System Concept . . . . . . . . . . . . 873.2.4 System Design & Assembly . . . . . . . . . . . . . . . 943.2.5 Subject Recruitment . . . . . . . . . . . . . . . . . . 1023.2.6 Test Preparation . . . . . . . . . . . . . . . . . . . . . 1033.2.7 Test Procedure . . . . . . . . . . . . . . . . . . . . . . 1063.3 Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 1083.3.1 Acceleration Analysis . . . . . . . . . . . . . . . . . . 1083.3.2 Muscle Activity Analysis . . . . . . . . . . . . . . . . 108vTable of Contents3.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1113.4.1 Acceleration Results . . . . . . . . . . . . . . . . . . . 1113.4.2 Muscle Activity Results . . . . . . . . . . . . . . . . . 1133.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1163.5.1 Limitations . . . . . . . . . . . . . . . . . . . . . . . . 1223.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1223.6.1 Recommendations for Future Work . . . . . . . . . . 1233.7 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . 1244 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1274.1 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1354.2 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . 137AppendicesA CWBV System Equation of Motion: Complete Solution . 138B CWBV System Manual . . . . . . . . . . . . . . . . . . . . . . 142C Ethics Approval Certificate . . . . . . . . . . . . . . . . . . . 149D CWBV Study Consent Form . . . . . . . . . . . . . . . . . . 152E DASC Study Questionnaire . . . . . . . . . . . . . . . . . . . 156F CWBV Study Recruitment Flyer . . . . . . . . . . . . . . . . 159viList of Tables1.1 ISO 2631-1 comfort reactions to vibration environments . . . 112.1 Elastic modulus & density of PBO and steel fibers. . . . . . . 242.2 DASC subjects’ baseline data . . . . . . . . . . . . . . . . . . 272.3 Daily activities simulation course obstacle description . . . . 282.4 Accelerometer (MDS 210U) specifications . . . . . . . . . . . 312.5 Octave frequency ranges . . . . . . . . . . . . . . . . . . . . . 462.6 Statistically significant differences in frequency content of vi-bration during the DASC study . . . . . . . . . . . . . . . . . 502.7 MSCS average wheelchair velocities. . . . . . . . . . . . . . . 512.8 SSCS average wheelchair velocities. . . . . . . . . . . . . . . . 512.9 SSCS mean acceleration peak and RMS results. . . . . . . . . 522.10 SSCS vibration FPS peak power mean, standard deviationand relative standard deviation results. . . . . . . . . . . . . . 532.11 SSCS vibration FPS total power mean, standard deviationand relative standard deviation results. . . . . . . . . . . . . . 543.1 The Penn spasm frequency scale . . . . . . . . . . . . . . . . 663.2 The modified Asworth scale . . . . . . . . . . . . . . . . . . . 673.3 Principal frequency weightings in 13 octaves (ISO 2631-1) . . . 783.4 Instantaneous accelerations, displacement amplitudes and forceoutput . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 823.5 Required and available vibration system specifications . . . . 873.6 Frequency range, spring constant value, natural frequencyand maximum load per spring set . . . . . . . . . . . . . . . . 933.7 Baseline data of the CWBV study subjects. . . . . . . . . . . 1033.8 Baseline data results. . . . . . . . . . . . . . . . . . . . . . . . 1043.9 Order of frequency of vibration . . . . . . . . . . . . . . . . . 1073.10 Skoo¨ld et al. study results . . . . . . . . . . . . . . . . . . . . 1103.11 Ratio between peak muscle activity and baseline muscle activity1103.12 Dominant frequency of vibration at seating system’s footrests 1133.13 CWBV study EMG Results . . . . . . . . . . . . . . . . . . . 115viiList of Figures1.1 ISO 2631-1 Health Guidance Caution Zone . . . . . . . . . . 111.2 Summary of past studies of comfort levels . . . . . . . . . . . 132.1 PBO fibers that make up a Spinergy R© wheel spoke . . . . . . 252.2 The daily activities simulation course obstacle layout . . . . . 282.3 The obstacles of the daily activities simulation course . . . . 292.4 Mechsense Digital MDS 210U accelerometer . . . . . . . . . . 302.5 Main axle accelerometer mounting method . . . . . . . . . . . 312.6 Footrest accelerometer mounting method . . . . . . . . . . . 322.7 Defined directions. . . . . . . . . . . . . . . . . . . . . . . . . 332.8 Obstacle course data recording . . . . . . . . . . . . . . . . . 362.9 Acceleration platform setup . . . . . . . . . . . . . . . . . . . 382.10 Typical vibration recordings . . . . . . . . . . . . . . . . . . . 422.11 Typical Time-Domain to Frequency-Domain Conversion . . . 452.12 Typical frequency power spectrum . . . . . . . . . . . . . . . 472.13 DASC subjective spasticity visual analog scale results. . . . . 482.14 DASC acceleration results. . . . . . . . . . . . . . . . . . . . . 492.15 DASC peak and total power per bin . . . . . . . . . . . . . . 502.16 SSCS mean acceleration peak (a) and RMS (b) results. . . . . 522.17 Vibration frequency content mean peak power per 10 Hz bin. 552.18 Vibration frequency content mean total power per 10 Hz bin. 553.1 The visual analog scale for subjective assessment of spasticity. 663.2 ISO 2631-1 health guidance caution zones. . . . . . . . . . . . 743.3 The distance between the marked value of weighted acceler-ation level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 753.4 Axes of the human body as defined by ISO 2631-1 . . . . . . 773.5 Commercially available fitness vibration platforms . . . . . . 833.6 Rotating eccentric discs system . . . . . . . . . . . . . . . . . 853.7 Large electrodynamic shaker . . . . . . . . . . . . . . . . . . . 863.8 The controlled whole body vibration system concept and itsmathematical model . . . . . . . . . . . . . . . . . . . . . . . 88viiiList of Figures3.9 System frequency response . . . . . . . . . . . . . . . . . . . . 903.10 Compression spring types: (a) coil spring. (b) disc spring. . . 923.11 CWBV system springs . . . . . . . . . . . . . . . . . . . . . . 933.12 CWBV system required force and frequency response . . . . . 943.13 Schematic layout of the controlled WBV system . . . . . . . . 953.14 CAD model of the controlled whole body vibration system’sseating station . . . . . . . . . . . . . . . . . . . . . . . . . . 963.15 The vibration system seating station elements . . . . . . . . . 963.16 The seating station support frame . . . . . . . . . . . . . . . 973.17 Vibration generation unit: The seating station . . . . . . . . 983.18 Bolt connecting support platform to electrodynamic shaker’sarmature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 983.19 Vibration generation unit: Electrodynamic shaker . . . . . . 993.20 Vibration generation unit: LDS P1000L amplifier . . . . . . . 993.21 The system control unit. . . . . . . . . . . . . . . . . . . . . . 1003.22 System control unit: Controller (LDS VLL1) . . . . . . . . . 1003.23 Vibration system testing . . . . . . . . . . . . . . . . . . . . . 1023.24 Muscles from which activity was recorded, using surface EMG,during the CWBV study . . . . . . . . . . . . . . . . . . . . . 1053.25 Seated subject with EMG sensors attached. . . . . . . . . . . 1063.26 Footrest vibration frequency power spectrum . . . . . . . . . 1123.27 Examples of muscle activity produced in response to the ap-plied WBV . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1143.28 CWBV study EMG results . . . . . . . . . . . . . . . . . . . 1163.29 The CWBV system’s frequency response . . . . . . . . . . . . 1184.1 Combined research results . . . . . . . . . . . . . . . . . . . . 135A.1 Mathematical model of the vibration system. . . . . . . . . . 138B.1 Supporting Springs . . . . . . . . . . . . . . . . . . . . . . . . 143B.2 Support platform resting on the support springs . . . . . . . 143B.3 Fastened wooden support platform - Top view. . . . . . . . . 144B.4 Feedback accelerometer bolted to seat-plate. . . . . . . . . . . 144B.5 Complete control system. . . . . . . . . . . . . . . . . . . . . 145B.6 Amplifier - LDS P1000L. . . . . . . . . . . . . . . . . . . . . . 146B.7 Controller - LDS VLL1. . . . . . . . . . . . . . . . . . . . . . 146ixList of AbbreviationsADL – Activities of Daily LivingAS – As SpunCWBV – Controlled Whole Body VibrationDASC – Daily Activities Simulation CourseDFT – Discrete Fourier TransformDRU – Data Recording UnitEMG – ElectromyographyFFT – Fast Fourier TransformFPS – Frequency Power SpectrumHM – High ModulusISO – International Organization for StandardizationMAS – Modified Ashworth ScaleMSCS – Multiple Subject Controlled Speed StudyPBO – Polyphenylene BensobisoxazoleRMS – Root Mean SquareSCI – Spinal Cord InjurySCU – System Control UnitSRC – Simulated Road CourseSRI – Stanford Research InstituteSSCS – Single Subject Controlled Speed StudyUBC – University of British ColumbiaVAS – Visual Analog ScaleVGU – Vibration Generation UnitWBV – Whole Body VibrationxList of Unitscm – CentimeterDC – Direct Currentg – Acceleration due to gravity (approximately 9.81 m/s2)Hz – Hertzkg – KilogramkN – Kilo Newtonm – Metermm – MillimeterN – Newtonrad – Radians – SecondV – VoltVAC – Volts of Alternating CurrentxiAcknowledgementsI would like to thank my supervisors, Dr. Peter Cripton and Dr. BonitaSawatzky for giving me the opportunity to conduct this research and fortheir academic and moral support. I am especially grateful for Dr. Cripton’sencouragement when I was very close to the breaking point.I would also like to thank:• Dr. John Howie and Dr. Carlos Ventura for allowing me to use theirelectrodynamic shaker system and for allowing me to test my appara-tus in the very busy earthquake lab.• Mr. Glenn Jolly and Mr. Scott Jackson for helping me set up, testand operate the electrodynamic shaker system.• Mr. Markus Fengler for helping me design and build my apparatus.• Dr. Jean-Se´bastien Blouin for his assistance and advice on Electromyo-graphy.• The members of the Injury Biomechanics Lab for all their help, espe-cially Jake and Seth for helping me transport all my equipment to andfrom the testing facility.• Dr. Barbara Hughes and Dr. Heather Finlayson for conducting themedical examinations and supervising the tests.• Everyone who participated as a subject in my research.Last but not least, I wish to thank my family:My parents, for their endless love and support.My wife, for her patience, love and encouragement.My daughter, for filling my life with joy.xiiTo my dear parents, loving wife and wonderful daughter.xiiiCo-Authorship StatementPart of the work presented in this thesis was a collaboration between the au-thor and Sigrid Vorrink (Sigrid Vorrink MSc. - Faculty of Human MovementSciences, Vrije Universiteit, Amsterdam, the Netherlands).Author’s Contribution to Work Described inChapter 2:The author’s contribution to performing the research included:• Adaption of the instrumentation mounting system• Daily Activities Simulation Course setup• Instrumentation calibration & installation• Data collectionThe author’s contribution to the data analysis included:• MATLAB routines code writing• Acceleration & frequency content analysisThe author’s contribution to the preparation of the published version ofChapter 2 included capture and editing of the included images. The datafrom the DASC study and Controlled Speed study was reanalyzed by theauthor, using a slightly varied method than that utilized during the collab-oration and presented in the published article.Author’s Contribution to Work Described inChapter 3:The work described in Chapter 3, in its entirety, was the sole effort of theauthor.xiv1. IntroductionIt is estimated that 300,000 people in North America were living with SpinalCord Injury (SCI) in 2008. There are approximately 12,000 new cases ofspinal cord injury (SCI) each year in the United States1. In Canada, morethan 1000 individuals suffer a spinal cord injury every year2. These newcases of SCI add to the population of individuals already living with chronicSCI (chronic SCI: more than 1 year since the injury).1.1 Spasticity in Spinal Cord Injury & itsExperimental DetectionHuman movement is very complex and involves different regions of the brainand spinal cord. Muscle contraction speed, timing, direction and force areeach controlled by different regions in the brain and several pathways de-scend through the spinal cord, carrying inhibitory and excitatory signalsfrom the brain to the rest of the body. A spinal cord injury can damagethe spinal pathways at the lesion site, disrupting or severing the flow ofregulatory impulses through the pathways to spinal levels below the site ofinjury.Spinal cord injury has three defined phases: Acute, Secondary and Chronic3.The acute phase is described as the time from initial injury up to severaldays after injury. The secondary phase lasts up to several weeks after injuryand the chronic phase begins roughly around 1 year after initial injury.In the acute phase, mechanical damage of neural and soft tissue results inlocal cell death. Additional neural injury is caused by increased compressionof the spinal cord due to edema and hemorrhaging. The secondary phaseis characterized by programmable cell death (Apoptosis). Inflammatory11.1. Spasticity in Spinal Cord Injurycells invade the spinal cord tissue, the lesion grows and cells in the regionaround the original lesion also begin to die. During the chronic phase,cell death continues to spread at the lesion and scarring of the spinal cordoccurs. Damaged axons (nerve fibers that conduct electrical impulses) andsurrounding undamaged axons may begin to regenerate but grow only upto 1 mm per day3.It is estimated that 65–78% of individuals with chronic spinal cord injuryexperience spasticity. Spasticity can impair function and degrade qual-ity of life by causing pain and fatigue, contributing to the development ofcontractures, pressure ulcers, infections, negative self-image and interferingwith a wheelchair user’s seating, transfers, and wheeling4;5;6. Spasticity is amotor disorder, characterized by an exaggeration of the stretch reflexes andpersistent muscle contraction that cannot be relaxed7. There are 3 distincttypes of spasms: Clonus, Flexor spasms & Extensor spasms.Clonus, which can be elicited in individuals with SCI by a rapid movementof the ankle (passive dorsiflexion), is a rhythmic contraction and relaxationof the calf muscles, and results in rapid and repetitive motions of the foot,causing a seated individual’s knee to bounce up and down8. An individualexperiencing Clonus while in a wheelchair will often have to readjust his/herfoot in the footrest as the repeated bouncing will cause the foot to shift itsposition, sometimes falling completely out of the footrest. The uncontrol-lable bouncing of the knee during clonus can also unbalance any items thatthe individual might be carrying on his/her lap, potentially causing the itemto fall to the ground.Flexor spasms consist of coordinated flexion of the limb at multiple jointswhile Extensor spasms are observed clinically in SCI as an uncontrolled,multi-joint extension of the limb9. Both Flexor and Extensor spasms occurin response to painful stimuli10. A seated individual experiencing extensorspasms may experience a sudden extension of one or both legs. This caneither cause the foot to leave the footrest (as the leg kicks forward) or forcethe individual’s trunk backwards (as the leg suddenly straightens), causinghim/her to lose balance. Flexor spasms will usually cause an uncomfortabletightening of an individual’s muscles.There are several methods commonly used for assessment of spasticity.These methods include self-report questionnaires, clinical assessments andbiomechanical methods. Self report questionnaires usually include a set of21.1. Spasticity in Spinal Cord Injurydiscrete scores or a continuous scale, and are designed to assess the individ-ual’s subjective perception of spasticity.The Modified Ashworth scale (MAS) is the most frequently used clinicalspasticity assessment method11;12. The MAS assessment is simple and quick,is easy to carry out (although experience is required) and requires no instru-mentation. Assessing spasticity with the MAS involves an imposed stretchof the tested muscle and assigning a score, from 0 to 4, to the muscle tone(see table 3.2 for MAS score definitions).A commonly used biomechanical method for assessment of spasticity is Elec-tromyography (EMG). EMG is a method for detection and recording ofmuscle activity by measuring the electrical potential generated by musclecells during muscle activity and rest. Surface EMG is conducted by placingsensors on the skin, directly over the muscle of interest, and has been shownto be a reliable and repeatable method for documenting muscle motor unitbehavior13;14;15;16;17.The EMG data can be used to determine when a muscle is active, the de-gree of activation and to estimate muscle force. The amplitude of the EMGsignal generally increases with an increase in muscle force or contractionvelocity of the muscle. However the relationship between muscle contrac-tion force/velocity and the EMG signal amplitude can only be qualitativelyassessed18. See section 3.1 on page 68 for additional details on spasticityassessment methods.Many people with SCI use wheelchairs as their main, if not only, means oftransport and are therefore in a wheelchair for many hours every day. Awheelchair in motion, like any other vehicle, undergoes vibration mainly dueto ride surface roughness and encountered obstacles (potholes, curbs, steps,door thresholds, etc). As a result, the wheelchair user experiences wholebody vibration (WBV) which, depending on the frequency, amplitude andduration of the vibration, may cause decreased comfort, activity interference,impaired health and pain19.Very little is known about how whole body vibration affects wheelchair userswith spinal cord injury (SCI) or how to minimize any adverse effects20. Incontrast, the International Organization for Standardization (ISO) created aset of standards (ISO 2631-1) for “people in normal health who are regularlyexposed to vibration”21. The recommendations on health risks in ISO 2631-31.2. Wheelchair Vibration & Spasticity1 are mainly based on exposures of uninjured individuals to WBV, withexposure durations in the range of 4–8 hours. The ISO 2631-1 standard isdescribed in more detail in subsection 1.4 on page 11.1.2 Wheelchair Vibration & SpasticityWheelchair users often adopt a slouched posture, due to improper lumbarsupport and trunk instability, which increases trunk moment and disc de-formation and leads to increased inter-vertebral disc pressure and pain22;23.This posture may be assumed for up to 14 hours a day, during which thewheelchair user experiences a variety of vibrations and shocks which can con-tribute to increased development of neck and back discomfort and pain24;25.It is because of the prolonged sitting and regular exposure to WBV thatwheelchair use presents one of the greatest risks of low back pain and in-jury26. Recently, researchers have begun studying how different wheelchairstructures, components and use influence wheelchair vibration.VanSickle et al. used the standard Double-Drum and Curb-Drop wheelchairfatigue tests (which both use a 100 kg dummy weight and not a humansubject), a simulated road course and field tests (both using human subjectswith SCI) to evaluate dynamic reaction forces and moments exerted on awheelchair and rider27;28. The Double-Drum test is designed to simulatecommon obstacles that a wheelchair user encounters daily (door thresholds,ride surface cracks, etc.) The test consists of two drums with slats mountedon each drum, rotating at different speeds beneath the wheelchair. The slatsimpact the wheels and induce vibration.The Curb-Drop test simulates a wheelchair dropping off the edge of a curb.During the test, the wheelchair is simply dropped from a 5 cm height onto theground. The simulated road course was also designed to simulate obstaclesthat a wheelchair user may encounter during activities of daily living andconsisted of eight obstacles, rigidly fixed to a flat concrete surface, overwhich the subject would wheel.The field test consisted of the subjects using a wheelchair outside of thelaboratory for their daily activities, as they normally do, for at least 4 hours.Sensors, fixed to the wheelchair’s wheel hubs and front casters, measureddynamic ground reaction forces and moments during all tests, except for the41.2. Wheelchair Vibration & Spasticityfield test in which no front caster sensors were used.VanSickle et al. found that the simulated road course produced force andmoment distributions and extreme values that were similar to those pro-duced during the field tests. This result validates the simulated road coursedesign as a model of ‘real world’ wheelchair use conditions. The standard-ized fatigue testing produced results that differed from both human trials.The Double-Drum and Curb-Drop tests produced maximal forces and mo-ments that were significantly higher than are produced during normal use.VanSickle concluded that the wheelchair is exposed to infrequent, but highmagnitude vertical forces and that the rider absorbs most of the energy ofthese shocks.DiGiovine et al. examined wheelchair user perceived ride comfort duringpropulsion by having 30 subjects (25 of the subjects had SCI) wheel throughan “Activities of Daily Living” course with 7 different wheelchairs29. A vi-sual analog scale based survey, completed by the subjects, was used to assesssubjective perception of ride comfort. Their results showed that wheelchairshaving a high degree of adjustability were perceived as more comfortable andhaving better basic ergonomics than wheelchairs having minimal adjustabil-ity.VanSickle et al. also had 16 manual wheelchair users (most with SCI) wheelthrough a simulated road course and use the same wheelchair in their regulardaily activities to study how dynamic acceleration affects wheelchair-ridercomfort30. Two 3-dimensional accelerometers, mounted to the frame of awheelchair and to a bite bar, measured acceleration at the wheelchair frameand head of the user.The field test results showed that subjects and their wheelchairs were ex-posed to a few high-acceleration events rather than consistent, low magni-tude, accelerations. The acceleration at the wheelchair frame and at thehead, for the simulated road course, was found to exceed the 8-h “fatigue-decreased performance boundary” defined by ISO 2631-1. Considerabledampening beyond approximately 20 Hz, probably due to vibration absorp-tion by the subjects’ bodies, was also found. VanSickle et al. concluded thatvibration may be a contributing factor to fatigue among manual wheelchairusers and that this could lead to injury.Maeda surveyed 33 manual wheelchair users and found that wheelchair vi-51.2. Wheelchair Vibration & Spasticitybration affected the users’ ride comfort and that the effect of different ridesurfaces played an important role31. Maeda found that 30 subjects wereusing a wheelchair more than 8 hours per day, 32 subjects felt discomfortduring a wheelchair ride in outside use and 31 subjects complained aboutwheelchair vibration from from tiled walkways, gravel walkways and walkingblocks. Maeda recorded complaints about transmitted vibrations from thewheelchair to the neck and head, lower back and buttocks of the subjects,with the most affected location being the neck. Vertical vibrations werefound to be the most common vibration transmitted from the wheelchair tothe user and vertical vibrations seem to be the main cause of complaints bywheelchair users.Several studies examined how seat cushions and back supports influence thetransmission of vibrations during manual wheelchair propulsion. One studyhad 10 non-SCI subjects wheel through an activities of daily living courseusing 4 seat cushions and 4 back supports and measured vibrations using3-axis accelerometers32. The study found differences among the seat cush-ions and back supports that appeared to be due to the seat cushion/backsupport design and postural support. The study also found that vibrationsgenerated by single-event shocks and repeated shocks (as opposed to oscil-latory motions and self-generated vibrations) tend to reside in the rangeof frequencies most sensitive to humans and that vibrations in this rangeof frequencies have the greatest effect on the transmission of whole-bodyvibration during manual wheelchair propulsion.Another study by DiGiovine used the same methods to compare 16 ran-domly selected seating systems as well as the 32 participating subjects’ ownsystems. The study found the same results and concluded that that theindividuals were not using the most appropriate seating system in terms ofthe reduction of vibration transmission33.The effect of the surface over which the wheelchair rolls on wheelchair vibra-tion has also been explored. Cooper et al. had 10 able bodied subjects wheelover 6 different sidewalk surfaces, at 1 m/s and 2 m/s, in both a manualand electric wheelchair equipped with 3-axis accelerometers mounted at theseat and footrests34. They found significant differences in peak accelerationsbetween the seat and footrest and between the sidewalk surfaces at 1 m/sbut no difference in the work required to propel over the surfaces. Ridingover the tested surfaces with an electric powered wheelchair induced wholebody vibration that exceeded the ISO 2631-1 exposure limit. At 1 m/s vi-61.2. Wheelchair Vibration & Spasticitybration from 4 of the 6 surfaces exceeded the exposure limit after more than8 hours of driving. At 2 m/s all 6 surfaces induced whole body vibrationthat exceeded the exposure limit after less than 3 hours.Koontz et al. had 11 manual wheelchair users wheel over 8 different surfacesusing wheelchairs equipped with a 3-dimensional force and torque-sensingpush-rim35. Koontz found that propulsion forces and moments are con-siderably higher when wheelchair users are starting up, propelling uphill,or pushing over surfaces that imposed greater resistance to propulsion andconcluded that frequent starting and stopping and regularly propelling onoutdoor and inclined surfaces may increase the likelihood of users develop-ing upper-limb injuries because of increased and cumulative loading on thearms.Other studies have focused on the effectiveness of suspension systems. Cooperet al. examined how suspension caster forks and rear-suspension systemsinfluence the shock and vibration transmitted to an occupant of a manualwheelchair36. Six manual wheelchairs were fitted with a 3-axis accelerometerat the footrest and at the seat. The wheelchairs were tested on a double-drum wheelchair test machine with a 75 kg and a 100 kg test dummy. Cooperfound that wheelchairs with suspension had significantly different frequen-cies at which the peak accelerations occurred for both the seat and thefootrest. They also found that rear suspension systems are not clearly su-perior to traditional designs, despite reducing some of the factors related toshock and vibration exposure. They recommended that manual wheelchairusers seriously consider suspension caster forks, especially if they are activeor experiencing chronic pain, because suspension caster forks reduce theshock and vibration exposure to the user.Kwarciak et al. investigated the ability of suspension manual wheelchairsto reduce seat accelerations during curb descents of various heights26. Onesubject descended 3 simulated curbs (curb height: 5 cm, 10 cm & 15 cm)with 16 manual wheelchairs fitted with a 3-axis accelerometer at the seat.Wheelchair orientation was measured using a camera & marker system (Op-totrak). Kwarciak found that suspension manual wheelchairs provide somelevel of vibration suppression but that the orientation of the wheelchairduring the activity limits the suppression of vibration. Kwarciak’s resultsshowed that suspension wheelchairs transmitted significantly lower peak seataccelerations than folding wheelchairs during the 5 cm curb descents andsignificantly lower frequency weighted peak seat accelerations during the 571.3. Whole Body Vibration & Spasticitycm and 10 cm curb descents. When weight was considered, the suspensionwheelchairs had significantly lower peak seat accelerations than the lighterrigid wheelchairs during 5 cm curb descents. Kwarciak concluded that theangle at which a wheelchair impacts its landing surface can notably influencethe effectiveness of its suspension.Requejo et al. exposed 10 manual wheelchair users to repeated bumps whileseated on wheelchairs with either rigid frames or with rear suspension anddetermined the forces transmitted from the seat and the accelerations ex-perienced by the wheelchair user20. Four wheelchairs were fitted with 7load cells below the seat and backrest frame. Accelerometers (1-axis) weremounted on the wheelchairs’ wheel hub and seat frame and on a bicyclehelmet worn by the test subject. Vibrations were generated by placing thewheelchairs’ rear wheels over a rotating drum with a small metal rod fixedalong the drum length.Requejo et al. found that wheelchair suspension reduced the force and accel-erations experienced by manual wheelchair users. They also found that, athigher speeds, wheelchair suspension may improve the function of manualwheelchair users in terms of comfort and that this was especially true forusers without trunk muscle innervations. They determined that suspensionconfiguration, trunk muscle innervations, and speed may influence vibra-tion exposure and that minimizing the forces transmitted to the rider isimportant for ride quality.To date, no investigators have studied the relationship between wheelchairwheel design and wheelchair vibration or the relationship between wheelchairvibration and muscle spasticity in individuals with spinal cord injury.1.3 Whole Body Vibration & SpasticityThe effects of Whole Body Vibration (WBV) on non-SCI individuals havebeen extensively researched and are well defined due primarily to the highincidence of injuries correlated with sustained exposure to WBV19. Manystudies have examined the effects of occupational exposure to WBV, mostlyon drivers and operators of heavy equipment, and several literature reviewsand standards have been dedicated to the exposure of humans without SCIto WBV21;37;38;39.81.3. Whole Body Vibration & SpasticityHuman response to whole body vibration may be considered to involve de-graded comfort, activity interference, low magnitude vibration perceptionand motion sickness19. Combinations of these effects will often appear si-multaneously. The dynamic interaction at the interface of the vibratingsurface and the body, as well as body posture, direction, frequency, magni-tude and duration affect the transmission of vibration to the body. WBVhas also been associated with impaired health and studies have found a vari-ety of health problems that are either caused by or exacerbated by exposureto WBV19.Systems adversely affected by WBV include the gastrointestinal system,male & female reproductive systems, visual and vestibular (balance andspatial positioning) systems40;19;41;42;43. There is also evidence of exposureto WBV resulting in spine pathologies such as vertebral degeneration, her-niated discs and lower back pain44;45;46;47;48The degree to which vibration is transmitted to the body depends on thevibration’s frequency. The range of frequencies most often associated witheffects of WBV on health, activities and comfort is approximately 0.5–100Hz19. Exposure to vibration in the 4–10 Hz range has been observed tocause pain in the chest and abdomen. Back pain commonly occurs with vi-bration in the 8–12 Hz range and vibration at 10–20 Hz can cause headaches,eyestrain, and irritations in the intestines and bladder49. The magnitudes ofinterest with WBV are within the range 0.01–10 m/s2. At frequencies below1 Hz and above 20 Hz higher magnitudes are required for perception19.Like any physical structure or mechanical system, individual parts and sys-tems of the human body have resonant frequencies. When the body isexposed to vibration at frequencies that correspond to the resonant fre-quencies of certain body parts/systems, the vibration magnitude in thoseparts/systems will be greater than in non–resonating body parts/systems.WBV in the 2–5 Hz frequency range can often be amplified by up to 240%in the neck and lumbar regions. Trunk organs can experience vibrationamplification of over 150% at the 4–6 Hz frequency range while at the 20–30 Hz range vibration at the head and shoulders can be amplified by over300%50. The effect of the vibration will therefore be more profound in theresonating body parts/systems.As the impact of vibration on the human body varies at different frequencies,91.4. The ISO 2631-1 Standardthe influence of vibration frequency on the human body must be accountedfor. This is now commonly achieved by frequency weighting. The frequencyweightings are values by which the vibration magnitude at each frequencyis to be multiplied in order to ’weight’ it according to its effect on the body(See table 3.3 on page 78 for a complete list of weightings).1.4 The ISO 2631-1 StandardThe ISO 2631-1 standard was created with the intention of defining methodsof quantifying WBV in relation to health and comfort, probability of vibra-tion perception and incidence of motion sickness in healthy humans. Thestandard is applicable to vibrations transmitted through supporting sur-faces to the human body as a whole21. ISO 2631-1 uses frequency-weightedRoot-Mean-Square (RMS) acceleration to evaluate comfort & health effectsof vibration on seated humans.The RMS value of a vibration signal is an important measure of its ampli-tude. It is the square root of the average of the squared values of amplitude(The instantaneous amplitude values of the waveform are squared and thesesquared values are averaged over a time interval of at least one period ofthe wave). The RMS value of a weighted acceleration signal is defined inEquation 1.1aRMS =√1T∫ T0a2w(t)dt (1.1)where aw(t) is the weighted acceleration as a function of time. See Table 3.3on page 78 for additional information on acceleration frequency weighting.The ISO 2631-1 standard does not provide a quantitative relationship be-tween vibration exposure and health effects but does provide a graph con-taining Health Guidance Caution Zones (Figure 1.1). The Health GuidanceCaution Zones are defined in terms of frequency-weighted acceleration mag-nitude and exposure duration. Within the caution zone itself, shown in gray,caution with respect to potential health risks is advised. For exposures belowthe zone, shown in green, health effects have not been clearly documentedand/or objectively observed. Health risks are likely for exposures above the101.4. The ISO 2631-1 Standardzone, shown in red.Figure 1.1: ISO 2631-1 Health Guidance Caution Zone (Image enhancedby author - colors do not appear in original image). Green Zone: Healtheffects undocumented. Gray Zone: Caution advised. Red Zone: Healthrisks likely.The comfort levels, as defined in ISO 2631-1 (shown in Table 1.1) are notclearly defined as they overlap.Acceleration Level [m/s2] Comfort Level> 2 Extremely uncomfortable1.25 – 1.5 Very uncomfortable0.8 – 1.6 Uncomfortable0.5 – 1 Fairly uncomfortable0.315 – 0.63 A little uncomfortable< 0.315 Not uncomfortableTable 1.1: ISO 2631-1 comfort reactions to vibration environmentsSeveral researchers have also studied comfort levels associated with expo-sure to WBV. Whitham and Griffin showed that for vertical vibration thereis a distinct difference in the locations of discomfort on the body at differ-111.4. The ISO 2631-1 Standardent frequencies and that vibration level does not have much effect on thelocations51. They also found the maximum sensitivity to vertical vibrationacceleration to be in the 4–16 Hz range.Kiiski et al. exposed four healthy males, in a standing position, to vibrationsat frequencies from 10 to 90 Hz in 5 Hz increments and noted that everysubject felt some discomfort specifically in the 20–25 Hz frequency rangewhen the amplitude was 0.5 mm or greater52. Fothergill and Griffin hadseated subjects adjust the level of vibration, at a constant frequency, andcategorize their degree of comfort53;54. A summary, by Kaneko et al55,of several other studies on comfort levels during exposure to whole bodyvibration is shown in Figure 1.2.Kaneko55 set out to clarify the relationship between whole body vibrationand perceived comfort by vibrating seated subjects with random stimuli be-tween 1 and 100 Hz. The signals were frequency-weighted in accordance withISO 2631-1 and the RMS values were adjusted to be equal. Kaneko foundthat signals with different frequency content will illicit differing evaluationsof comfort even if the frequency-weighted RMS acceleration is equal.Kaneko produced a vibration comfort scale with categories that do not over-lap (unlike the ISO2631-1 comfort levels scale). The vibration level in eachof Kaneko’s categories was higher than the corresponding ISO 2631-1 cat-egories. Kaneko also found that the degree of discomfort increased whenthe signal contained a higher percentage of lower frequency components andthat the differences in comfort perception increased in proportion to theincrease in frequency-weighted RMS acceleration.121.5. Conclusion                                                                                                                                                                                                                                                                                                                               Source (year) Scale Mean magnitude Standard deviation Situation(1) Fothergill Very unpleasant 2.5 1 Seated subjects(1972) Unpleasant 1.7 0.83 Levels of 8 Hz sinusoidMildly unpleasant 1.1 0.5Not unpleasant 0.7 0.35Noticeable 0.3 0.14(2) Jones and Saunders Very unpleasant 3.7 – Seated subjects(1974) Very uncomfortable 2.2 – Equivalent levels ofUncomfortable 1.2 – 8 Hz sinusoidMean threshold of discomfort 0.7 –Not uncomfortable 0.33             –(3) Oborne and Clarke Very uncomfortable > 2.3 – Standing subjects(1974) Uncomfortable   1.2–2.3 – Levels of 10 Hz sinusoidFairly uncomfortable   0.5–1.2 –Fairly comfortable 0.23–0.5 –Very comfortable <0.23 –(4) Fothergill and Griffin Very uncomfortable 2.7 0.91 Seated subjects(1977) Uncomfortable 1.8 0.77 Levels of 10 Hz sinusoidMildly uncomfortable 1.1 0.47Noticeable, but not uncomfortable 0.4 0.16(5) ISO2631-1 Extremely uncomfortable Greater than 2 –(1997) Very uncomfortable 1.25–2.5 –Uncomfortable   0.8–1.6 –Fairly uncomfortable 0.5–1 –A little uncomfortable 0.315–0.63 –Not uncomfortable Less than 0.315Figure 1.2: Summary of past studies of comfort levels during exposure towhole body vibration (summarized by Kaneko et al.55).1.5 ConclusionWheelchair use presents one of the greatest risks of pain and injury dueto prolonged sitting, improper posture and regular exposure to whole bodyvibration22;23;26. Although the effects of whole body vibration (WBV) onhealthy individuals have been extensively researched and an ISO standarddedicated to people in normal health who are regularly exposed to vibration,very little is known about how WBV affects wheelchair users with spinal cordinjury (SCI)21;20. It has been shown however, that regular wheelchair usecan expose the user to WBV that exceed the levels defined as safe by theISO standard30;32;34.The few studies that have been dedicated to wheelchair vibration have ex-amined dynamic reaction forces and moments exerted on the wheelchair anduser, and the perceived ride comfort during propulsion and in response to vi-bration. However, the spasticity of an individual with SCI during wheelchair131.5. Conclusionuse has not been studied27;29;30;31.Other studies have examined the influence of individual components suchas cushions and back-supports, ride surfaces and suspension systems onwheelchair vibration and have shown that individual elements of the wheelchairsystem can affect wheelchair vibration and transmissibility of vibration tothe user. However, none of these studies have examined the influence ofwheel design on wheelchair vibration32;33;36;34;35;26;20.The simulated road course, which has been used by many of the wheelchairvibration studies to induce wheelchair vibration, has been shown to pro-duce force and moment distributions and extreme values that were simi-lar to those produced during regular (out of lab) wheelchair use27. Otherwheelchair vibration inducing methods, such as the Double-Drum and Curb-Drop methods, have been shown to produce forces and moments that weresignificantly higher than are produced during normal use.Suspension systems have been shown to provide a better ride by reducingthe accelerations experienced by the wheelchair user. However, a suspensionsystem adds weight to the wheelchair and a heavier wheelchair is more dif-ficult to transport. Lifting and storing/retrieving a heavy wheelchair whentransferring in and out of a car, for example, may contribute to the increasedrisk of upper-limb pathology20. Wheelchair suspension systems may also ab-sorb not only the energy from impacts and vibration, but also the energyintended for propulsion.Reducing wheelchair vibration via an alternative wheel design, that doesnot add to the wheelchair’s weight, instead of a suspension system couldprovide the benefit of reducing the risk of pain and injury due to wholebody vibration without adding the risks and inconvenience associated witha heavier wheelchair.Anecdotal reports by manual wheelchair users of lower limb muscle spasmsbeing triggered by wheelchair vibration suggest a possible correlation be-tween WBV and spasticity.1 However, no study to date has looked into apossible correlation between wheelchair vibration (and the resulting WBVexperienced by the user) and spasticity in individuals with SCI.1Reports made in private conversations with Dr. Bonita Sawatzy PhD, of the Universityof British Columbia.141.6. Research Objectives1.6 Research ObjectivesThis thesis had 2 objectives:1. Examine the influence of wheelchair wheel design on wheelchair vibra-tion magnitude and frequencies and to study the influence of wheelchairvibration on subjective estimates of spasticity by manual wheelchairusers with spinal cord injury2. Design and construct an apparatus that allows non-injurious exposureof individuals with spinal cord injury to magnitude and frequency con-trolled whole body vibration and to record muscle activity in responseto the exposure, using electromyography, for analysis and identifica-tion of spasticity.151.7. Bibliography1.7 Bibliography[1] National Spinal Cord Injury Statistical Center. Spinal cord injury factsand figures at a glance. Technical report, University of Alabama, June2009.[2] Gwynedd E Pickett, Mauricio Campos-Benitez, Jana L Keller, and NeilDuggal. Epidemiology of traumatic spinal cord injury in canada. Spine(Phila Pa 1976), 31(7):799–805, Apr 2006.[3] Claire E Hulsebosch. Recent advances in pathophysiology and treat-ment of spinal cord injury. Adv Physiol Educ, 26(1-4):238–255, Dec2002.[4] C. Skld, R. Levi, and A. Seiger. Spasticity after traumatic spinalcord injury: nature, severity, and location. Arch Phys Med Rehabil,80(12):1548–1557, Dec 1999.[5] F. M. Maynard, R. S. Karunas, and W. P. Waring. Epidemiologyof spasticity following traumatic spinal cord injury. Arch Phys MedRehabil, 71(8):566–569, Jul 1990.[6] M. M. Adams and A. L. Hicks. Spasticity after spinal cord injury.Spinal Cord, 43(10):577–586, Oct 2005.[7] R. R. Young. Spasticity: a review. Neurology, 44(11 Suppl 9):S12–S20,Nov 1994.[8] J. M. Hidler and W. Z. Rymer. A simulation study of reflex instabilityin spasticity: origins of clonus. IEEE Transactions on RehabilitationEngineering, 7(3):327–340, Sept. 1999.[9] Ming Wu, T. George Hornby, Jennifer Hilb, and Brian D Schmit. Ex-tensor spasms triggered by imposed knee extension in chronic humanspinal cord injury. Exp Brain Res, 162(2):239–249, Apr 2005.[10] G. Sheean. The pathophysiology of spasticity. Eur J Neurol, 9 Suppl1:3–9; dicussion 53–61, May 2002.[11] R. W. Bohannon and M. B. Smith. Interrater reliability of a modifiedashworth scale of muscle spasticity. Phys Ther, 67(2):206–207, Feb1987.161.7. Bibliography[12] F. Biering-Srensen, J. B. Nielsen, and K. Klinge. Spasticity-assessment:a review. Spinal Cord, 44(12):708–722, Dec 2006.[13] M. P. Kadaba, M. E. Wootten, J. Gainey, and G. V. Cochran. Re-peatability of phasic muscle activity: performance of surface and intra-muscular wire electrodes in gait analysis. J Orthop Res, 3(3):350–359,1985.[14] A. Gollhofer, G. A. Horstmann, D. Schmidtbleicher, and D. Schnthal.Reproducibility of electromyographic patterns in stretch-shorteningtype contractions. Eur J Appl Physiol Occup Physiol, 60(1):7–14, 1990.[15] A. M. Sherwood, W. B. McKay, and M. R. Dimitrijevi. Motor controlafter spinal cord injury: assessment using surface emg. Muscle Nerve,19(8):966–979, Aug 1996.[16] C. Skld, K. Harms-Ringdahl, C. Hultling, R. Levi, and A. Seiger. Si-multaneous ashworth measurements and electromyographic recordingsin tetraplegic patients. Arch Phys Med Rehabil, 79(8):959–965, Aug1998.[17] B. Zupan, D. S. Stoki, M. Bohanec, M. M. Priebe, and A. M. Sher-wood. Relating clinical and neurophysiological assessment of spasticityby machine learning. Int J Med Inform, 49(2):243–251, Apr 1998.[18] Carlo J. De Luca. The use of surface electromyography in biomechanics.Journal of Applied Biomechanics, 13 (2):135–163, 1997.[19] M. J. Griffin. Handbook of Human Vibration. Academic Press Inc.,1990.[20] Philip S Requejo, Grigor Kerdanyan, Jean Minkel, Rodney Adkins, andRobert Waters. Effect of rear suspension and speed on seat forces andhead accelerations experienced by manual wheelchair riders with spinalcord injury. J Rehabil Res Dev, 45(7):985–996, 2008.[21] International Organization for Standardization. Mechanical vibrationand shock – Evaluation of human exposure to whole-body vibration –Part 1: General requirements (ISO 2631-1), 1997.[22] Samuelsson K, Larsson H, Thyberg M, and Tropp H. Back pain andspinal deformity common among wheelchair users with spinal cord in-juries. Scandinavian Journal of Occupational Therapy, 3:28–32, 1996.171.7. Bibliography[23] Malcolm H Pope, Kheng Lim Goh, and Marianne L Magnusson. Spineergonomics. Annu Rev Biomed Eng, 4:49–68, 2002.[24] Hoover AE, Cooper RA, Dan D, Dvorsnak M, Cooper R, FitzgeraldSG, and Boninger ML. Comparing driving habits of wheelchair users:Manual versus power. In Proceedings of the Rehabilitation Engineeringand Assistive Technology Society of North America (RESNA) 26th In-ternational Conference on Technology & Disability: Research, Design,Practice, and Policy, 2003.[25] Michael L Boninger, Rory A Cooper, Shirley G Fitzgerald, Jessica Lin,Rosemarie Cooper, Brad Dicianno, and Betty Liu. Investigating neckpain in wheelchair users. Am J Phys Med Rehabil, 82(3):197–202, Mar2003.[26] Andrew M Kwarciak, Rory A Cooper, and Shirley G Fitzgerald. Curbdescent testing of suspension manual wheelchairs. J Rehabil Res Dev,45(1):73–84, 2008.[27] D. P. VanSickle, R. A. Cooper, and M. L. Boninger. Road loads actingon manual wheelchairs. IEEE Transactions on Rehabilitation Engineer-ing, 8(3):371–384, Sept. 2000.[28] International Organization for Standardization. Wheelchairs – Part 8:Requirements and test methods for static, impact and fatigue strengths(ISO 7176-8), 1998.[29] M. M. DiGiovine, R. A. Cooper, M. L. Boninger, B. M. Lawrence, D. P.VanSickle, and A. J. Rentschler. User assessment of manual wheelchairride comfort and ergonomics. Arch Phys Med Rehabil, 81(4):490–494,Apr 2000.[30] D. P. VanSickle, R. A. Cooper, M. L. Boninger, and C. P. DiGiovine.Analysis of vibrations induced during wheelchair propulsion. J RehabilRes Dev, 38(4):409–421, 2001.[31] Setsuo Maeda, Makoto Futatsuka, Jiro Yonesaki, and Maki Ikeda. Re-lationship between survey results of vibration complaints of wheelchairusers and vibration transmissibility of manual wheelchair. Environmen-tal Health and Preventative Medicine, 8:82–89, July 2003.[32] Carmen P DiGiovine, Rory A Cooper, Erik Wolf, Shirley G Fitzger-ald, and Michael L Boninger. Analysis of whole-body vibration during181.7. Bibliographymanual wheelchair propulsion: a comparison of seat cushions and backsupports for individuals without a disability. Assist Technol, 15(2):129–144, 2003.[33] Carmen P DiGiovine, Rory A Cooper, Shirley G Fitzgerald, Michael LBoninger, Erik J Wolf, and Songfeng Guo. Whole-body vibration dur-ing manual wheelchair propulsion with selected seat cushions and backsupports. IEEE Trans Neural Syst Rehabil Eng, 11(3):311–322, Sep2003.[34] Rory A Cooper, Erik Wolf, Shirley G Fitzgerald, Annmarie Kellerher,William Ammer, Michael L Boninger, and Rosemarie Cooper. Evalua-tion of selected sidewalk pavement surfaces for vibration experiencedby users of manual and powered wheelchairs. J Spinal Cord Med,27(5):468–475, 2004.[35] Alicia M Koontz, Rory A Cooper, Michael L Boninger, Yusheng Yang,Bradley G Impink, and Lucas H V van der Woude. A kinetic analysisof manual wheelchair propulsion during start-up on select indoor andoutdoor surfaces. J Rehabil Res Dev, 42(4):447–458, 2005.[36] Rory A Cooper, Erik Wolf, Shirley G Fitzgerald, Michael L Boninger,Rhys Ulerich, and William A Ammer. Seat and footrest shocks andvibrations in manual wheelchairs with and without suspension. ArchPhys Med Rehabil, 84(1):96–102, Jan 2003.[37] International Organization for Standardization. Mechanical vibrationand shock – Human exposure – Vocabulary (ISO 5805), 1997.[38] International Organization for Standardization. Mechanical vibrationand shock – Guidance on safety aspects of tests and experiments withpeople – Part 1: Exposure to whole-body mechanical vibration and re-peated shock (ISO 13090-1), 1998.[39] British Standards Institution. Measurement and Evaluation of HumanExposure to Whole-Body Mechanical Vibration and Repeated Shock (BS6841), 1987.[40] E. Aantaa, E. Virolainen, and V. Karskela. Permanent effects of lowfrequency vibration on the vestibular system. Acta Otolaryngol, 83(5-6):470–474, 1977.[41] H. Seidel. Selected health risks caused by long-term, whole-body vibra-tion. Am J Ind Med, 23(4):589–604, Apr 1993.191.7. Bibliography[42] M. Bovenzi and C. T. Hulshof. An updated review of epidemiologicstudies on the relationship between exposure to whole-body vibra-tion and low back pain (1986-1997). Int Arch Occup Environ Health,72(6):351–365, Sep 1999.[43] M. Bovenzi. Health risks from occupational exposures to mechanicalvibration. Med Lav, 97(3):535–541, 2006.[44] J. L. Kelsey and R. J. Hardy. Driving of motor vehicles as a riskfactor for acute herniated lumbar intervertebral disc. Am J Epidemiol,102(1):63–73, Jul 1975.[45] H. Seidel, R. Bastek, D. Bruer, C. Buchholz, A. Meister, A. M. Metz,and R. Rothe. On human response to prolonged repeated whole-bodyvibration. Ergonomics, 23(3):191–211, Mar 1980.[46] AGARD Conference Proceedings No. 378. Beevis, D., & Forshaw, S.E.Back pain and discomfort resulting from exposure to vibration in trackedarmoured vehicles, 1985.[47] M. H. Pope, D. G. Wilder, and M. L. Magnusson. A review of studieson seated whole body vibration and low back pain. Proc Inst Mech Eng[H], 213(6):435–446, 1999.[48] N. Kumar Kittusamy and Bryan Buchholz. Whole-body vibration andpostural stress among operators of construction equipment: a literaturereview. J Safety Res, 35(3):255–261, 2004.[49] Etienne Grandjean. Fitting the task to the man: a textbook of occupa-tional ergonomics. Taylor & Francis, 4 edition, 1989.[50] Edward W. Finucane. Definitions, Conversions, and Calculations forOccupational Safety and Health Professionals, Third Edition. CRCPress, April 11, 2006.[51] M. J. Griffin E. M. Whitham. The effects of vibration frequency anddirection on the location of areas of discomfort caused by whole-bodyvibration. Applied Ergonomics, 9:231–239, December 1978.[52] Juha Kiiski, Ari Heinonen, Teppo L Jrvinen, Pekka Kannus, and HarriSievnen. Transmission of vertical whole body vibration to the humanbody. J Bone Miner Res, 23(8):1318–1325, Aug 2008.201.7. Bibliography[53] L. C. Fothergill. A study of subjective response to whole-body vibra-tion. Master’s thesis, University of Southampton, 1972.[54] L. C. Fothergill and M. J. Griffin. The subjective magnitude of whole-body vibration. Ergonomics, 20(5):521–533, Sep 1977.[55] Chikako Kaneko, Takahide Hagiwara, and Setsuo Maeda. Evaluationof whole-body vibration by the category judgment method. Ind Health,43(1):221–232, Jan 2005.212. Wheelchair WheelComparison Study22.1 IntroductionA wheelchair in motion is exposed to infrequent, but high magnitude verti-cal accelerations and most of the energy from these vibrations is thought tobe absorbed by the wheelchair user1. Studies involving able-bodied subjectshave found a variety of health problems that are either caused by or exac-erbated by exposure to whole body vibration (WBV). However, very littleis known about how WBV affects wheelchair users with spinal cord injury(SCI) or how to minimize any adverse effects2;3;4;5;6;7;8.The few studies that have focused on how various elements of the wheelchairsystem influence wheelchair vibration magnitudes and frequencies have lookedat the effect of wheelchair suspension9;10;11;8, cushions12;13 and wheeling sur-face14;15;16. These studies used standardized fatigue tests (the Double-Drumand Curb-Drop tests. See section 1.2 for details), obstacle courses simulat-ing a wheelchair user’s daily activities and “real-world” field tests to inducewheelchair vibrations. The various individual elements of the wheelchairwere all found to influence the wheelchair’s vibration and the user’s expo-sure to WBV. Several of the studies also found the level of WBV, to whichthe user is exposed, exceeds the level of WBV exposure that is defined as safeby the International Standardization Organization17;18;12;14. No study how-ever, has examined the influence of wheelchair wheel design on wheelchairvibration.2A version of this chapter has been published. Vorrink SN, Van der Woude LH, Messen-berg A, Cripton PA, Hughes B, Sawatzky BJ. Comparison of wheelchair wheels in termsof vibration and spasticity in people with spinal cord injury. Journal of RehabilitationResearch & Development. 2008; 45(9):1269-79. PMID: 19319752222.1. IntroductionOnly one study compared different wheelchair wheel designs. This studylooked at the influence of different wheel designs on the wheelchair user’scomfort and wheeling energy efficiency19. However, this study did not studythe wheel designs’ influence on vibration. The study had 20 subjects withSCI wheel over obstacles simulating those encountered by a wheelchair userduring daily activities with wheelchairs equipped with 2 different wheeldesigns - a standard steel spoked wheel and a composite material spoked(Spinergy R©) wheel. The subjects’ heart rate was monitored and a Physi-ological Cost Index (measured in heartbeats per meter) was calculated bysubtracting each subject’s resting heart rate from his/her wheeling heart rateand dividing the result by his/her wheeling speed. A questionnaire, filledout by each subject after completing the wheeling portion of the study, wasused to assess user comfort and wheeling preference.The study found no difference in wheeling energy efficiency between theSpinergy R© and the steel spoked wheels but did find that the Spinergy R©wheels were preferred over steel spoked wheels in terms of overall comfortand suggested that increased comfort may have important implications inmanagement of wheelchair users’ pain and spasticity.A few other studies have looked at how wheelchair vibration affects thewheelchair user’s comfort level20;18;21. However, no study has looked into apossible correlation between wheelchair vibration and spasticity in individ-uals with SCI. Anecdotal reports by manual wheelchair users of lower limbmuscle spasms being triggered by wheelchair vibration suggest a possiblecorrelation between WBV and spasticity.3The potential danger and discomfort caused by extensive exposure to WBVhave motivated the development of better-adapted wheelchairs. Some wheelchairmanufacturers market their products as having superior vibration suppres-sion capabilities and/or spasticity reduction qualities. One such manufac-turer is “Colours N Motion Inc.” (Colours N Motion Inc. USA) which manu-factures the Boing! R© wheelchair. The Boing! R© wheelchair includes a systemsuspension that consists of 4 independent coil springs (one at each wheel)and is marketed as being a more comfortable wheelchair that is perfect forusers who suffer from spasticity:“The innovative suspension system created by Colours Wheelchairs removes3Reports made in private conversations with Dr. Bonita Sawatzy PhD, of the Universityof British Columbia.232.1. Introductionmuch of the impact and vibration from the road making a more comfortableride without adding much weight to the wheelchair. The Boing! is perfectfor people affected by spasticity, pressure sores, pain or who are lookingfor more comfort”22.When comparing wheelchairs with different suspension systems, using loadsensors under the wheelchairs’ seat and backrest to measure forces and ac-celerometers at the wheelchairs’ wheel hub, seat frame and helmet worn bythe user to measure accelerations, Requejo et al. did find that the Boing! R©had the lowest overall forces and head accelerations out of all four wheelchairstested (in all speed conditions)8. Boing! R© users experienced less head mo-tion. Requejo et al. concluded that their findings suggested that the Boing! R©wheelchair exhibits a better overall vibration suppression performance andpotentially better overall ride, particularly for those with a high level ofSCI8.Requejo’s study indicates that the vibration reduction capabilities of theBoing! R© may be superior to other currently available wheelchairs that in-clude suspension systems, at least during some specific tasks. However,Requejo also did not study the effect of wheelchair vibration on spasticity.Another wheelchair wheel manufacturer, Spinergy R© (Spinergy Inc. USA.www.spinergy.com), has taken a wheel design that has been successful inbicycles and adapted it to wheelchairs. Spinergy’s R© wheelchair wheel de-signs feature a triple cavity rim, alloy hub with 1-piece construction, andcarbon fiber spokes which originate from the hub, a feature that Spinergycalls “reverse spoking”.Spinergy’s R© wheelchair wheel spokes are made from strands of a syntheticpolymer — Polyphenylene Bensobisoxazole (PBO), shown in Figure 2.1.PBO is a thermoset polyurethane synthetic polymer, developed by “SRIInternational” (Stanford Research Institute) and trademarked as “Zylon”by “Toyobo” (Toyobo Co. Ltd. Japan. www.toyobo.co.jp). Table 2.1 liststhe elastic modulus (E) and density of PBO and steel fibers.Material Elastic Modulus [GPa] Density [g/cm3]PBO (As Spun) 180 1.54PBO (High Modulus) 270 1.56Steel 200 7.8Table 2.1: Elastic modulus & density of PBO and steel fibers.242.1. IntroductionFigure 2.1: PBO fibers that make up a Spinergy R© wheel spoke. (Imagesource: spinergy.com)Like Colours N Motion, Spinergy R© also claims that its wheels’ vibrationdampening qualities can reduce muscle spasms and fatigue: “Spinergy wheelsare going to provide a much smoother ride. This means less vibration andshock reaches your body, which can reduce muscle spasms and body fa-tigue”.232.1.1 HypothesisThe hypothesis of this study was that composite material spoked wheelchairwheels absorb more vibration than standard steel-spoked wheelchair wheels.It was also hypothesized that the reduced vibration of wheelchairs equippedwith composite material wheels would result in a lower rate of muscle spas-ticity occurrence in wheelchair users with Spinal Cord Injury (SCI) whosuffer from spasticity.2.1.2 ObjectiveThe objective of the wheelchair wheel comparison study was to comparewheelchair wheel designs and the influence of wheelchair vibration on sub-jective estimates of spasticity. The study consisted of 2 parts:1. Daily activities simulation course (DASC) study - SCI subjects2. Controlled speed study - non-SCI subjects:(a) Multiple subject controlled speed (MSCS) study(b) Single subject controlled speed (SSCS) repeatability study252.2. Methods2.2 Methods2.2.1 Subject RecruitmentThirteen SCI subjects (10 males and 3 females) were recruited, throughflyers advertising the study, from the G.F. Strong Rehabilitation Centre(G.F. Strong Rehabilitation Centre, Vancouver, BC, Canada.) and fromthe G.F. Strong Outpatient Spinal Cord Injury Program.To be considered for participation in the study, subjects had to have metthe following criteria:• Between the ages of 16 to 65 years• Spinal cord injury at or below the level of C7• Spasticity of at least Ashworth grade 1 or Spasm frequency scale grade1 for at least one year• Manual wheelchair user with sufficient physical strength to wheel overobstacles typically encountered by manual wheelchair users in activi-ties of daily living• Unmodified wheelchair setup for at least 6 months prior to time ofstudy participation• Cognitive capability to understand and follow basic instructions andto give informed consentSubjects with any history of cardiovascular or pulmonary disease that wouldmake participation unsafe for the subject were excluded from the study.262.2. MethodsGender Age [years] Injury LevelMale 62 T12Male 60 T4-5Male 58 T5-6Male 54 C5Male 52 T3Male 50 T3-4Male 48 C7Male 46 T4Male 39 C6-7Female 38 T5Female 33 T8Female 30 T11Male 30 T5-6Table 2.2: DASC subjects’ baseline data: gender, age and level of spinalcord injury of subjects who participated in the daily activities simulationcourse study.Ethical approval was obtained from the University of British Columbia’sEthical Review Board before beginning the study and every subject receiveda detailed explanation about the study’s purpose, procedure and risks andsigned an informed consent form.2.2.2 Daily Activities Simulation Course DesignThe Daily Activities Simulation Course (DASC), shown in Figure 2.2, con-sisted of 9 obstacles that were designed to simulate obstacles that wheelchairusers encounter while performing daily activities. The DASC was basedon the courses used and validated in several earlier studies attempting tosimulate obstacles encountered by wheelchair users in their daily activi-ties20;1;18;12;13;24;19. The DASC used in this study consisted of the sameobstacles used in these earlier studies but differed in its arrangement as theobstacles of the DASC were arranged as a series of individual parallel lanesand not one continuous loop.The obstacles were positioned parallel to each other and secured in placewith adhesive tape. Two lines on the floor, spaced 5.2 meters apart with272.2. Methodsthe obstacles positioned between them, marked the start and end points ofeach run. The obstacles of the DASC are described in Table 2.3 and shownin Figure 2.3.Figure 2.2: The daily activities simulation course obstacle layoutObstacle Description Dimensions [cm]Rumble Strip13 strips of foam under a Strips (section): 1.5 X 2.5rubber mat. Figure 2.3(a) Mat: 170 (Length)Carpet Figure 2.3(b) 120 (Length) X 1 (Height)Door Threshold Figure 2.3(c) 80 X 1.5 (Height)RampIncline with vertical drop Slope: 5.7 ◦Figure 2.2 Height: 8Dimple Strip Figure 2.3(e) 120 X 1Speed BumpsSinusoidal bump.Low: 2.5 (Height)Medium: 5 (Height)Figure 2.3(f) High: 7.5 (Height)Floor Gym floor (no obstacle) 520 (Length)Table 2.3: Daily activities simulation course obstacle description282.2. Methods(a) Rumble strip (b) Carpet (c) Door threshold(d) Ramp (e) Dimple strip (f) Speed bumpFigure 2.3: The obstacles of the daily activities simulation course2.2.3 InstrumentationWheelchairsWheelchair users have their personal wheelchairs fitted specifically for them.The use of a wheelchair that is not professionally fitted for the user couldalter the user’s wheeling technique and provide less than optimal trunkand leg support. Improper support may cause the user to alter his/herposture which, in addition to sitting on an unfamiliar cushion, could triggerspasticity in a wheelchair user with SCI.It was very important for this study that elements that could act as triggersof spasticity, other than the obstacles of the DASC, be eliminated (or at leastreduced as much as possible). It was therefore decided that the subjectsparticipating in this study would use their own, custom-fitted, wheelchairsbut to remove their wheels and install the study’s test wheels.292.2. MethodsWheelsTwo sets of wheelchair wheels were prepared for this study. One set was apair of standard steel-spoked wheels. The second set was a pair of Spinergy R©composite material (PBO, see subsection 2.1 for details) spoked wheels.All identifying markings were removed from both sets of wheels (logos weremasked with black paint) and the steel spokes were painted black to resemblethe composite material spokes. The only visible difference between the twowheel sets was the different number of spokes. All test wheels (2 for eachwheel set) had their tires pumped to a pressure of 100 psi, as had been donein a previous wheelchair wheel comparison study19.AccelerometersTwo 2-axis, digital accelerometers (Mechsense Digital MDS210U, Mech-works Systems Inc., Ontario, Canada) were used to record vibration datafrom the wheelchairs (Figure 2.4). See Table 2.4 or accelerometer specifica-tions.Figure 2.4: Mechsense Digital MDS 210U accelerometer302.2. MethodsSpecification ValueMeasurement Range ±10 [g]Output Range 0–3.3 [V]Minimum Sensitivity 66 ±17 [mV/g]Alignment Error ±1 [deg]Sampling Rate (per channel) 1–2000 [Hz]Gain 1.00–10.00Dimensions 7.3 x 3.8 x 2.5 [cm]Weight 40 [gr]Table 2.4: Accelerometer (MDS 210U) specificationsA system of previously developed L-brackets and U-bolts was modified to ac-commodate a wide variety of wheelchair axle diameters (Original mountingsystem developed by Bernd Schro¨er, The Technical University of Darmstadt,Darmstadt, Germany). U-bolts of varying widths, along with backplateswith corresponding holes and a groove, shown in Figure 2.5, allowed theflat-faced accelerometer to be securely fastened to the various diameters ofthe cylindrical main axles on the different wheelchair models.Figure 2.5: Main axle accelerometer mounting methodA second L-bracket was used to mount the second accelerometer onto thewheelchairs’ footrest plate. Since footrest plates are fixed at varying angles,the second L-bracket had a pivot hole and guide groove, that enabled thefootrest accelerometer to be horizontally aligned. The bracket was usuallybolted to the subject’s wheelchair. When the subject’s wheelchair footrest312.2. Methodsplate did not allow bolting, the bracket was fixed to the plate using adhesivetape (seen in Figure 2.6). In both cases, the result was a tight connectionbetween the accelerometer and the wheelchair components such that norelative motion at that location was present.Figure 2.6: Footrest accelerometer mounting methodThe main axle was selected as an accelerometer mounting location becauseit is the wheelchair’s structural element that supports the seat and to whichthe wheels attach. Vibration, originating from the wheels, would necessarilybe transmitted to the main axle before continuing through to the rest of thewheelchair and to the user.The footrest was selected as a second accelerometer mounting location be-cause it is an interface region between the wheelchair and the user that isdirectly attached to the front casters. Vibration, originating from the frontcasters, would necessarily be transmitted to the footrest and from there tothe user’s feet, possibly triggering a stretch reflex or spasticity, as clinicalobservations suggest.The wheelchair’s forward direction (direction of travel) was defined as thepositive X direction and labeled “Longitudinal”. The vertical direction (per-pendicular to the floor) was defined as the Y direction and labeled “Perpen-dicular”. The accelerometer mounted on the main axle and the accelerom-eter mounted on the footrest plate could not be oriented in the exact samemanner due to the accelerometers’ cable. In order to avoid bending anddamaging the accelerometers’ cable, each accelerometer was mounted withits cable directed away from the wheelchair-accelerometer interface. The322.2. Methodspositive Y direction on the main axle was defined to be the downward di-rection (toward the floor) and the upward direction (toward the sky) on thefootrest (see Figure 2.7). Vibration in the lateral direction (sideways) wasnot measured.Figure 2.7: Defined directions.Both accelerometers were connected (via cable) to a laptop computer run-ning accelerometer management and data recording software (Mechmanager,Mechworks Systems Inc.).2.2.4 Test PreparationBaseline Data CollectionUpon the subject’s arrival to the test facility, a Modified Ashworth Scale(MAS) assessment was carried out by a physician and baseline data wascollected. The same medical physician carried out the MAS assessmentsfor all the subjects who participated in this study. Clinical baseline dataincluded the subject’s age, gender, diagnosis, co-morbidities and time sinceinjury. Information about the length of time of subject’s wheelchair use,percentage of time spent in the wheelchair, age of current wheelchair andany special wheelchair components was also collected.Subjects also completed a self report questionnaire which included a SpasmFrequency Scale, Spasm Severity Scale, Interference with Function Scale,Painful Spasm Scale and a Visual Analog Scale (VAS). This provided in-formation about the severity of the subject’s spasticity on the day of the332.2. Methodstest.Every subject received a detailed explanation on the test procedure andrisks. The objective of the study was not disclosed to the subject untilafter his/her participation so as to keep the subject as objective as possibleand not influence his/her wheeling technique. The subject then signed aninformed consent form.Accelerometer MountingThe accelerometers were calibrated, using the data-acquisition software’sbuilt in calibration application, before being mounted on the subject’s wheelchair.Calibration of the accelerometers consisted of aligning the axes (±x & ±y)parallel to the direction of gravity and allowing the application to define theperpendicular axes of measurement. A floating-bubble level was used foraligning the accelerometer axes.The accelerometers were then mounted on the subject’s wheelchair (the sub-ject had to transfer out of the wheelchair and on to a bench while the ac-celerometers were being attached). The subject’s personal wheelchair wheelswere replaced with a pair of test wheels. The order of wheel sets tested wasrandomized. The accelerometers were then leveled, using the floating-bubblelevel to ensure that the vertical axis is aligned perpendicularly to the ground,and secured. The accelerometer data cables were run backwards (beneaththe wheelchair), taped to the wheelchair frame so that they would not gettangled in the wheelchair wheels or otherwise interfere with the subject’swheeling, and connected to the laptop computer.Subject WeighingWeighing a person with SCI is not a trivial task as the person is unableto stand on a scale. Therefore, the weight of the subjects could not bemeasured directly and instead had to be calculated. This was achieved byfirst weighing the wheelchair without the subject. This was done while thesubject was seated on a bench, waiting for the accelerometers to be mountedon his/her wheelchair. The subject then transferred back into the wheelchairand the wheelchair and subject were weighed again together, resulting in thecombined weight (Wc+s). The weight of the subject (Ws) was calculated by342.2. Methodssubtracting the weight of the wheelchair (Wc) from the combined weight ofthe wheelchair and subject. The calculation is shown in Equation 2.1:Ws = Wc+s −Wc (2.1)2.2.5 Test ProcedureThe order of obstacles which the the subject wheeled over was randomized.The subject aligned his/her front caster wheels with the start line and waitedfor a verbal “start” command. The subject then wheeled at his/her naturalpace over the obstacle and continued wheeling until reaching the end line,a distance of 5.20 meters from the start line. This was considered a single“run”.Data RecordingAs the subject wheeled through the DASC, acceleration data from bothaccelerometers was recorded on a laptop computer (connected via cable tothe accelerometers) which was held by a researcher who followed the sub-ject through the course (Figure 2.8). Acceleration data was recorded usingMechworks Systems Inc’s MechManager software. The acceleration datawas collected at a sampling rate of 1000 Hz. The subject’s time to completeeach run was also recorded.352.3. Controlled Speed Wheel Comparison StudiesFigure 2.8: Obstacle course data recordingAfter completing each run, the subject rated his/her comfort level and theseverity of his/her spasticity at that particular moment on a VAS. Thiswas repeated for all 9 obstacles. After completing the 9th run, the subjectcompleted 5 Visual Analog Scales. The wheelchair wheels were then replacedwith the second wheel set and the 9 obstacle runs, in a randomized order,were completed again. Acceleration data and Visual Analog Scales wererecorded again.After completing all 18 runs (9 obstacles × 2 wheel sets), the subject’s ownwheels were reattached to his/her wheelchair (the subject had to transferout of the wheelchair while this was taking place) and the accelerometerswere unmounted. This concluded the study. Twelve subjects completed all9 obstacle runs. One subject did not feel comfortable with the ramp obstacleand therefore did not wheel over the ramp.2.3 Controlled Speed Wheel Comparison StudiesTo limit the variations introduced by different wheelchair systems, as usedin the DASC study, a single wheelchair was used in this study. Since theuse of a wheelchair that has not been specifically fitted for the user maycause discomfort and/or trigger spasms in users with SCI and due to thegreater acceleration expected in this study, it was decided that only healthynon-SCI individuals be accepted as subjects.362.3. Controlled Speed Wheel Comparison Studies2.3.1 Subject RecruitmentTwenty two able bodied volunteers were recruited at the GF Strong Reha-bilitation Centre (The subjects were mostly visitors and employees of therehabilitation center). None of the subjects had Spinal Cord Injury or anyother illness that required the use of a wheelchair. None of the subjects hadany previous wheelchair use experience.2.3.2 Study DesignTo limit the variations of speed at which the wheelchair riders negotiatedthe obstacles when wheeling through the Daily Activities Simulation Course(DASC), an acceleration platform (ramp) was designed. The purpose of theplatform was to allow gravity only to accelerate the wheelchair and rider,rather than the wheelchair user’s manual propulsion, thus maintaining arelatively constant speed for all runs.A study of mobility characteristics and activity levels of manual wheelchairusers, by Tolerico et al. , mounted data loggers on the wheelchairs of 52subjects and monitored their activity, at home and at a gym, over a pe-riod of 13 or 20 days25. The study found the average wheeling velocity ofmanual wheelchair users to be 0.96 ±0.17 m/s2 at the gym and 0.79 ±0.19m/s2 at home. The acceleration platform was designed to produce wheelingvelocities that were consistent with Tolerico’s findings.The acceleration platform consisted of a surface at an incline of 9◦. Twolines on the platform’s surface marked two different starting positions fromwhich the wheelchair would begin to accelerate before transitioning to thefloor and rolling over a vibration inducing obstacle. The vibration inducingobstacle consisted of a sinusoidal bump. The bump was 25 mm high with abase that was 80 mm wide. The obstacle was fixed to the floor and positionedso that the two starting lines were at a distance of 1.65 m and 2.00 m fromit (Figure 2.9). The platform included elevated edges, designed to preventthe wheelchair from rolling off the side of the platform.372.3. Controlled Speed Wheel Comparison StudiesFigure 2.9: Acceleration platform setup2.3.3 InstrumentationOne Invacare A4 wheelchair (Invacare Canada L.P.,Mississauga, Ontario,Canada) was used for this study. The same 2 Mechworks Mechsense Digi-tal MDS210U accelerometers used in the DASC study were used to recordvibration data from the wheelchair (See Table 2.4 for accelerometer specifi-cations).The same 2 sets of wheelchair wheels that were used in the DASC study(a pair of standard steel-spoked wheels and a pair of Spinergy R© compositematerial spoked wheels) were used for this study. The only visible differ-ence between the two wheel sets was the different number of spokes as allidentifying markings had been removed from both sets of wheels (logos weremasked with black paint and the steel spokes were painted black to resemblethe composite material spokes). All test wheels (2 for each wheel set) hadtheir tires pumped to a pressure of 100 psi (the same tire pressure as in theDASC study).The main axle and footrest were again selected as accelerometer mountinglocations for the same reasons as described for the DASC study. Vibration,originating from the wheels, would be transmitted to the wheelchair and the382.3. Controlled Speed Wheel Comparison Studiesuser through the main axle. Vibration, originating from the front casters,would be transmitted to the user’s lower limbs through the footrest, possiblytriggering a stretch reflex or spasticity.The accelerometers were calibrated, using the same method as described insubsection 2.2.4 (page 34), before being mounted on the test wheelchair. Theaccelerometers’ axes were aligned in the same orientation as in the DASCstudy (see subsection 2.2.3). Vibration in the lateral direction (sideways)was again not measured. Both accelerometers were connected (via cable) toa laptop computer running accelerometer management and data recordingsoftware (Mechmanager, Mechworks Systems Inc.).2.3.4 Test PreparationBaseline Data CollectionEvery subject received a detailed explanation on the test procedure andrisks. The subject then signed an informed consent form. The subject’s ageand weight were recorded. The average participant weight was 71.5 kg (±11.5 kg SD).2.3.5 Test ProcedureMultiple Subject Controlled Speed Wheel Comparison StudyOne set of test wheels was installed on the test wheelchair. The subjectsat in the test wheelchair and was asked to remain passive and relaxed.The wheelchair was then positioned on the acceleration platform with thefront casters aligned with the lower starting line and held in place by aresearcher. The wheelchair was then released and allowed to acceleratedown the platform with the aid of gravity only. The wheelchair rolled on tothe floor and over the vibration inducing obstacle.This process was then repeated with the wheelchair positioned with its frontcasters aligned with the higher starting line. The wheelchair wheels werethen replaced with the second set of test wheels and the process was re-peated. Every subject rolled down the acceleration platform and over the392.4. Data Analysisobstacle 4 times (from both starting lines with both sets of wheels).Single Subject Controlled Speed Wheel Comparison StudyThe controlled speed wheel comparison study procedure was also conductedwith a single subject. The purpose of the single subject controlled speedstudy (SSCS) was to test the repeatability of the controlled speed wheel com-parison method. The single subject completed 16 runs, using a “KuschallChampion” wheelchair, with each set of wheels at each velocity, for a totalof 64 runs. This was done in order to minimize, as much as possible, thevariability due to the subjects’ physical characteristics, posture and bracing(tensing of the body in anticipation for the impact) as a single subject couldmaintain a consistent seating posture and impact bracing, throughout allruns, more reliably than several different subjects.Data RecordingAs the wheelchair rolled down the platform and over the obstacle, accelera-tion data from both accelerometers was recorded on a laptop computer (con-nected via cable to the accelerometers) which was held by a researcher whofollowed the wheelchair. Acceleration data was recorded using accelerom-eter management and data recording software (MechManager, MechworksSystems Inc.). The acceleration data was collected at a sampling frequencyof 1000 Hz for the MSCS and 2000 Hz for the SSCS. The subject’s time tocomplete each run was also recorded.2.4 Data Analysis2.4.1 Acceleration AnalysisThe recorded acceleration signal from each run (while the wheelchair was inmotion, but before the wheelchair encountered the obstacle) was averagedand the resulting value was subtracted from the entire signal. This was donein an effort to reduce any baseline “noise” and\or “DC offset”.402.4. Data AnalysisA “DC offset” is a constant value shift and occurs when a signal’s baselinevalue does not equal zero. If the mean of a signal’s values equal’s a valueother than zero, then there is a “DC offset” and every sample point valuehas been shifted by that mean value. The acceleration signal in the Y-axisdirection, at both the main axle and the footrests, always had a “DC offset”value of 1 g due to gravity.This method of “noise” removal differs from the method used in the originaldata analysis method (described in the published version of this chapter.26)where one separate measurement was recorded from each subject while thewheelchair was not in motion and the mean value from this single mea-surement was subtracted from every acceleration signal recorded with thatsubject.The justification for this method is that it was assumed that the ambientvibration “noise” and any DC-offset values may vary throughout the testingday and even from one obstacle run to the next. It was also thought thatby subtracting the mean value of the signal while the wheelchair was inmotion would reduce not only ambient vibration “noise” and any DC-offsetbut also vibration due to the ride surface and the subject’s propulsion of thewheelchair, thus isolating the vibration due to impact with the obstacle.The peak acceleration and Root Mean Square (RMS) values were then cal-culated, using Matlab (Matlab version 7. The MathWorks Inc., Natick, MA,USA). Typical vibration recordings are shown in Figure 2.10.412.4. Data Analysis-2.5-2-1.5-1-0.500.50 1 2 3 4 5Acceleration [g] Time [s] Rumble Strip (a) Rumble strip-2.5-2-1.5-1-0.500.50 1 2 3 4 5Acceleration [g] Time [s] Carpet  (b) Carpet-2.5-2-1.5-1-0.500.50 1 2 3 4 5 6Acceleration [g] Time [s] Door Threshold  (c) Door threshold-2.5-2-1.5-1-0.500.50 1 2 3 4 5 6 7 8 9 10Acceleration [g] Time [s] Ramp (d) Ramp-2.5-2-1.5-1-0.500.50 1 2 3 4 5Acceleration [g] Time [s] Dimple Strip  (e) Dimple strip-2.5-2-1.5-1-0.500.50 1 2 3 4 5 6Acceleration [g] Time [s] Low Sinusoidal Bump  (f) Low speed bump-2.5-2-1.5-1-0.500.50 1 2 3 4 5 6 7 8Acceleration [g] Time [s] Medium Sinusoidal Bump  (g) Medium speed bump-2.5-2-1.5-1-0.500.50 1 2 3 4 5 6 7Acceleration [g] Time [s] High Sinusoidal Bump  (h) High speed bump-2.5-2-1.5-1-0.500.50 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5Acceleration [g] Time [s] Floor  (i) FloorFigure 2.10: Typical vibration recordingsRoot Mean SquareRMS is a measure of the magnitude of a varying quantity and is especiallyuseful when the values vary between positive and negative. RMS is anaverage of the square root of the sum of squared values. For a continuous(known) function, the RMS value can be calculated using Equation 2.2.fRMS =√1T∫ T0f2(t)dt (2.2)422.4. Data AnalysisSince the recorded acceleration data was a series of discrete sampled points,the RMS value was calculated using Equation 2.3.xRMS =√√√√ 1nn∑i=1x2i (2.3)The RMS values for each obstacle were compared between the two types ofwheelchair wheels used during the run (composite material spoked and steelspoked) and between the vibration measurement location (main axle andfootrest).2.4.2 Frequency Power Spectrum AnalysisVibration signals can differ from each other not only in magnitude but alsoin frequency. A single vibration signal, which may appear random andunintelligible when viewed in the time domain, may in fact be composed ofmultiple single-frequency signals at varying magnitudes. Revealing a signal’sfrequency content requires a transformation of the signal from the timedomain to the frequency domain. This transformation can be achieved byapplying a mathematical process known as a Fourier Analysis.The Fourier Analysis & Fast Fourier TransformsA Fourier analysis is the decomposition of a function into a sum of sinusoidalfunctions, by determining the amplitude and phase of each component, thatcan be recombined to obtain the original function. The Fourier Transformis defined in Equation 2.4.X(ω) =∫ ∞−∞x(t)e−iωt dt (2.4)A signal can be either continuous or discrete, and it can be either periodicor aperiodic. Combining these options generates four possible signal cate-gories: Aperiodic-Continuous, Periodic-Continuous, Aperiodic-Discrete andPeriodic-Discrete. These four signal categories extend from −∞ to∞. Since432.4. Data AnalysisSine and Cosine waves extend from −∞ to ∞, and infinitely long signalscannot be used to synthesize a signal of finite length, the Fourier Transformcannot handle signals of finite length27.By artificially inserting an infinite number of zero value sample points, beforeand after the signal, an Aperiodic-Discrete signal is created. This methodis called “Zero Padding” and allows for the application of the Discrete TimeFourier Transform (Equation 2.5).X(ω) =∞∑n=−∞x(n)e−iωn (2.5)Another method is infinitely duplicating and concatenating the duplicatesto the left and right of the signal. This creates a Periodic-Discrete signaland allows for the application of the Discrete Fourier Transform (Equation2.6).Xk =N−1∑n=0xn e−2piiN kn =N−1∑n=0xn e−iωn (2.6)k = 0, ..., N − 1Xk represent the amplitude and phase of the sinusoidal components of theinput signal (xn).An FFT is an algorithm that allows efficient computation of the DiscreteFourier Transform (DFT), and its inverse.BinningAn FFT was applied to the time-domain vibration signals, resulting infrequency-domain data. A typical time-domain to frequency domain con-version is shown in Figure 2.11.442.4. Data AnalysisFigure 2.11: Typical Time-Domain to Frequency-Domain ConversionThe Frequency Power Spectrum (FPS) of each signal (range 0–500 Hz) wasthen divided into frequency octaves for human vibration exposure, usingequation 2.7, as defined by ISO 2631-1 and used in a previous study byCooper et al.10:f2 = 223 × f1 (2.7)In equation 2.7, f2 represents the highest frequency in the FPS. Dividing f2by 223 gives f1 which is now the bottom frequency of the last octave. Thiscalculation is repeated, with f1 now becoming the top frequency (f2) of thenext (lower) octave, until f1 is equal to or lower than 1.25 Hz.Equation 2.7 produces frequency octaves of varying width, shown in Table2.5. The octaves become wider as one goes up in frequency. This resultsin greater resolution at the lower frequencies, to which the human body ismore sensitive. Table 2.5 lists the octaves and their corresponding frequencyranges.452.4. Data AnalysisOctave Frequency Range [Hz]1 0 – 1.232 1.23 – 1.953 1.95 – 2.464 2.46 – 3.105 3.10 – 3.916 3.91 – 4.927 4.92 – 6.208 6.20 – 7.819 7.81 – 9.8410 9.84 – 12.4011 12.40 – 15.6312 15.63 – 19.6913 19.69 – 24.8014 24.80 – 31.2515 31.25 – 39.3716 39.37 – 49.6117 49.61 – 62.5018 62.50 – 78.7519 78.75 – 99.2120 99.21 – 125.0021 125.00 – 157.4922 157.49 – 198.4323 198.43 – 250.0024 250.00 – 314.9825 314.98 – 396.8526 396.85 – 500.00Table 2.5: Octave frequency rangesFigure 2.12 shows a typical FPS. The FPS of each signal was also dividedinto bins of equal frequency width in order to create data sets of an equalnumber of data points for analysis.Since perception of vibration at frequencies below 1 Hz and above 20 Hzrequires relatively higher magnitudes of vibration3 and the average octavewidth between 1 and 20 Hz is approximately 2 Hz, a frequency bandwidthof 2 Hz per bin was used (each bin was 2 Hz wide). The frequency powerspectrum of the SSCS, however, was sectioned into bins of 10 Hz width.462.4. Data Analysis00.20.40.60.811.21.41.60 50 100 150 200 250 300 350 400 450 500Power Frequency [Hz]  Frequency Power Spectrum  Figure 2.12: Typical frequency power spectrumIt was clear from all computed FPS’s that the power at frequencies above60 Hz is negligible in comparison with the 0–60 Hz range. It was thereforedecided to analyze only the 0–60 Hz range of the FPS’s.The peak power value within each octave/bin was taken as the measure ofpeak power per octave/bin. The area under the FPS curve represents thetotal vibration power10. The total vibration power within each octave/binwas obtained by integrating the FPS curve over the octave/bin width. Theresulting values were taken as the measure of total vibration power peroctave/bin.2.4.3 Statistical AnalysisAnalyses of variance (two-way repeated measures) were conducted on thepeak acceleration values, peak power per octave, peak power per bin, totalpower per octave and total power per bin (dependent variables). For theDASC data, wheel type and obstacle were set as the independent variables.For the MSCS data, wheel type and velocity were set as the independentvariables. The α level for all statistic significance tests was set to 0.05. Asthe SSCS was a repeatability study, no analysis for statistical significancewas conducted on the SSCS data.472.5. Results2.5 ResultsAs the DASC and Controlled Speed Study data has been previously analyzed(excluding the MSCS study data) and the results published26, only a briefsummary of the DASC and Controlled Speed Study results is presentedhere. Additionally, since vertical vibration is the most common vibrationtransmitted from the wheelchair to the user and vertical vibrations seemto be the main cause of complaints by wheelchair users21), only results ofanalyses in the Y axis direction are presented here.2.5.1 Daily Activities Simulation Course ResultsSubjective Spasticity & Comfort ResultsNo significant differences were detected between Spinergy R© and steel spokedwheels, in any of the obstacles, for subjective spasticity, comfort or overallassessment. Figure 2.13 shows the subjective spasticity visual analog scaleresults.0123456Rumbl e Str i p Ca r pet DoorTh r eshold Ram p Di m pl e Str i p B um p (S) B um p (M) B um p (L) FloorVAS Spasticity Score Obstacle  DASC Subjective Spasticity  Spi ner gySteelFigure 2.13: DASC subjective spasticity visual analog scale results.482.5. ResultsAcceleration ResultsNo statistically significant differences were detected between Spinergy R© andsteel spoked wheels for peak acceleration (Figure 2.14(a)) or RMS accelera-tion (Figure 2.14(b)) values.024681012Rumbl e Str i p Ca r pet DoorTh r eshold Ram p Di m pl e Str i p B um p (S) B um p (M) B um p (L) FloorAcceleration [m/s2 ] Obstacle  DASC Peak Acceleration (Y)  Spi ner gySteel(a) Peak acceleration00.20.40.60.811.21.41.61.82Rumbl e Str i p Ca r pet DoorTh r eshold Ram p Di m pl e Str i p B um p (S) B um p (M) B um p (L) FloorAcceleration [m/s2 ] Obstacle  DASC Acceleration RMS (Y)  Spi ner gySteel(b) Acceleration RMSFigure 2.14: DASC acceleration results.Frequency Power Spectrum Analysis ResultsNo significant differences in peak power per octave or total power per octavewere found. Table 2.6 lists the statistically significant differences found inthe peak power per bin and total power per bin frequency analysis results.Figures 2.15(a) and 2.15(b) show the peak power and total power per binresults.492.5. ResultsPeak Power per BINFrequency [Hz]Mean ValueP-valueInteractionSpinergy Steel (wheel-obstacle)52–54 0.103 0.074 0.044 No54–56 0.065 0.050 0.023 No58–60 0.032 0.027 0.035 NoTotal Power per BINFrequency [Hz]Mean ValueP-valueInteractionSpinergy Steel (wheel-obstacle)58-60 0.028 0.021 0.030 NoTable 2.6: Statistically significant differences in frequency content of vibra-tion during the DASC study, between Spinergy R© and steel spoked wheels.00.511.522.533.51 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30Frequency Power Frequency Bin DASC Peak Power per Bin (Y)  Spi ner gySteel(a) Peak power00.511.522.533.51 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30Frequency Power Frequency Bin DASC Total Power per Bin (Y)  Spi ner gySteel(b) Total powerFigure 2.15: DASC peak and total power per bin: (a) Peak power. (b)total power. Each bin number represents a 2 Hz wide frequency range. Bin1: 0–2 Hz, Bin 2: 2–4 Hz . . . Bin 30: 58–60 Hz.502.5. Results2.5.2 Multiple Subject Controlled Speed Study ResultsTable 2.7 lists the average wheelchair velocities achieved, from both rampheights, with each set of wheels.VelocityWheeltypeActual Standard RelativeCategory Velocity Deviation Std. Dev.FastSpinergy 0.92 m/s 0.04 m/s 4%Steel 0.93 m/s 0.04 m/s 4%SlowSpinergy 0.83 m/s 0.04 m/s 5%Steel 0.85 m/s 0.05 m/s 5%Table 2.7: MSCS average wheelchair velocities.No statistically significant differences were detected between Spinergy R© andsteel spoked wheels for peak acceleration, RMS acceleration values, peakpower per bin/octave or total power per bin/octave.2.5.3 Single Subject Controlled Speed Study ResultsAs the single subject controlled speed (SSCS) study was designed to testthe repeatability of the multiple subject controlled speed study (MSCS),the results of interest were the standard deviation and the relative standarddeviation. The frequency power spectrum was sectioned into bins of 10Hz width and no analysis for statistical significance was conducted on theresults. Table 2.8 lists the average wheelchair velocities achieved, from bothramp heights, with each set of wheels.VelocityWheeltypeActual Standard RelativeCategory Velocity Deviation Std. Dev.FastSpinergy 0.86 m/s 0.05 m/s 5%Steel 0.88 m/s 0.02 m/s 3%SlowSpinergy 0.80 m/s 0.02 m/s 3%Steel 0.79 m/s 0.03 m/s 4%Table 2.8: SSCS average wheelchair velocities.512.5. ResultsAcceleration ResultsTable 2.9 lists the SSCS mean acceleration peak & RMS results. Figure 2.16displays the results.Peak AccelerationVelocityWheeltype MeanStandard RelativeCategory Deviation Std. Dev.FastSpinergy 2.76 m/s2 0.97 m/s2 35%Steel 3.00 m/s2 0.86 m/s2 29%SlowSpinergy 1.49 m/s2 0.23 m/s2 15%Steel 1.54 m/s2 0.17 m/s2 11%Acceleration RMSVelocity Wheeltype MeanStandard RelativeDeviation Std. Dev.FastSpinergy 0.40 m/s2 0.07 m/s2 17%Steel 0.41 m/s2 0.05 m/s2 13%SlowSpinergy 0.26 m/s2 0.025 m/s2 10%Steel 0.26 m/s2 0.016 m/s2 6%Table 2.9: SSCS mean acceleration peak and RMS results.00.511.522.533.544.5Fast SlowAcceleration [m/s2 ] Velocity  SSCS Peak Acceleration (Y) Spi ner gySteel(a) Peak acceleration00.050.10.150.20.250.30.350.40.450.5Fast SlowAcceleration [m/s2 ] Velocity  SSCS Acceleration RMS (Y)  Spi ner gySteel(b) Acceleration RMSFigure 2.16: SSCS mean acceleration peak (a) and RMS (b) results.522.5. ResultsFrequency Content Analysis ResultsTables 2.10 and 2.11 list the vibration frequency power spectrum (FPS)mean, standard deviation and relative standard deviation for SSCS peakpower per bin and total power per bin results (respectively). Figures 2.17and 2.18 display the vibration FPS peak power per bin and total power perbin mean results (respectively).Peak Power Per BinBin VelocityWheeltype MeanStandard RelativeRange [Hz] Category Deviation Std. Dev.0–10FastSpinergy 4.30 0.59 14%Steel 4.85 0.43 9%SlowSpinergy 3.03 0.35 11%Steel 3.29 0.39 12%10–20FastSpinergy 1.84 0.66 36%Steel 2.09 0.97 47%SlowSpinergy 1.14 0.48 42%Steel 1.27 0.70 55%20–30FastSpinergy 3.67 1.20 33%Steel 2.61 1.15 44%SlowSpinergy 2.05 0.82 40%Steel 1.71 0.69 40%30–40FastSpinergy 4.41 2.03 46%Steel 4.29 1.02 24%SlowSpinergy 2.26 0.57 25%Steel 1.97 0.76 39%40–50FastSpinergy 2.50 1.16 46%Steel 2.29 1.02 44%SlowSpinergy 0.98 0.48 49%Steel 1.11 0.34 30%50–60FastSpinergy 0.30 0.18 58%Steel 0.37 0.19 51%SlowSpinergy 0.29 0.16 55%Steel 0.26 0.07 26%Table 2.10: SSCS vibration FPS peak power mean, standard deviationand relative standard deviation results.532.5. ResultsTotal Power Per BinBin VelocityWheeltype MeanStandard RelativeRange [Hz] Category Deviation Std. Dev.0–10FastSpinergy 12.31 2.62 21%Steel 15.57 3.06 20%SlowSpinergy 7.81 1.85 24%Steel 8.88 2.71 31%10–20FastSpinergy 6.43 3.74 58%Steel 7.89 3.72 47%SlowSpinergy 3.59 1.87 52%Steel 3.79 1.82 48%20–30FastSpinergy 10.78 4.76 44%Steel 9.75 5.30 54%SlowSpinergy 5.13 1.91 37%Steel 4.76 1.36 29%30–40FastSpinergy 14.24 6.92 49%Steel 13.76 7.00 51%SlowSpinergy 6.16 1.58 26%Steel 5.50 1.61 29%40–50FastSpinergy 4.78 1.90 40%Steel 6.08 2.68 44%SlowSpinergy 2.38 1.17 49%Steel 2.41 0.76 32%50–60FastSpinergy 0.63 0.19 30%Steel 0.83 0.33 40%SlowSpinergy 0.70 0.28 41%Steel 0.65 0.18 28%Table 2.11: SSCS vibration FPS total power mean, standard deviationand relative standard deviation results.542.5. Results01234567891 2 3 4 5 6Frequency Power Frequency Bin SSCS Peak Power per Bin (Y)  Fast  Spi ner gyFast  SteelSlow Spi ner gySlow St eelFigure 2.17: Vibration frequency content mean peak power per 10 Hz bin.05101520251 2 3 4 5 6Frequency Power Frequency Bin  SSCS Total Power per Bin (Y)  Fast  Spi ner gyFast  SteelSlow Spi ner gySlow St eelFigure 2.18: Vibration frequency content mean total power per 10 Hz bin.552.6. Discussion2.6 DiscussionThe daily activities simulation course (DASC) and multiple subject con-trolled speed (MSCS) study results revealed no statistically significant ac-celeration differences between the two tested wheel designs. There was noclear trend in the DASC acceleration results as some obstacles resulted inlarger accelerations with Spinergy R© wheels while others produced largeraccelerations with steel spoked wheels. It is interesting that the mediumsinusoidal bump resulted in larger accelerations than the large sinusoidalbump, for both wheel types. This is most likely due to the subjects adapt-ing their wheeling speed and style when traversing the different obstacles.The subjects probably reduced their wheeling speed before contacting thelarge sinusoidal bump significantly more than they did before contacting themedium bump. The subjects may have also maintained a more controlleddescent from the large sinusoidal bump than they did from the mediumbump.The subjective spasticity visual analog scale (VAS) results also revealedno significant differences between the two tested wheels but a trend wasapparent. The subjects perceived more spasticity with the steel spokedwheels than with the Spinergy R© wheels from all but one obstacle. The onlyobstacle from which subjects perceived more spasticity with the Spinergy R©wheels was the rumble-strip. This is particularly interesting as the rumble-strip, which produced relatively low levels of acceleration, resulted in a levelof perceived spasticity that was only exceeded by the level of perceivedspasticity from the medium and large sinusoidal bumps. This could bedue to the fact that the rumble-strip produced repeated low level vibrationwhile the other DASC obstacles produced a single shock (not including thedimple-strip, carpet and floor which seem to produce insignificant levels ofvibration).The DASC frequency power spectrum (FPS) results revealed statisticallysignificant differences in the 52–56 Hz and 58–60 Hz ranges. These findingsdiffer from the results of the original analysis (which detected no differ-ences). This is odd as the original analysis utilized matched pair t-tests (onthe DASC data) which, when applied without a correction for dependentvariables, should (theoretically) detect differences that an analysis of vari-ance would not detect. It is possible that the different methods of noisereduction between the original analysis and current analysis resulted in dif-562.6. Discussionferences being detected by one method of analysis and not the other.The original method of noise reduction included a single measurement calledthe “zero-measurement” and considered to represent the ambient vibration“noise”. During the DASC study, before each subject began his/her partic-ipation in the study, a measurement was recorded from both accelerometerswhile the wheelchair was not in motion. During the MSCS study, thismeasurement was taken only once at the start of the testing day. The zero-measurement’s mean value was calculated and this one mean value (persubject) was then subtracted from every subject’s acceleration data in anattempt to remove any ambient vibration “noise” and DC-offset. In thedata re-analysis method, the “zero-measurement” data was not used. In-stead, the mean value of every acceleration data measurement, while thewheelchair was in motion but before impacting the obstacle, was calculatedfor each signal recorded. each of these mean values was then subtractedfrom the acceleration signal from which it was calculated.In any case, the power of vibration in the 52–60 Hz range is very low incomparison with the vibration power at lower frequencies. So while somedifferences were detected as statistically significant, these differences aremost likely physically insignificant. Because of the relatively low powerand physical insignificance of vibration in the 52–60 Hz frequency range, noadditional analysis for statistical differences was conducted on the data.The wheeling velocities achieved in the MSCS are consistent with averagewheeling velocities of manual wheelchair users and the relative standarddeviations in velocities (4% for the “Fast” velocity and 5% for the “Slow”velocity) were quite small. The SSCS wheeling velocities were slightly lowerthan those in the MSCSS and the relative standard deviations in velocitieswere also slightly lower (except for “Fast-Spinergy”). These velocity resultssuggest that the method used for controlling the wheelchair’s velocity duringthe MSCS wheel comparison study was effective in maintaining consistentvelocities that were similar to those experienced by manual wheelchair users.The slight difference in velocities between the MSCS and SSCS could be dueto the use of different wheelchair models for the MSCS and SSCS studies (an“Invacare A4” wheelchair was used for the MSCS and a “Kuschall Cham-pion” wheelchair was used for the SSCS).The DASC results had relative standard deviations that were extremely high— as high as 61% for peak acceleration and 247% for peak power per bin.572.6. DiscussionLarge standard deviations in the DASC results are to be expected as manydifferent wheelchair systems were used and the wheeling speeds and obsta-cle negotiation techniques were uncontrolled. The controlled speed wheelcomparison study successfully controlled wheeling speed and used only onewheelchair system but the standard deviations were, surprisingly, still ex-tremely large — as high as 47% for peak acceleration and 167% for peakpower per bin. The standard deviations in the SSCS results are also dis-appointing. Even with a single subject, maintaining a constant posture (asaccurately as humanly possible), participating in all test runs and wheelingat speeds that were shown to be consistent, the relative standard deviationswere as high as 35% for peak acceleration results and 58% for peak & totalpower per FPS bin.It is possible that differences in acceleration magnitudes and frequency powerbetween Spinergy R© wheels and standard steel spoked wheels may actuallyexist and that these differences were obscured by the large standard devi-ations. The deviations in acceleration magnitudes and vibration frequencypower values that were present in the study results are undoubtedly presentin “real world” situations. It is highly unlikely that if no significant differ-ences were detected between the two wheel designs in either of the wheelcomparison studies, that manual wheelchair users would be able to detectany differences, even if they do exist. Therefore, it would be logical to con-clude that Spinergy R© wheels do not differ from standard steel spoked wheelsin terms of vibration acceleration magnitudes and frequency power.The apparent trend in the subjective spasticity VAS results which showsless perceived spasticity with the Spinergy R© wheels, coupled with resultsfrom a Hughes et al. ’s 2005 study which found that Spinergy R© wheels areperceived as more comfortable than standard steel spoked wheels, suggeststhat while no objective mechanical differences could be identified betweenthe two tested wheels, some other difference between the two wheel designsmay still exist.582.6. Discussion2.6.1 LimitationsDaily Activities Simulation CourseDuring the daily activities simulation course (DASC) study, wheeling speedwas not controlled and every subject wheeled through the course at his/herown pace. Negotiation of the individual obstacles of the course was also un-controlled, allowing subjects to use their own style when encountering theobstacles. Some subjects would pause while on the obstacles and controltheir descent while others wheeled right over the obstacles without hesita-tion. Some subjects would raise their front casters before contacting theobstacles while others did not.Allowing each subject to use his/her own wheelchair during the study in-troduced many uncontrolled variables as each wheelchair was structurallydifferent and included different suspension systems, cushions and backrests -all of which have been shown to influence wheelchair vibration and its trans-missibility to the user. The fact that each subject used the same wheelchair(his/her own) while testing both sets of wheels probably reduced the overalleffect of the use of many different wheelchairs for this study but it would havebeen better to have used a single wheelchair for all participating subjects.The DASC obstacles have been shown to accurately represent the obstaclesthat a wheelchair user encounters during daily activities however there wasno way to control the acceleration magnitudes and frequency range to whichthe subjects were exposed. Additionally, the visual analog scale methodof assessing spasticity is highly subjective. Therefore, definite conclusionsregarding a possible correlation between wheelchair vibration (frequencies& acceleration magnitudes) and spasticity cannot be made with confidence.Controlled Speed Wheel ComparisonThe controlled speed wheel comparison eliminated the variability betweenwheelchairs by using a single wheelchair throughout the study. The speed ofwheeling, along with the method of contacting and traversing the obstaclehad been successfully controlled. Despite this, very large standard devia-tions in the results were present. Even the single subject controlled speedstudy results exhibited large standard deviations. These standard devia-592.7. Conclusiontions greatly reduced the ability to draw conclusions based on the statisticalanalyses of the results.2.7 ConclusionBoth wheelchair wheel comparison studies failed to detect any significantdifferences in peak acceleration between the two wheel designs. The indi-viduals with SCI who participated in the study, and use manual wheelchairson a regular basis, also found no differences in comfort or subjective sen-sation of spasticity when using the two wheel designs. Any differences invibration frequency between the two wheel designs are practically insignif-icant as they are either in frequency ranges where the power of vibrationis very low or can only be detected under conditions that would never beencountered by manual wheelchair users.Even if differences that were undetected during the wheelchair wheel com-parison studies do exist between the two wheel designs, these differencesare likely to be physically insignificant. A wheelchair user would probablybe unable feel any difference between the wheel designs under “real-world”conditions and therefore not benefit from the use of one wheel design overthe other.Considering these results and the results of a previous study19 which foundno difference in wheeling efficiency between the same two wheel designsthat were compared in this study, it seems pretty clear that the compositematerial spoked wheel design (Spinergy R©) offers no mechanical advantageover the standard steel spoked wheel design.Whether or not any alternative wheel design can significantly influencewheelchair vibration cannot be concluded from these wheel comparison stud-ies as only two wheel designs were compared and only conclusions aboutthose two designs can be made. It can however be concluded, that in orderfor a wheel design to have a different effect on wheelchair vibration thanthe effect of standard steel spoked wheels, the design would have to be moreradically different than having composite material spokes instead of steelspokes.602.7. Conclusion2.7.1 Recommendations for Future WorkFuture studies, examining the influence of wheel design on wheelchair vi-bration, should study additional wheel designs. It is not recommended thatSpinergy R© wheels be tested again for influence on wheelchair vibration, un-less the Spinergy R© company introduces another wheelchair wheel that isfundamentally different, either in the wheel design or the materials it ismade of, than the wheel design tested in this study. Studies on wheelchairvibration as a trigger of muscle spasticity should incorporate an objectivemethod of measuring spasticity, such as surface EMG, and not rely solelyon the highly subjective self report questionnaires.612.8. Bibliography2.8 Bibliography[1] D. P. VanSickle, R. A. Cooper, and M. L. Boninger. Road loads actingon manual wheelchairs. IEEE Transactions on Rehabilitation Engineer-ing, 8(3):371–384, Sept. 2000.[2] Etienne Grandjean. Fitting the task to the man: a textbook of occupa-tional ergonomics. Taylor & Francis, 4 edition, 1989.[3] M. J. Griffin. Handbook of Human Vibration. Academic Press Inc.,1990.[4] M. H. Pope, D. G. Wilder, and M. L. Magnusson. A review of studieson seated whole body vibration and low back pain. Proc Inst Mech Eng[H], 213(6):435–446, 1999.[5] M. Bovenzi. Health effects of mechanical vibration. G Ital Med LavErgon, 27(1):58–64, 2005.[6] M. Bovenzi. Health risks from occupational exposures to mechanicalvibration. Med Lav, 97(3):535–541, 2006.[7] Edward W. Finucane. Definitions, Conversions, and Calculations forOccupational Safety and Health Professionals, Third Edition. CRCPress, April 11, 2006.[8] Philip S Requejo, Grigor Kerdanyan, Jean Minkel, Rodney Adkins, andRobert Waters. Effect of rear suspension and speed on seat forces andhead accelerations experienced by manual wheelchair riders with spinalcord injury. J Rehabil Res Dev, 45(7):985–996, 2008.[9] Wolf EJ, Cooper RA, DiGiovine CP, and Ammer ML. Analysis ofwhole-body vibrations on manual wheelchairs using a hybrid iii testdummy. In Proceedings of the Annual RESNA Conference. RESNAPress, 2001.[10] Rory A Cooper, Erik Wolf, Shirley G Fitzgerald, Michael L Boninger,Rhys Ulerich, and William A Ammer. Seat and footrest shocks andvibrations in manual wheelchairs with and without suspension. ArchPhys Med Rehabil, 84(1):96–102, Jan 2003.[11] Andrew M Kwarciak, Rory A Cooper, and Shirley G Fitzgerald. Curbdescent testing of suspension manual wheelchairs. J Rehabil Res Dev,45(1):73–84, 2008.622.8. Bibliography[12] Carmen P DiGiovine, Rory A Cooper, Erik Wolf, Shirley G Fitzger-ald, and Michael L Boninger. Analysis of whole-body vibration duringmanual wheelchair propulsion: a comparison of seat cushions and backsupports for individuals without a disability. Assist Technol, 15(2):129–144, 2003.[13] Carmen P DiGiovine, Rory A Cooper, Shirley G Fitzgerald, Michael LBoninger, Erik J Wolf, and Songfeng Guo. Whole-body vibration dur-ing manual wheelchair propulsion with selected seat cushions and backsupports. IEEE Trans Neural Syst Rehabil Eng, 11(3):311–322, Sep2003.[14] Rory A Cooper, Erik Wolf, Shirley G Fitzgerald, Annmarie Kellerher,William Ammer, Michael L Boninger, and Rosemarie Cooper. Evalua-tion of selected sidewalk pavement surfaces for vibration experiencedby users of manual and powered wheelchairs. J Spinal Cord Med,27(5):468–475, 2004.[15] Erik Wolf, Jonathan Pearlman, Rory A Cooper, Shirley G Fitzgerald,Annmarie Kelleher, Diane M Collins, Michael L Boninger, and Rose-marie Cooper. Vibration exposure of individuals using wheelchairs oversidewalk surfaces. Disabil Rehabil, 27(23):1443–1449, Dec 2005.[16] Alicia M Koontz, Rory A Cooper, Michael L Boninger, Yusheng Yang,Bradley G Impink, and Lucas H V van der Woude. A kinetic analysisof manual wheelchair propulsion during start-up on select indoor andoutdoor surfaces. J Rehabil Res Dev, 42(4):447–458, 2005.[17] International Organization for Standardization. Mechanical vibrationand shock – Evaluation of human exposure to whole-body vibration –Part 1: General requirements (ISO 2631-1), 1997.[18] D. P. VanSickle, R. A. Cooper, M. L. Boninger, and C. P. DiGiovine.Analysis of vibrations induced during wheelchair propulsion. J RehabilRes Dev, 38(4):409–421, 2001.[19] Barbara Hughes, Bonita J Sawatzky, and Adrienne T Hol. A compar-ison of spinergy versus standard steel-spoke wheelchair wheels. ArchPhys Med Rehabil, 86(3):596–601, Mar 2005.[20] M. M. DiGiovine, R. A. Cooper, M. L. Boninger, B. M. Lawrence, D. P.VanSickle, and A. J. Rentschler. User assessment of manual wheelchair632.8. Bibliographyride comfort and ergonomics. Arch Phys Med Rehabil, 81(4):490–494,Apr 2000.[21] Setsuo Maeda, Makoto Futatsuka, Jiro Yonesaki, and Maki Ikeda. Re-lationship between survey results of vibration complaints of wheelchairusers and vibration transmissibility of manual wheelchair. Environmen-tal Health and Preventative Medicine, 8:82–89, July 2003.[22] http://www.newdisability.com/colourswheelchairs.htm.[23] www.spinergy.com/wheelchair/faq.aspx. Online.[24] Erik J Wolf, M. S Rory A Cooper, Carmen P DiGiovine, Michael LBoninger, and Songfeng Guo. Using the absorbed power method toevaluate effectiveness of vibration absorption of selected seat cushionsduring manual wheelchair propulsion. Med Eng Phys, 26(9):799–806,Nov 2004.[25] Michelle L Tolerico, Dan Ding, Rory A Cooper, Donald M Spaeth,Shirley G Fitzgerald, Rosemarie Cooper, Annmarie Kelleher, andMichael L Boninger. Assessing mobility characteristics and activitylevels of manual wheelchair users. J Rehabil Res Dev, 44(4):561–572,2007.[26] Sigrid N W Vorrink, Lucas H V Van der Woude, Allon Messenberg,Peter A Cripton, Barbara Hughes, and Bonita J Sawatzky. Comparisonof wheelchair wheels in terms of vibration and spasticity in people withspinal cord injury. J Rehabil Res Dev, 45(9):1269–1279, 2008.[27] Steven W. Smith. The Scientist and Engineer’s Guide to Digital SignalProcessing. California Technical Pub., 1997.643. Controlled Whole BodyVibration Study43.1 IntroductionThe daily activities simulation course study and the controlled speed wheelcomparison study examined the effect of wheel design on wheelchair vi-bration (and thus the effect of wheel design on the whole body vibrationto which wheelchair users are exposed) and the subjective perception ofspasticity. These studies could not objectively examine the relationship be-tween whole body vibration (WBV) and muscle spasticity in individualswith spinal cord injury (SCI) as:1. It was not possible to control the acceleration magnitudes to whichthe subjects were exposed2. The subjects were exposed to a limited range of vibration frequencies3. The subjects’ actual muscle response to the WBV exposure could notbe recorded for analysisIn order to study the relationship between WBV and spasticity and whetherWBV can be a trigger of spasticity in individuals with SCI, a method of ap-plying controlled WBV (in terms of acceleration amplitude and frequencypower spectrum) and objectively quantifying the individuals’ muscle re-sponse had to be developed.4A version of this chapter will be submitted for publication. Messenberg A, FinlaysonH, Sawatzky BJ, Cripton PA. Whole body vibration as a trigger of spasticity in individualswith spinal cord injury.653.1. IntroductionThere are several methods commonly used for spasticity assessment: self-report questionnaires, clinical assessments and biomechanical methods. Selfreport questionnaires are designed to assess the individual’s subjective per-ception of spasticity and usually include a set of discrete scores, such as inthe Penn Spasm Frequency Scale (Table 3.1), or a continuous scale (visualanalog scale. Figure 3.1).Score Description0 No spasms1 Mild spasms induced by stimulation2 Infrequent strong spasms occurring less than once per hour3 Spasms occurring more than once per hour4 Spasms occurring more than 10 times per hourTable 3.1: The Penn spasm frequency scaleFigure 3.1: The visual analog scale for subjective assessment of spasticity.The most frequently used clinical spasticity assessment method is the Mod-ified Ashworth Scale (MAS)1;2. Assessing spasticity with the MAS involvesan imposed stretch of the tested muscle and assigning a score, from 0 to 4,to the muscle tone (see table 3.2 for MAS score definitions).663.1. IntroductionScore Description0 No increase in muscle tone1 Slight increase in tone with a catch and release or minimal resis-tance at end of the range of motion when the affected part(s) ismoved in flexion or extension1+ Slight increase in muscle tone, manifested by a catch, followedby minimal resistance throughout the remainder (less than half)of the range of motion2 More marked increase in muscle tone through most of the rangeof motion, but the affected part(s) is easily moved3 Considerable increase in tone, passive movement difficult4 Affected part rigid in flexion or extensionTable 3.2: The modified Asworth scaleBiomechanical methods of assessing spasticity include isokinetic dynamome-ters, electrogoniometers, and electromyography (EMG). Isokinetic dynamome-ters are passive devices which resist applied forces and control the speed ofmotion. The devices record the applied force throughout the joint’s rangeof motion and allow the muscle’s resistance to passive movement to be cal-culated.Electrogoniometers are tools for measuring joint angles. By allowing a limbto swing freely about a joint and calculating the ratio between the initial andfinal position of the limb, spasticity of the limb’s muscles can be assessed asa reduction of the swing is generally found in individuals with spasticity2.Electromyography (EMG) is a method for detection and recording ofmuscle activity by measuring the electrical potential generated by musclecells during muscle activity and rest. Surface EMG is a method of EMGthat is conducted by placing sensors on the surface directly over the muscleof interest (the skin). Surface EMG has been shown to be a reliable andrepeatable method for documenting muscle motor unit behavior3;4;5;6;7.The raw EMG signal can be used to determine the muscle’s activation timing(is the muscle active or not?). However, in order to determine the degreeof activation and/or estimate muscle force, the Root Mean Square (RMS)and the average rectified value are commonly used. The RMS is the squareroot of the sum of squared values, averaged over the duration of the signal(see subsection 2.4.1 for additional details) and the average rectified value is673.1. Introductionan average of the signal’s absolute values. The RMS value represents signalpower but there is no specific physical meaning to the average rectifiedvalue8.The amplitude of the EMG signal generally increases with an increase inmuscle force or contraction velocity of the muscle. However, the relationshipbetween muscle contraction force/velocity and the EMG signal amplitudecan only be qualitatively assessed and a comparison between two EMGsignals to quantitatively determine how two muscle forces vary is not possibleto conduct with precision8.Many studies have been dedicated to WBV exposure and its effects on able-bodied individuals. A standard was created with the specific intention ofdefining methods of quantifying WBV in relation to health and comfort,probability of vibration perception and incidence of motion sickness in un-injured humans9. Very little, however, is known about how WBV affectsindividuals with spinal cord injury (SCI)10.Previous studies have looked at the vibration that individuals with SCImight experience from their wheelchair and how different elements of thewheelchair and ride surface affect that vibration11;12;13;14;15;16;17;18;19. Otherstudies have exposed non-SCI subjects to controlled WBV and examinedtheir comfort level20;21;22;23.Several studies have attempted to trigger muscle spasms in individuals withSCI:Zupan et al. set out to find the relationship between the Ashworth scaleand surface EMG based assessment of spasticity7. The spasticity level of98 subjects with SCI was first assessed using the Ashworth scale and thenpassive hip, knee and ankle movements were imposed while muscle activitywas recorded using surface EMG. Zupan considered any non-zero EMG ac-tivity resulting from passive maneuvers as an abnormal or spastic response.Zupan’s results indicated that surface EMG can be used to predict the Ash-worth score.Sko¨ld et al. also investigated the validity of the MAS score as a measureof spasticity in SCI individuals6. Fifteen SCI subjects underwent flexingand then extending of the knee during while being assessed a MAS scoreand simultaneous surface EMG recordings from their thigh muscles (RectusFemoris and Biceps Femoris). The EMG registration lasted 10 seconds and683.1. Introductionpeak activity was defined as the highest voltage after 1 to 3 seconds fromthe start of EMG registration. The EMG values of the agonist muscle (re-sponsible for generating the movement) was subtracted from those of theantagonist muscle (opposes the agonist muscle). The difference between theEMG values of the agonist and antagonist muscles was correlated with thesimultaneously estimated MAS score.Sko¨ld found that the EMG recordings correlated positively with simultane-ous MAS assessment of spastic muscle contraction and concluded that theAshworth scale may accurately reflect spasticity, triggered by movement, inSCI patients.McKay et al. used surface EMG recordings of lower limb muscles from 67SCI subjects to study muscle responses to reinforcement maneuvers, volun-tary movement attempts, vibration and the ability to volitionally suppresswithdrawal evoked by stimulation of the foot sole24. The RMS value ofthe EMG data was taken as an estimate of the corresponding spinal motorneuron pool activity but only EMG data from muscles of the leg passivelymoved were used for the study of spasticity. The average activity occurringin the 1 second period immediately preceding each maneuver was subtractedfrom each EMG recording.McKay defined different surface EMG threshold values to be considered aresponse for each of the applied maneuvers. Peak values greater than 1µVwere identified as responses to the reinforcement maneuvers, a minimumpeak value of 2.5µV was used as the threshold for a plantar reflex suppressionresponse and a minimum magnitude value of 3µV was required from theEMG data recorded during voluntary movement. A continued activation atan amplitude of 1µV or more from the muscle vibrated during the 30 secondsof vibration was required in order to be considered a tonic vibratory response(tonic stretch reflex).The tonic stretch reflex can be triggered through the application of a con-stant rate of stretch or by continuous activation of muscle receptors, such aswhen a vibrator is applied over the tendon or muscle25. When a vibrator isapplied over the muscle of an individual with an uninjured nervous system, asteady increase in motor unit activation can be seen in EMG recordings after1–2 seconds. After several additional seconds, the muscle activity reaches aplateauthat persists as long as the vibratory stimulus is applied. When avibrator is applied over a spastic muscle, a response will appear immediately693.1. Introductionand there is a sudden onset that may decrease in amplitude or be sustainedwhile the vibration is applied.McKay found a strong correlation between the subjects’ Ashworth Scalespasticity assessment scores and muscle EMG activity recorded during pas-sive limb movement. The study also found that the most common muscleresponse was the tonic vibration response, present in 48% of the subjects.The tonic vibration muscle response also correlated significantly with thesubjects’ Ashworth Scale spasticity assessment scores.A study by Sherwood et al. had 40 SCI subjects undergo reinforcementmaneuvers (deep breath, tightly squeeze eyes closed, clench jaw, lift head& maximal grip effort), voluntary movements (hip, knee and ankle flexionand extension), passive hip, knee and ankle movements, taps to the knee(patellar) and heel (Achilles) tendons, stroking of the foot sole and vibrationof the knee and ankle tendons5. Vibration of the thigh and ankle tendonswas accomplished by placing a pneumatic vibrator over the tendons for atleast 30 seconds. The thigh (quadriceps, adductors and hamstrings) andcalf (tibialis anterior, and triceps surae) muscle activity was monitored andrecorded using surface EMG.All of Sherwood’s surface EMG data (except for the tendon taps data) wasanalyzed by calculating the signals’ root mean square (RMS) values. Novalues of EMG activity were defined as thresholds for spasticity levels (ex-cept for the “low” activity level which was defined as values below 50µV )and response descriptions were based on the presence or absence of activ-ity. Sherwood found that vibration often failed to elicit any response inhealthy subjects but evoked activity in 44% of the stimulated muscles ofSCI subjects.In another study, Schmit et al. used EMG recordings from 10 SCI subjects,who were subjected to imposed hip movements at different velocities, to in-vestigate the role that hip proprioceptors (sensors of the muscle’s positionin space) play in triggering an extensor reflex response26. Hip movementswere imposed, with the ankle and knee held isometrically,using a speciallydesigned motorized leg brace. The resulting response to hip movement wascharacterized using hip, knee and ankle joint torque measurements and sur-face EMG from the calf muscles (Tibialis Anterior, Soleus and medial Gas-trocnemius) and thigh muscles (Vastus Medialis, Rectus Remoris, BicepsFemoris and Adductors). The rectified EMG recordings were evaluated to703.1. Introductiondetect the timing of muscle activity during and following the imposed hipmovement but no quantitative analysis of the EMG activity was conducted.Schmit found that the onset of flexor muscle activity typically precededextensor muscle activity and that the EMG signals of the extensor musclesoutlasted those of the flexor muscles. Schmit also found that EMG signalswere absent in cases where knee or ankle torques were not produced.Wu et al. used EMG to record activity of 6 leg muscles during controlledknee extension in fifteen SCI subjects27. A motorized leg brace, with torquetransducers at the knee and ankle positions, imposed knee flexion and ex-tension movements at 4 different speeds while the subjects were in a sittingposition and a supine (laying on back) position. Surface EMG data wasrecorded from calf muscles (Tibialis Anterior, medial Gastrocnemius) andthigh muscles (Vastus Medialis, Rectus Femoris and Hamstrings).Wu calculated the area of the rectified and smoothed EMG activity andnormalized the EMG data to the data obtained from the sitting position atone movement speed. Wu found that the knee flexor muscles (Hamstrings)were usually activated by the stretch itself and maintained throughout theensuing hold period. The muscle activity recorded from the hip flexor (Vas-tus Medialis), knee extensor (Rectus Femoris) and ankle extensor (medialGastrocnemius) muscles was consistent with torque responses of extensorspasms.Summary: The effect of individual wheelchair elements on wheelchair vi-bration and the wheelchair user’s exposure to WBV has been examined byseveral studies. Studies have exposed individuals with SCI to localized vi-bration by placing vibrators over muscle tendons and using surface EMGto record and analyze muscle activity response. Other studies have alsoused surface EMG to examine muscle activity, in individuals with SCI, inresponse to passive muscle stretch by imposing joint movements in an effortto trigger muscle spasms.However, no study has exposed subjects with SCI to frequency and accel-eration controlled WBV while using objective methods to examine whetherWBV may act as a trigger for muscle spasms in individuals with SCI. Thiswas the goal of the current study.713.2. Methods3.1.1 HypothesisThe hypothesis of this study was that Whole Body Vibrations (WBV) canelicit lower limb muscle spasms in people with Spinal Cord Injury (SCI).3.1.2 ObjectiveThe objective of this study was to design and build a system that wouldenable the application of WBV at controlled frequencies and amplitudesto seated human subjects with SCI and allow for the recording of muscleactivity in response to the applied WBV.This study was a pilot study. The goal of this study was to test the feasibilityof such a system. This system would be used for more extensive clinicalresearch.3.1.3 Ethics ApprovalApproval for this study was granted by the Vancouver Coastal Health Au-thority on 7/October/2008 (research study #V08-0268. See Appendix C forcertificate).3.2 MethodsSince vertical vibration is the most common vibration transmitted from thewheelchair to the user and also seems to be the main cause of complaintsby wheelchair users, the controlled Whole Body Vibration (CWBV) testingsystem was designed to expose seated subjects with spinal cord injury tovertical vibrations. The system that was to produce the vibrations had tomeet specific requirements for acceleration magnitude, vibration frequency,displacement amplitude and force output.The requirements for acceleration magnitude and vibration frequency weredefined by relying on the ISO 2631-1 standard’s guide for effects of expo-sure to whole body vibration on health to select the desired frequencies and723.2. Methodsaccelerations to be applied. The system’s displacement amplitude require-ment was defined by calculating the appropriate displacement amplitudesthat would result in the desired acceleration magnitudes at the specifiedvibration frequencies. The system’s force output requirement was definedby calculating the necessary force required to displace the estimated massbeing tested (subject mass + weight of supporting equipment) at the desiredfrequencies.The system also had to meet several functional, physical and safety criteria.The system would have to be relatively lightweight so as to allow it to betransported to and from the testing facility. The size of the system wouldhave to allow it to be assembled in a standard sized room in the testingfacility and the system’s configuration would have to allow the potentialsubjects, who would be wheelchair-bound, to access and transfer to and fromthe system’s test seat. A dedicated power generator would be unavailable sothe system would have to be plugged into a standard electric power outletand would therefore need to be powered by 110 VAC (single phase, 60 Hz).Naturally, the system would also have to be mechanically and electricallysafe to operate and be tested in. To meet these requirements (see subsection3.2.1 for a requirements summary), several different concepts were consid-ered. The following sections detail the the process of defining the system’srequirements and the various design concepts that were considered.3.2.1 Specification of System RequirementsVibration Exposure Level SelectionMaintaining a consistent vibration amplitude while varying the vibrationfrequency would allow examination of the effect of WBV frequency on mus-cle activity (and spasticity, if detected) and the possibility of a frequency–amplitude interaction affecting the subject’s muscle response would be elim-inated. It was therefore decided that this study’s subjects would be exposedto a consistent frequency weighted acceleration at all applied frequencies ofvibration.Relying on the ISO 2631-1 standard’s Health Guidance Caution Zones,shown in Figure 3.2, a vibration exposure level was selected. As the ISO2631-1 standard is defined for exposure of persons in normal health to733.2. Methodswhole body vibration, a vibration exposure level was selected and severalsafety factors were applied in order to ensure an exposure level that was safefor persons with spinal cord injury as well.Figure 3.2: ISO 2631-1 health guidance caution zones.ISO 2631-1’s health guidance caution zones’ borders remain at constant lev-els for exposures lasting 10 minutes or less. This study’s intended vibrationexposure consisted of 10 sessions, each lasting 20 seconds, for a total ex-posure time of 200 seconds (see subsection 3.2.7 for details). An exposureduration of 200 seconds, which includes periods of rest with no exposure atall, falls well below ISO 2631-1’s Health Guidance Caution Zones’ minimalvibration exposure duration definition of 10 minutes. It was decided thatfor this study the upper limit of the “safe” zone for exposures lasting 10minutes (and under) would be an acceptable starting point as a vibrationexposure level.The upper limit of the “safe” zone is not marked on ISO 2631-1’s guide andtherefore had to be interpolated between the two marked levels above andbelow the “safe” zone limit. The exposure level axis (vertical axis) on theHealth Guidance Caution Zones graph represents values of weighted ac-celeration (see subsection 3.2.1 for details), given in units of m/s2 howeverthe increments are non-linear. The exposure level axis increments are in-creases in Gain between weighted acceleration levels. Gain is a logarithmic743.2. Methodsratio between two given values and has units of decibel (dB). The definitionfor Gain (for amplitudes) in units of dB is:LdB = 20log(A1A0)(3.1)Where LdB is the amplitude ratio value (Gain) in units of dB and A1 & A0are the two amplitudes (the amplitude units cancel out as they are dividedby each other in the fraction). Applying equation 3.1 to the given weightedacceleration values, above and below the unmarked level of interest (theupper limit of the “safe” zone), on the Health Guidance Caution Zonesgraph reveals the gain per marked increment:LdB = 20log(42.5)≈ 4 dBThe increase in gain from the marked value of 2.5 m/s2 to the unmarkedvalue of the upper limit of the “safe” zone was found by measuring thedistance from the marked value to the unmarked value (see Figure 3.3).15.63mm9.01mmFigure 3.3: The distance between the marked value of weighted accelera-tion level to the unmarked value on the ISO2631-1 health guidance cautionzones figure.The total distance between two marked values is 15.63 mm. The distancebetween the marked value of 2.5 m/s2 to the unmarked value of the upper753.2. Methodslimit of the “safe” zone is 9.01 mm. Therefore, the increase in gain, GdBfrom 2.5 m/s2 to the unmarked value is the distance between the two levelsas a fraction of the gain between two marked distances (4 dB):GdB =(9.0115.63)× 4 = 2.3 dBLetting the unmarked value of the upper limit of the “safe” zone be A1 andreapplying equation 3.1, the value of the upper limit of the “safe” zone wasfound to be 3.26 m/s2:2.3 = 20log(A12.5)⇒2.320= log(A12.5)⇒ 102.320 =A12.5⇒ A1 = 2.5× 102.320 = 3.26 m/s2As a safety factor, a reduced value of 3 m/s2 was selected as the consis-tent vibration amplitude level to which the subjects in this study would beexposed. This reduced level of weighted acceleration is a 0.72dB reductionin Gain from the “safe” zone upper limit. Since 3 m/s2 is the weightedacceleration value, the corresponding instantaneous accelerations at eachvibration frequency had to be calculated.Acceleration Frequency WeightingAs the impact of vibration on the human body varies at different frequencies(see subsection 1.3 on page 10 for details), when calculating an individual’svibration exposure, the varying impact of the frequency of vibration mustbe accounted for by incorporating frequency weightings into the calculation.Frequency weightings are values by which the vibration magnitude at eachfrequency is to be multiplied in order to ’weight’ it according to its effect onthe body.The ISO 2631-1 standard defines different frequency weightings for the dif-ferent axes of vibration. The vibration frequency spectrum, recommended763.2. Methodsby ISO 2631-1 to be up to 80 Hz, is divided into bands (frequency ranges)of varying width and each band is assigned an appropriate weighting value.These frequency bands are sectioned according to the “13 -octave” calcula-tion method (see equation 2.7 on page 45 for frequency band calculationmethod). For vibration in the Z direction of a subject in the seated posi-tion, shown in Figure 3.4, the frequency weightings (Wk) are shown in table3.3.Figure 3.4: Axes of the human body as defined by ISO 2631-1According to ISO 2631-1, the frequency-weighted acceleration, aw, is to bedetermined according to the equation 3.2:aw =√∑(wi · ai)2 (3.2)Where ai is the instantaneous acceleration and wi is the weighting factor forthe ith frequency band, given in table 3.3.773.2. MethodsFrequency Band Frequency [Hz] Weighting Factor0 1 0.4821 1.25 0.4842 1.6 0.4943 2 0.5314 2.5 0.6315 3.15 0.8046 4 0.9677 5 1.0398 6.3 1.0549 8 1.03610 10 0.98811 12.5 0.90212 16 0.76813 20 0.63614 25 0.51315 31.5 0.40516 40 0.31417 50 0.24618 63 0.18619 80 0.132Table 3.3: Principal frequency weightings in 13 octaves (ISO 2631-1)Instantaneous Acceleration CalculationEquation 3.2 was used to calculate the instantaneous acceleration requiredat each frequency to maintain a consistent weighted acceleration exposure.Since vibration exposure (in this study) is only in the Z direction and consistsof a single frequency (per exposure), wk is the only weighting factor and thereis no summation. Equation 3.2 therefore becomes:aw =√(wk · ai)2⇒ aw = wk · ai(3.3)and the instantaneous acceleration, ai is then:ai =awwk(3.4)783.2. MethodsSetting the weighted acceleration, aw, to the selected 3 m/s2 and using table3.3 for the appropriate wk for each frequency, the appropriate instantaneousacceleration at each frequency was calculated. For frequencies for which aspecific weighting, wk, was not listed in table 3.3, the appropriate weightingfactor was interpolated. The displacements required to produce the calcu-lated instantaneous accelerations then had to be calculated.Displacement CalculationThe displacements required to produce the desired instantaneous accelera-tions were calculated, assuming sinusoidal vibration would be applied, usingequation 3.5:X(t) = Asin(ωt) (3.5)Where X is the position and A is the displacement amplitude. Calculatingthe second derivative (with respect to time), the acceleration is found:X˙(t) = Aω(cosωt)X¨(t) = −Aω2sin(ωt)(3.6)The effective value of a sinusoidal function is the root-mean-square (RMS)value of the function. The RMS value of a continuous function, f(t), is:fRMS =√1T2 − T1∫ T2T1[f(t)]2dt (3.7)So the effective value of the sine function, f(x) = Asin(θ), is:fRMS =√12pi∫ 2pi0[Asin(θ)]2dθ= A√12pi∫ 2pi0sin2(θ)dθ793.2. MethodsAnd since: sin2θ =1− cos(2θ)2:fRMS = A√12pi∫ 2pi01− cos(2θ)2dθ=A√2√12pi∫ 2pi0[1− cos(2θ)]=A√2√12pi(θ −sin(2θ)2)∣∣∣∣2pi0=A√2√12pi(2pi − 0− 0 + 0)=A√2Therefore, the effective value of the sinusoidal acceleration X¨(t) = −Aω2sin(ωt)is:X¨RMS =Aω2√2(3.8)Rearranging equation 3.8 to find the displacement amplitude (A):A =X¨√2ω2(3.9)As an additional safety factor, the maximal acceleration and not the effec-tive acceleration was used to calculate the displacement amplitudes.X¨(t) = −Aω2sin(ωt) is maximal when sinωt = −1, therefore:X¨max = Aω2 (3.10)And the displacement amplitude (for maximum acceleration) is:A =X¨ω2(3.11)The displacement amplitudes were calculated using equation 3.11, replacingX¨ with the instantaneous acceleration ai and ω with 2pif , where f is theappropriate vibration frequency:A =ai(2pif)2(3.12)803.2. MethodsForce CalculationThe force required to produce the desired accelerations was calculated usingequation 3.13:F = mx¨ (3.13)where x¨ is the instantaneous acceleration, ai.In order to calculate the exact force output, the mass to be acceleratedmust be known. As the exact mass to be accelerated was still unknown, anapproximate mass was estimated.The mass to be vibrated would include an adult human subject (estimatedto weigh up to 80 kg), a seat and a supporting structure (estimated to weighup to 20 kg) for a total estimated mass of 100 kg. The force output requiredper vibration frequency was then calculated by applying equation 3.4, witha mass of 100 kg, into equation 3.13 with the selected weighted acceleration(aw) of 3 m/s2:F = 100 · ai =100 · awwk=300wk(3.14)where wk is the appropriate weighting factor at each frequency (from Table3.3).System Requirements SummaryIn order to expose the estimated 100 kg load to the selected weighted accel-eration of 3 m/s2 at a frequency range of 5–80 Hz, the controlled whole bodyvibration system would have to meet the instantaneous accelerations, dis-placement amplitudes and system force output requirements listed in Table3.4.813.2. MethodsFrequency Acceleration Displacement Force5 Hz 2.89 m/s2 2.926 mm 288.7 N10 Hz 3.04 m/s2 0.769 mm 303.6 N15 Hz 3.72 m/s2 0.419 mm 372.1 N20 Hz 4.72 m/s2 0.299 mm 471.7 N25 Hz 5.85 m/s2 0.237 mm 584.8 N30 Hz 6.98 m/s2 0.196 mm 697.8 N35 Hz 8.16 m/s2 0.169 mm 816.3 N40 Hz 9.55 m/s2 0.151 mm 955.4 N45 Hz 10.71 m/s2 0.134 mm 1071.4 N50 Hz 12.20 m/s2 0.124 mm 1219.5 N55 Hz 13.46 m/s2 0.113 mm 1345.8 N60 Hz 15.01 m/s2 0.106 mm 1501.2 N65 Hz 16.70 m/s2 0.100 mm 1669.9 N70 Hz 18.32 m/s2 0.095 mm 1831.9 N75 Hz 20.29 m/s2 0.091 mm 2028.6 N80 Hz 22.73 m/s2 0.090 mm 2272.7 NTable 3.4: Instantaneous accelerations, displacement amplitudes and forceoutput required to achieve a weighted acceleration vibration exposure of 3m/s2 of a 100 kg load at frequencies between 5–80 Hz (in 5 Hz increments)In addition to the following performance requirements:Vibration frequency: 5–80 Hz (in 5 Hz increments)Displacement amplitudes: 0.09–2.9 mmForce output: > 2273 Nthe vibration system would also have to meet the following criteria:Weight - Allowing transport to/from the test facilityDimensions - Allowing:• Assembly at test facility• Access by subjects in wheelchairs823.2. MethodsPower consumption: 110 VACSafety - The system must:• Be electrically & mechanically safe for operator & subjects• Allow instant shutdown in case of emergencyTo meet these requirements, several different concepts were considered. Thefollowing section details the various preliminary concepts that were consid-ered.3.2.2 Preliminary WBV System ConceptsConcept 1: Fitness Vibration PlatformsVibration platforms are used mainly in fitness centers. The idea behind theseplatforms is that muscle vibration increases the maximum voluntary mus-cle contraction (by recruitment of more muscle spindles than contractionwithout vibration) and therefore exercising while standing on a vibratingplatform can increase one’s strength and allow better performance. Exam-ples of fitness vibration platforms currently available are shown in Figure3.5.(a) Pneumex (b) Pro Vibe (c) Galileo (d) Power PlateFigure 3.5: Commercially available fitness vibration platforms833.2. MethodsDisadvantages:• Limited control - Fitness vibration platforms commercially availabletoday offer very limited control over the frequency and amplitude ofvibration• Cost - The cost acquiring a commercially available fitness vibrationplatform (can be several thousands of dollars) was restrictive for thisstudy’s budgetThe fitness vibration platform concept was rejected.Concept 2: Rotating Discs SystemThe rotation discs system (Figure 3.6) would include a seating station, sup-ported by springs, and two disc shaped masses mounted eccentrically on twomotors. Each motor would rotate in the opposite direction. The rotating ec-centric discs would produce vibrations in all directions however due to theireccentric positioning and rotation in opposite direction from each other, thehorizontal component of acceleration from each rotating mass would cancelthe horizontal component from the other rotating mass, leaving only verticalacceleration (see Figure 3.6(a)).843.2. Methods(a) Vibration mechanism (b) Seating stationFigure 3.6: Rotating eccentric discs system: Vertical vibration generationmechanism (a) and seating station (b).Disadvantages:• Amplitude control - Controlling the system’s vibration amplitudewould require adjustment of the discs’ eccentricity. Adhering to ISO2631-1’s guidelines on allowable vibration amplitude per frequencywould require very high precision adjustments of eccentricity. Suchprecision would likely not be realistically achieved using the machin-ing tools and budget available for this study. Correction of vibrationamplitude while the system is in operation would be impossible and theneed to stop the system completely in order correct amplitude errors(by readjusting disc eccentricity) would render this system impracticalfor human subject testing purposes• Frequency control - Rotation startup and shutdown would result intransient vibration frequencies and amplitudes that would make it verydifficult (if not impossible) to expose subjects to only one vibrationfrequency and amplitude for a desired amount of time per test runThe rotating discs system concept was rejected.853.2. MethodsConcept 3: Large Electrodynamic ShakerA large electrodynamic shaker system (MB Dynamics C25HB, Figure 3.7)was made available for this study. The shaker system has a frequency rangeof 5-2000 Hz and a maximum force output of 22.3 KN. These specifica-tions offer the ability to directly vibrate large masses (more than adequateto vibrate adult human subjects) with good control of the frequency andamplitude.Figure 3.7: Large electrodynamic shakerThe proposed concept was to mount a seat directly on the electrodynamicshaker’s armature. The subject would be strapped into the seat and bevibrated directly by the armature.Disadvantages:• Missing\damaged components - The system amplifier and con-troller were inoperable due to missing and damaged internal compo-nents. The MB Dynamics C25HB electrodynamic shaker was manu-factured in the 1950’s and required parts that are considered techno-logically ancient and no longer available. The manufacturing company(MB Dynamics) was also no longer in existence• Cost - The cost of purchasing an adequate amplifier proved to berestrictive• Size - The height of the armature face to which the test seat would be863.2. Methodsattached would require a crane/hoist system to transfer the subjects(all with SCI) from their own wheelchairs into the test system seat.The height at which the subjects would be seated (on top of the elec-trodynamic shaker) would require that the testing facility have extrahigh ceilings. The width of the electrodynamic shaker would require atesting facility with an extra wide entrance for system delivery/removal• Weight - The weight of the system (over 2.5 Tons) posed a risk ofdamaging the test facility’s floor (during transfer and operation) andwould also make transferring the system to and from the test locationextremely difficultDespite managing to successfully operate and control the system (using anamplification and control system that was specifically designed and assem-bled for this system), the large electrodynamic shaker system concept wasrejected.3.2.3 Final Vibration System ConceptSmall Electrodynamic ShakerA small electrodynamic shaker system was also made available for this study(See subsection 3.2.4 for additional shaker system details). The availablesystem’s shaker weighed 82 kg and was relatively compact (shaker dimen-sions: 37.5 cm x 27.5 cm x 39.5 cm). Table 3.5 lists the shaker system’sspecifications along with the required specifications (defined previously).Specification Required AvailableFrequency Range 5–80 Hz 5–7500 HzDisplacement Amplitude 0.09–2.9 mm 0–19 mmForce Output 2273 N 489 NPower Voltage 110 VAC 110 VACTable 3.5: Required and available vibration system specificationsThe frequency range and armature displacement capabilities of the availableshaker were more than adequate for this study’s performance requirements873.2. Methodsand the size and weight of the shaker system allowed it to be transportedto any test location with relative ease.However, the electrodynamic shaker produced maximum force of only 489N. As 489 N is less than the force that would be required to accelerate anadult human subject at frequencies above 20 Hz (see table 3.4) it would beimpossible to use this shaker to directly vibrate the subjects. The test sys-tem was therefore designed to include a set of springs which would supportthe static mass of the subject and seating system being vibrated (Figure3.8(a)). The designed vibration system setup can be described by the modelshown in Figure 3.8(b).(a) Vibration system concept (b) Mathematical modelFigure 3.8: The controlled whole body vibration system concept and itsmathematical model. k is the equivalent spring constant, equal to the sumof the spring constants of each individual spring (k = 4ki).CWBV System Driving Force CalculationThe system’s governing equation of motion (assuming a sinusoidal drivingforce) is:mx¨+ cx˙+ kx = F sin(ωt) (3.15)The natural frequency of a system is the frequency at which the systemnaturally vibrates when displaced or set in motion and allowed to oscillatefreely. When a mechanical system is vibrated at a frequency that matches883.2. Methodsthe system’s natural frequency, the system tends to absorb more energythan it does when vibrated at other frequencies. The absorption of vibra-tional energy at the natural frequency is called resonance and results inan amplification of the system’s vibration amplitude.The damping ratio, ζ, is a measure of the decay of oscillations in a me-chanical system that has been displaced or set in motion. While no externaldamping system has been designed into the system, dissipation of vibrationenergy and the consequent decay of oscillations is unavoidable (due to fric-tion between the springs and their housing and due to the internal dampingof the system’s structural components).The natural frequency (ωn) and viscous damping coefficient (c) of a mechan-ical system consisting of a mass and spring are:ωn =√km⇒ m =kω2nc = 2mωnζ ⇒ ζ =c2mωnEquation 3.15 can therefore be rewritten as:x¨+ 2ζωnx˙+ ω2nx =Fkω2n sin(ωt) (3.16)The steady-state solution of equation 3.16 is:x(t) =Fk·1√(1− ( ωωn )2)2 + (2ζ ωωn )2sin(ωt− φ) (3.17)where φ is the phase shift between the driving force and the displacement:φ = tan−1(2ζ( ωωn )1− ( ωωn )2)(3.18)See appendix A for the complete solution of the CWBV system’s governingdifferential equation of motion.The displacement amplitude, X, is:X =Fk·1√(1− ( ωωn )2)2 + (2ζ ωωn )2(3.19)893.2. MethodsRearranging equation 3.19 provides the following relationship:XkF=1√(1− ( ωωn )2)2 + (2ζ ωωn )2(3.20)The ratioXkFis known as the amplification ratio and represents themechanical system’s frequency response. Figure 3.9 shows a plot of theamplification ratio as a function of the frequency ratio for varying dampingfactor values. The frequency ratio is the ratio between the frequency ofthe driving force and the natural frequency of the system.0246810120 0.5 1 1.5 2 2.5 3Amplification Ratio Xk/F Frequency Ratio f/fn  Mechanical System Frequency Response  ζ=0 ζ=0.1 ζ=0.2 ζ=0.5 ζ=1 Figure 3.9: System frequency response: vibration amplification ratio ver-sus frequency ratio for varying damping factor values.Differentiating x(t) twice produces the acceleration:x¨(t) = −Fk·ω2√(1− ( ωωn )2)2 + (2ζ ωωn )2sin(ωt− φ) (3.21)Rearranging equation 3.21 in order to find the force, F :F (t) = −x¨kω2 sin(ωt− φ)√(1−( ωωn)2)2+(2ζωωn)2(3.22)903.2. MethodsDifferentiating F (t) and equating to zero in order to find the maximumforce:F ′(t) =x¨k cos(ωt− φ)ω sin2(ωt− φ)√(1−( ωωn)2)2+(2ζωωn)2(3.23)F ′(t) = 0 ⇒ cos(ωt− φ) = 0 ⇒ (ωt− φ) = ±pi2Therefore:Fmax =x¨kω2√(1−( ωωn)2)2+(2ζωωn)2(3.24)CWBV System Spring SelectionEquation 3.24 was used in selecting appropriate springs to be used in theproposed controlled whole body vibration (CWBV) system. As the acceler-ation, x¨, at each frequency has previously been calculated and is now known,springs with appropriate values of k could now be selected so that the natu-ral frequency of the system would be such that the force required to producevibration at each desired frequency would not be larger than the availableforce output of the electrodynamic shaker.The system’s damping ratio, ζ, was still unknown. Calculating ζ requiresthat the system be allowed to vibrate freely and the decrease in displacementamplitude between oscillations be measured. This was not possible at thedesign stage as the system had not yet been assembled. Therefore, thedamping ratio had to be estimated. As no external damping mechanismwas included in the system and any damping would be the result of internaldamping of the system’s structural components, a damping ratio value of 0.1was used for all calculations of required force output per frequency range.A single spring that would be adequate for vibration throughout the entire5–80 Hz range while requiring a driving force under 489 N was impossible.The frequency range therefore had to be partitioned into several ranges anda different set of springs, with appropriate k values, selected for each range.Relying on available stock at a large local engineering equipment supplier(McMaster-Carr, USA. mcmaster.com), several spring sets were selectedand the frequency range was partitioned into the following ranges: 5–25 Hz,30–50 Hz, 55–65 Hz and 70–80 Hz.913.2. MethodsSprings that possessed adequate mechanical characteristics were available,however they were not all of the same design. The only spring adequatefor the 5–25 Hz range was a coil spring (Figure 3.10(a)) while the springsfor the 30–50 Hz, 55–65 Hz, and 70–80 Hz were all Belleville disc springs(Figure 3.10(b)). The vibration system would therefore have to be able toaccommodate both coil springs and Belleville disc springs. Figures 3.11(a)–(d) list the specifications of each spring set selected to be used in the CWBVsystem.                                                       (a) Coil spring                                                  (b) Belleville disc springFigure 3.10: Compression spring types: (a) coil spring. (b) disc spring.923.2. MethodsPart Number:  94125K624  $10.97 per Pack of 5Type Precision Compression SpringsMaterial SteelSteel Type Steel Music WireSystem of Measurement MetricOutside Diameter 11 mmOutside Diameter Tolerance ±0.4 mmWire Size 2.2 mmOverall Length 23 mmCompressed Length 16.5 mmEnds Closed and GroundWire Type Round WireLoad 337.35 NDeflection at Load 5.3 mmRate 63.55 N/mmRate Tolerance ± 5% N/mmSpecifications Met Deutsche Industrie Normen (DIN)DIN Specification DIN 17223, Class C, 1.1200, DIN 2095      (a) 5–25 HzPart Number:  96445K257  $3.91 per Pack of 12Type Belleville Disc SpringsMaterial SteelSteel Type High-Carbon SteelMinimum Inside Diameter 8.2 mmMaximum Outside Diameter 23.0 mmThickness .80 mmOverall Height 1.55 mmLoad 718 NLoad Tolerance ±25%, -7.5%Deflection at Load .562 mmFlat Load 842 NSpecifications Met Deutsche Industrie Normen (DIN)DIN Specifications (DIN) DIN 2093Notes Springs are cold formed.           (b) 30–50 HzPart Number:  96445K284  $6.70 per Pack of 12Type Belleville Disc SpringsMaterial SteelSteel Type Grade 6150 Chrome-Vanadium SteelMinimum Inside Diameter 16.3 mmMaximum Outside Diameter 31.5 mmThickness 1.25 mmOverall Height 2.15 mmLoad 1,913 NLoad Tolerance +15%, -7.5%Deflection at Load .675 mmFlat Load 2,359 NSpecifications Met Deutsche Industrie Normen (DIN)DIN Specifications (DIN) DIN 2093Notes Springs are cold formed and fineblanked. IDs and ODs are machined.           (c) 55–65 HzPart Number:  96445K251  $5.06 per Pack of 12Type Belleville Disc SpringsMaterial SteelSteel Type High-Carbon SteelMinimum Inside Diameter 10.2 mmMaximum Outside Diameter 20.0 mmThickness 1.1 mmOverall Height 1.55 mmLoad 1,521 NLoad Tolerance ±25%, -7.5%Deflection at Load .337 mmFlat Load 1,976 NSpecifications Met Deutsche Industrie Normen (DIN)DIN Specifications (DIN) DIN 2093Notes Springs are cold formed.           (d) 70–80 HzFigure 3.11: CWBV system springs: specifications of the spring sets to beused, per frequency range, in the CWBV system.In addition to partitioning the frequency range and using a different springset for each range, the estimated load to be vibrated also had to be reducedso that the entire 5-80 Hz range be covered. Table 3.6 lists each spring set’sk value, the natural frequency of the system and the maximum load thatcan be vibrated.Frequency k Value Natural MaximumRange (per spring) Frequency Load5–25 Hz 63.55 N/mm 8.5 Hz 90 kg30–50 Hz 1277.6 N/mm 39 Hz 85 kg55–65 Hz 2834 N/mm 56.5 Hz 90 kg70–80 Hz 4513 N/mm 73 Hz 85 kgTable 3.6: Frequency range, spring constant value, natural frequency andmaximum load per spring set selected to be used in the CWBV system.933.2. MethodsFigure 3.12(a) displays the force required at each frequency range and Figure3.12(b) shows the system’s frequency response with each set of springs.0501001502002503003504004505000 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85Required Force [N] Frequency [Hz]  CWBV Required Force per Frequency  5-25 Hz30-50 Hz55-65 Hz70-80 HzMa x(a) Required force01234560 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100Amplification Ratio Xk/F Frequency [Hz]  CWBV System Frequency Response  5-25 Hz30-50 Hz55-65 Hz70-80 Hz(b) Frequency responseFigure 3.12: CWBV system required force and frequency response: (a)- Required force at each frequency range (the dashed line represents theavailable electrodynamic shaker’s maximal force). (b) - System frequencyresponse at each frequency range.3.2.4 System Design & AssemblyThe testing system consisted of the following units:1. Vibration generation unit (VGU)2. System control unit (SCU)3. Data recording unit (DRU)Figure 3.13 shows the schematic layout of the CWBV system (a systemassembly and operation manual is included as appendix B).943.2. MethodsFigure 3.13: Schematic layout of the controlled WBV systemThe Vibration Generation Unit (VGU)The vibration generation unit consisted of a seating station (Figure 3.17),an LDS V456 elctrodynamic shaker (LDS Test and Measurement Ltd., Eng-land, www.lds-group.com, Figure 3.19) and an LDS P1000L power amplifier(Figure 3.20). The seating station was designed using Unigraphics-NX6 solidmodeling software (Siemens PLM Software, USA) and consisted of a seat(with footrests) rigidly fixed to a supporting platform (Figure 3.14).953.2. Methods(a) Side view (b) Top view (c) Isometric viewFigure 3.14: CAD model of the controlled whole body vibration system’sseating station (The seat is not shown). (a) Side view. (b) Top view (plat-form removed for clarity). (c) Isometric view.The platform was supported by the support frame via 4 springs, one springat each corner. The supporting springs were positioned in a groove inside aspecially machined cap. The horizontal and vertical beams of the supportframe were constructed from C-channel beams. The beams (Figure 3.15(a),bases (Figure 3.15(b), fittings (Figure 3.15(c)) and bolts were ordered from ametal framing system manufacturer (Unistrut, Canada. www.unistrut.com).                                                                                                                                                                      (a) Beam                                                                        1⁄4"(6.4)3 1⁄2"(88.9)                (b) Base                                                                                   (95.3)3 3⁄4"5 3⁄8"(136.5)(95.3)3 3⁄4"                       Fitting notched forcontinuous vertical.(c) FittingFigure 3.15: The vibration system seating station elements ordered fromUnistrut.The cap was designed to fit in the cavity of the support frame’s verticalpillar, and contained two pockets to allow installation of either coil springsor Belleville disc springs (the same cap can accommodate both spring types).Four caps were machined from a block of aluminum at the University ofBritish Columbia’s machine shop. Figure 3.16 shows the support frame’s963.2. Methodsvertical pillar (Figure 3.16(a)), the cap (Figure 3.16(b)) and the installedcap with a coil spring (Figure 3.16(c)) and Belleville disc spring (Figure3.16(d)) installed. Figure 3.17 shows the seat installed on the support frameand platform.(a) Vertical pillar (b) Cap(c) Installed coil spring (d) Installed Belleville disc springFigure 3.16: The seating station support frame: (a) Vertical pillar withoutinstalled cap. (b) Cap that fits in the frame’s vertical pillar and houses thecoil/disc spring. (c) The installed cap housing a coil spring. (d) The installedcap housing a Belleville disc spring.973.2. MethodsFigure 3.17: Vibration generation unit: The seating stationThe electrodynamic shaker (Figure 3.19) was positioned below the supportplatform and was powered by an LDS P1000L power amplifier (Figure 3.20).The electrodynamic shaker’s armature was bolted to the support platform,allowing accelerations in both the up and down directions to be transferredto the platform (see Figure 3.18).(a) Platform-shaker bolt: Top (b) Platform-shaker bolt: BottomFigure 3.18: Bolt connecting support platform to electrodynamic shaker’sarmature983.2. MethodsFigure 3.19: Vibration generation unit: Electrodynamic shakerFigure 3.20: Vibration generation unit: LDS P1000L amplifierThe System Control Unit (SCU)The system control center consisted of a signal generator (Figure 3.21(a)),a feedback accelerometer (Figure 3.21(b), an oscilloscope and a controller(LDS VLL1, Figure 3.22) that regulated the amplifier and controlled thevibration amplitude.The signal generator provided an acceleration signal and controlled the vi-bration’s frequency and wave form. The feedback accelerometer was fixedto the seating system’s seat and measured the instantaneous accelerationexperienced at the seat. The feedback acceleration data was fed into thecontroller and into the oscilloscope (for visual inspection).993.2. Methods(a) Signal generator (b) Feedback accelerometerFigure 3.21: The system control unit.The controller modulated the vibration amplitude. The desired accelerationamplitude was programmed into the controller (The desired amplitude wasselected by adjusting panel dials to the appropriate level). The controllerreceived the actual acceleration amplitude data from the feedback accelerom-eter and compared it with the desired acceleration amplitude, adjusting thethe amplifier’s output to correct any discrepancy.Figure 3.22: System control unit: Controller (LDS VLL1)The Data Recording Unit (DRU)The data recording unit consisted of an Electromyograph (EMG) recorder(Noraxon Myosystem 1400A), two 2-axis accelerometers (Mechworks Mech-sense MDS210U) and a computer. The EMG recorder was connected via1003.2. Methodswires to the subject’s muscles (externally only. See the following section forwire connection details) and received muscle activity signals from the sub-ject’s muscles and relayed the data to the computer for display and storage.The vibration controller was also connected to the EMG data recorder.This would allow synchronization of the start, duration and end of the vi-bration period with the muscle activity of the subject in response to thatvibration. The two accelerometers were mounted on the seating system’sfootplate and relayed foot acceleration data to the computer for display andstorage. The computer received, displayed and stored the EMG data (usingNoraxon’s MyoResearchXP software) and footrests’ acceleration data (usingMechworks’ Mechsense software).System TestingTwo separate methods were used to test the system for both vibration fre-quency and amplitude accuracy. The MDS210U digital accelerometer wasmounted onto the support platform and the analog feedback accelerometerwas mounted on the seat frame. The digital accelerometer was connectedto a computer and the analog accelerometer was connected to the systemcontroller and to an oscilloscope.A vibration frequency and amplitude was set and the system was allowedto vibrate. During the system’s operation, the acceleration data from thedigital accelerometer was recorded for later analysis and the data from theanalog accelerometer was visually inspected. The analog accelerometer out-puts a voltage of 100 mV/g. The oscilloscope peak-peak voltage readingwas recorded and the corresponding acceleration level was calculated by di-viding the oscilloscope reading by 200. The oscilloscope also reported thesignal frequency. An FFT was then performed on the data from the digitalaccelerometer to reveal the vibration frequency.The frequency and amplitude results from the digital and analog accelerom-eters was then compared to the frequency and amplitude that was dialedinto the system. This testing procedure was repeated multiple times withdifferent frequencies and amplitudes. During the first series of tests, ballastweights were used to simulate a test subject. During the second series oftests, an able bodied adult human was seated in the system (see Figure3.23).1013.2. MethodsFigure 3.23: Vibration system testingThe system accurately produced vibration frequencies and amplitudes as setby the operator and the system was considered ready for operation in thisstudy.3.2.5 Subject RecruitmentPotential subjects were recruited, through flyers advertising the study, fromthe G.F. Strong Rehabilitation Centre (www.vch.ca/gfstrong/) OutpatientSpinal Cord Injury Program and from the University of British Columbia’sPoint Grey campus. See appendix F for the recruitment flyer.To be considered for participation in the study, subjects had to have metthe following criteria:• 18 years of age or older• Spinal cord injury at or below the level of C7• Spasticity of at least Ashworth grade 1 or Spasm frequency scale grade1 for at least one year• Cognitive capability to understand and follow basic instructions andto give informed consent1023.2. MethodsSubjects with a current pressure sore, any history of cardiovascular or pul-monary disease that would make participation unsafe for the subject wereexcluded from the study. Two subjects, 1 female and 1 male, participatedin this study. See table 3.7 for details3.2.6 Test PreparationThe accelerometers were calibrated before being mounted on the seatingstation. Calibration of the accelerometers consisted of aligning the axes(±x, ±y) parallel to the direction of gravity and allowing the application todefine the perpendicular axes of measurement.Baseline Data CollectionUpon the subject’s arrival to the test facility, a Modified Ashworth Scale(MAS) assessment was carried out by a physician and baseline data wascollected. The same medical physician carried out the MAS assessmentsfor both subjects who participated in this study. Clinical baseline dataincluded the subject’s age, gender, diagnosis, co-morbidities and time sinceinjury. The participating subjects’ baseline data is listed in Table 3.7.Gender Age Injury Level Time Since InjurySubject 1 Female 54 C5 (Incomplete) 3 yearsSubject 2 Male 24 C5 (Incomplete) 7 yearsTable 3.7: Baseline data of the CWBV study subjects.Subjects also completed a self report questionnaire (included as appendix E)which included a Spasm Frequency Scale, Spasm Severity Scale, Interferencewith Function Scale, Painful Spasm Scale and a Visual Analog Scale (VAS).The “Spasm Severity Scale” and “Painful Spasm Scale” data provided infor-mation about the severity and pain/discomfort of the subjects’ spasticity ingeneral (not specifically on the day of testing). The “Visual Analog Scale”data indicated the severity of spasticity at the time of testing. Table 3.8lists the subjects’ self report questionnaire results.1033.2. MethodsSubject 1 Subject 2Day 1 Day 2 Day 1 Day 2Visual Analog Scale 7.8 0.7 1.9 4.7Spasm Severity Scale Mild Moderate Moderate ModeratePainful Spasm Scale None None None NoneTable 3.8: Baseline data results.Each subject received a detailed explanation on the test procedure and risks.The subjects then signed an informed consent form (attached as appendixD).EMG PreparationThe subject would then transfer from his/her wheelchair into the seatingstation seat. A lap belt was used to secure the subject in the seat. Thephysician then placed EMG sensor stickers on the subject’s skin, over themuscles of interest. The muscles of interest were the Rectus Femoris (Figure3.24(a)), Adductors (Figure 3.24(b)), Biceps Femoris (Figure 3.24(c)) andmedial Gastrocnemius (Figure 3.24(d)). These muscles were chosen as theyare involved in lower limb activity associated with spasms and because theirlocation, close to the skin, allows for muscle activity signal recording withthe use of surface EMG sensors rather than invasive methods.1043.2. Methods(a) Rectus Femoris (b) Adductors (c) Biceps Femoris (d) M. GastrocsFigure 3.24: Muscles from which activity was recorded, using surfaceEMG, during the CWBV study. (a) Rectus Femoris. (b) Adductors. (c)Biceps Femoris. (d) medial Gastrocnemius.Figure 3.25 shows a subject seated in the testing system with surface EMGsensor stickers connected to the subject’s skin, over the muscles of interest.The EMG sensors were then individually connected, via wires, to the EMGrecording system.1053.2. Methods(a) EMG sensors over muscles of interest (b) Subject seated in systemFigure 3.25: Seated subject with EMG sensors attached.3.2.7 Test ProcedureThe frequencies at which the subject would be vibrated were randomly se-lected (between 5–80 Hz, in 5 Hz increments) and the order of applicationwas also randomized. A vibration frequency would be selected and the sig-nal generator would be adjusted appropriately (The oscilloscope was used toverify that the signal produced by the signal generator was accurate). Thecorresponding acceleration amplitude (using table 3.4) was dialed into thevibration controller. The system was then ready for operation.The operator would start the system and there would be a delay before theelectrodynamic shaker began to vibrate. This delay was of varying durations(10–30 seconds) and prevented the subject from anticipating exactly whenthe vibrations would start. The delay was built into the controller and couldnot be adjusted by the operator.The subject would be vibrated for 20 seconds before the vibration wouldbe halted by the operator. The vibration halting was instantaneous. Thesystem would then be prepared for vibration at the next frequency and1063.2. Methodsamplitude. Preparing the system for the next vibration took approximately1 minute so the subject had approximately 1 minute rests (periods with novibration) between each vibration session.The goal was to expose each subject to at least 10 different frequencies ofvibration, re-exposing the subject on the second day to the same vibrationfrequencies as the first day, in random order. However, due to subject 1’stime constraints, only 7 of the 10 vibration frequencies applied to subject1 on the first day of testing could be applied on the second day of testing.Table 3.9 lists the order of frequencies at which each subject was vibrated.Order Frequency [Hz]of Subject 1 Subject 2Application Day 1 Day 2 Day 1 Day 21 50 25 60 702 25 55 55 153 70 70 10 554 35 50 50 305 80 15 15 506 10 20 25 107 55 10 35 358 20 NA 20 809 60 NA 80 2510 15 NA 70 6011 NA NA 30 2012 NA NA 5 5Table 3.9: Order of frequency of vibration to which each subject wasexposed during each day of testing (NA - Not Applied).The EMG sensor stickers were then removed and the subject transferredfrom the seating station back into his/her own wheelchair. This concludedthe first day of testing. The subject returned several weeks later and thevibration procedure was repeated. The frequencies (and corresponding am-plitudes) of vibration on the second day of testing were the same as thoseapplied in the first day of testing however, their order was again randomized.1073.3. Data AnalysisNOTE: The calculations conducted during the design of the CWBV test-ing system indicated that it would be necessary to use different spring setsfor different vibration frequency ranges in order to not exceed the electrody-namic shaker’s maximal force output capability (see subsection 3.2.3 on page94 for details). However, during testing, vibration was applied throughoutthe 5–80 Hz range using a single spring set and the electrodynamic shaker’sforce output capability was not exceeded.Data RecordingAs soon as the operator pressed the start button, recording of accelerationdata from the footplate mounted accelerometers and EMG data from thesurface EMG sensors would begin. As there was a delay before actual vi-bration would start, the recorded data included data while the subject wasat rest (no vibration) and while being vibrated.The signal from the controller to the amplifier was also fed into the EMGdata recorder. This allowed the vibration signal to be synchronized to themuscle activity response to the vibration.3.3 Data Analysis3.3.1 Acceleration AnalysisThe acceleration data from the seating system footrests was analyzed usingMatlab (Mathworks Matlab version 7). A Fast Fourier Transform (FFT, seesection 2.4.2 for details) was conducted on the acceleration signal, resultingin a Frequency Power Spectrum (FPS). The frequency of vibration (revealedby the FPS) at the footrests was then compared to vibration frequency atthe seat.3.3.2 Muscle Activity AnalysisThe EMG data was rectified and visually inspected for regions of increasedmuscle activity. The Root Mean Square (RMS) value of the regions of1083.3. Data Analysisincreased muscle activity and of the baseline muscle activity region (activityof the muscle at rest, before vibration was applied) was calculated. Theduration of each episode of increased activity was also calculated. The ratiobetween the RMS value of each region of increased muscle activity and theRMS value of the baseline muscle activity was then calculated.There is no standard definition for what amplitude and duration of invol-untary muscle activity constitutes a spasm. A previous study, by Læssøe etal, on the possible anti-spastic effect of penile vibratory stimulation defineda spasm as EMG activity of an amplitude equaling 4 times the baseline am-plitude, with a duration of more than 5 seconds28. It is not clear if Læssøeused absolute, average or RMS values in calculating muscle activity valuesand the reasoning for Læssøe’s definition of spasticity is not given.A study by Sko¨ld et al, that found that surface EMG recordings correlatepositively with MAS assessment of spasticity in SCI patients, correlatedthe EMG values from thigh muscles (Rectus Femoris and Biceps Femoris)during imposed knee movements to an appropriate MAS score6. Table 3.10lists Sko¨ld’s results. Sko¨ld’s results were used to calculate the ratio betweenpeak muscle activity and baseline muscle activity, per MAS score. ThePeak/Baseline ratios, calculated from Sko¨ld’s results, are shown in table3.11.1093.3. Data AnalysisFlexionMAS Score 1.00 1.27 2.17 2.38RFBaseline 17.69 µV 20.87 µV 30.27 µV 34.01 µVPeak 137.92 µV 146.6 µV 274.33 µV 259.61 µVDuration 1.09 s 1.53 s 1.67 s 2.76 sBFBaseline 11.69 µV 15 µV 15.3 µV 20.11 µVPeak 59.85 µV 73.13 µV 81.03 µV 111.86 µVDuration 0.88 s 1.38 s 1.37 s 1.93 sExtensionMAS Score 1.00 1.08 1.75 2.14BFBaseline 27.13 µV 26.54 µV 36.68 µV 47.27 µVPeak 103 µV 116.69 µV 137.88 µV 160.5 µVDuration 3.82 s 2.83 s 2.68 s 7.38 sRFBaseline 14.8 µV 12.92 µV 16.7 µV 23.13 µVPeak 45.07 µV 47.85 µV 52.12 µV 60.5 µVDuration 3.21 s 2.3 s 2.82 s 5.99 sTable 3.10: Skoo¨ld et al. study results: Simultaneous MAS measurementsand surface EMG recordings during knee movements. RF - Rectus Femorismuscle (Antagonist during flexion. Agonist during extension). BF - Bi-ceps Femoris muscle (Agonist during flexion. Antagonist during extension).“Baseline” is the mean EMG data values recorded before and after the im-posed knee movements. “Peak” is the mean EMG data peak values duringthe knee movements. “Duration” is the duration of continuous electricalactivity above the baseline activity level.FlexionMAS Score 1.00 1.27 2.17 2.38RF Peak/Baseline 7.80 7.02 9.06 7.63BF Peak/Baseline 5.12 4.88 5.30 5.56ExtensionMAS Score 1.00 1.08 1.75 2.14BF Peak/Baseline 3.80 4.40 3.76 3.40RF Peak/Baseline 3.05 3.70 3.12 2.62Table 3.11: Ratio between peak muscle activity and baseline muscle ac-tivity, per MAS score, calculated from Sko¨ld’s simultaneous MAS measure-ments and surface EMG recordings study. RF - Rectus Femoris muscle. BF- Biceps Femoris muscle.1103.4. ResultsThe average Peak/Baseline muscle activity ratio, calculated from Sko¨ld’sstudy, for muscle activity correlating to MAS scores below 2 (1.00–1.75) is4.76. The average Peak/Baseline muscle activity ratio for muscle activitycorrelating to MAS scores above 2 (2.14–2.38) is 5.57.Since there is no current standard regarding what EMG values define amuscle spasm, in terms of absolute values or ratio between peak and baselinevalues, a standard had to be defined for this study. Based on both Læssøe’sdefinition and the ratios calculated from Sko¨ld’s results, for this study, anEMG RMS Peak/Baseline muscle activity ratio of 4 and above was definedas a spasm.3.4 Results3.4.1 Acceleration ResultsAn FFT analysis of the acceleration at the footrests revealed that duringvirtually all the vibration sessions (with very few exceptions), both footrestsexperienced the same vibration frequency. The frequency of vibration at thefootrests was the same as the frequency of applied vibration (applied at theseat and verified in real-time).In some instances, more than one frequency was experienced at the footrestsdespite only one frequency being applied to the system. This occurred duringvibration in the 5–15 Hz frequency range.1113.4. Results00.511.522.533.5 01020304050607080Power Frequency [Hz] CWBV Footrest Frequency Power Spectrum Subject 1 ‐ Day 2 ‐ 10 Hz ‐ Sensor 1 (a) Applied frequency: 10 Hz00.511.522.533.50 10 20 30 40 50 60 70 80Power Frequency [Hz]  CWBV Footrest Frequency Power Spectrum  Subject 2 -  Day 2 -  15 Hz -  Sensor 1  (b) Applied frequency: 15 HzFigure 3.26: Footrest vibration frequency power spectrum. Several fre-quencies are present despite only a single frequency of vibration being ap-plied: (a) 10 Hz. (b) 15 Hz.The frequencies that were present in addition to the applied frequency werewhole multiples of the applied frequency (see Figure 3.26 for examples). Inmost cases however, the applied frequency was either the only or the mostdominant frequency experienced at the footrests. Table 3.12 lists the appliedfrequencies of vibration and the corresponding frequencies measured at bothfootrests.1123.4. ResultsVibration Frequency [Hz]AppliedMeasuredSubject 1 Subject 2Day 1 Day 2 Day 1 Day 2Sensor Sensor Sensor Sensor1 2 1 2 1 2 1 25 NA NA NA NA 13 0 6 310 10 10 10 10 20 10 19 1015 15 15 15 15 15 15 15 1520 20 20 20 20 20 20 20 2025 25 25 0 25 25 25 25 2530 NA NA NA NA 30 30 30 3035 35 35 NA NA 35 35 35 3550 50 51 50 50 50 50 50 5055 55 56 57 57 53 55 55 5560 60 60 NA NA 61 60 60 6070 70 71 70 70 70 70 70 7075 75 76 NA NA NA NA NA NA80 80 81 NA NA 80 80 79 79Table 3.12: Dominant frequency of vibration at seating system’s footrests.The first column lists the frequencies of the applied vibration. Sensor 1 and 2are the accelerometers mounted on the left and right footrests (respectively).Data listed as “NA” represents frequencies that were not applied.3.4.2 Muscle Activity ResultsEMG Visual InspectionA visual inspection of the EMG data showed that several vibration sessionsproduced discrete periods of muscle activation while other vibration sessionsproduced a continuous period of muscle activation that began as the vibra-tion started and lasted until the vibration was stopped. Many vibrationsessions did not trigger any muscle activity. Figure 3.27 shows examples ofmuscle activity responses.1133.4. Results0 10 20 30 40 50 60 700100200Subject 2, Day 1, 10Hz, Rectus FemorisTime [s]Muscle Activity [ μV]0 10 20 30 40 50 60 7005001000Subject 2, Day 1, 10Hz, AdductorsTime [s]Muscle Activity [ μV]0 10 20 30 40 50 60 700204060 Subject 2, Day 1, 10Hz, Biceps FemorisTime [s]Muscle Activity [ μV]0 10 20 30 40 50 60 700100200Subject 2, Day 1, 10Hz, Medial GastrocsTime [s]Muscle Activity [ μV]0 10 20 30 40 50 60 70−5000500Subject 2, Day 1, 10Hz, ShakerTime [s]Muscle Activity [ μV](a) Discrete activation0 5 10 15 20 25 30 35 40 4502040Subject 2, Day 1, 15Hz, Rectus FemorisTime [s]Muscle Activity [ μV]0 5 10 15 20 25 30 35 40 450204060 Subject 2, Day 1, 15Hz, AdductorsTime [s]Muscle Activity [ μV]0 5 10 15 20 25 30 35 40 450102030 Subject 2, Day 1, 15Hz, Biceps FemorisTime [s]Muscle Activity [ μV]0 5 10 15 20 25 30 35 40 450204060 Subject 2, Day 1, 15Hz, Medial GastrocsTime [s]Muscle Activity [ μV]0 5 10 15 20 25 30 35 40 45−5000500Subject 2, Day 1, 15Hz, ShakerTime [s]Muscle Activity [ μV](b) Continuous activationFigure 3.27: Examples of muscle activity produced in response to theapplied WBV. The top 4 graphs in each figure are plots of the muscle activitywhile the bottom graph (black color) is a plot of the applied vibration.Figure 3.27(a): Discrete muscle activity periods. Figure 3.27(b): Continuousmuscle activity.Due to a technical error during the first day of testing, the vibration signalwas not properly synchronized with the muscle activity data. Therefore, itis impossible to know exactly when the vibration began and ended for allsessions conducted with subject 1 on day 1. Several vibration sessions withsubject 1 on day 1 produced discrete instances of muscle activation but sincethe seat acceleration was not synchronized to the EMG recording, it is notclear how long after vibration began or ended, this activity occurred. Thiserror was then corrected and all following sessions were properly synchro-nized.EMG RMS AnalysisSeveral sessions of WBV resulted in increased muscle activity. There isno standard definition for spasticity in terms of muscle activity amplitudeand duration, however for this study, based on the definition of spasms ina previous study28 and ratios between peak and baseline muscle activitycalculated from another study’s results6, periods of muscle activity with anRMS value that was at least 4 times the RMS value of the baseline muscleactivity were considered to be spasms. Table 3.13 lists the instances ofmuscle activity where the peak RMS values were 4 or more times greaterthan the baseline RMS value.1143.4. ResultsSubject Day Frequency P/B Duration Muscle1110 Hz4.4 7.4 s Rectus Femoris4.3 2.2 sBiceps Femoris5.4 3.0 s20 Hz 4.9 1.2 s Rectus Femoris80 Hz4.3 2.5 sRectus Femoris4.5 2.3 s215 Hz 4.3 23.7 s Adductors20 Hz 89.8 2.1 s Rectus Femoris215 Hz13.7 19.3 s Adductors4.1 14.6 s Rectus Femoris13.3 29.9 s Medial Gastrocs10 Hz76.7 14.7 sAdductors4.7 9.1 s5.3 11.0 s Rectus Femoris16.5 22.5 sMedial Gastrocs6.0 11.6 s15 Hz6.2 22.7 s Adductors4.8 23.1 s Medial Gastrocs55 Hz 4.1 24.6 s Medial Gastrocs60 Hz8.3 7.9 sMedial Gastrocs5.5 21.9 s2 No Spasticity TriggeredTable 3.13: CWBV study EMG Results: Periods of muscle activity withEMG RMS P/B (Peak/Baseline) ratios of at least 4 and their duration.EMG results that showed an increase in a certain muscles activity for severalseconds, followed by a return to baseline activity for several more seconds,and then another increase in activity are shown as multiple Peak/Baselineratios and durations for a single muscle at a specific vibration frequency (persubject, per day).Figure 3.28 shows the total number of vibration exposures that were appliedto both subjects on both days of testing and the total number of spasmstriggered per muscle at each frequency (multiple bursts of increased muscleactivity during a single exposure were considered as a single spasm).1153.5. Discussion012345 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80Number of Events Frequency [Hz] Total Number of Vibration Exposures & Spasms Triggered  Ex posur esRectus Fem or i sB i ceps Fem or i sAddu ctor sGast r ocnem i usFigure 3.28: CWBV study EMG results: Total number of applied vibra-tion exposures at each frequency (on both days of testing) and total numberof spasms triggered at each applied vibration.3.5 DiscussionThis study has shown that it is possible to safely expose individuals withSCI to frequency and amplitude controlled whole body vibration (WBV)and to record muscle activity responses to the vibration using surface elec-tromyography. These responses can then be analyzed for timing, durationand magnitude (relative to each muscle’s activity level at rest). It may alsobe possible to apply selected/defined criteria, in terms of muscle activitymagnitude and duration, in order to identify muscle activity which can becharacterized as spastic.The fact that the system was able to vibrate a human subject through theentire 5–80 Hz frequency range without the need to switch spring sets issomewhat puzzling. Calculations of the force required to vibrate a 100 kg1163.5. Discussion(estimated to be the total mass of the subject and supporting structure) in-dicated that different spring sets would be required when applying vibrationat different frequency ranges (as detailed in subsection 3.2.3 on page 94).Without installing appropriate spring sets for each frequency range , theforce required to produce the desired accelerations in the frequency rangeswith inappropriate spring sets installed would be greater than the force thatthe electrodynamic shaker can provide. There are several possible explana-tions for this.It is possible that the electrodynamic shaker that was used is actually capa-ble of a higher force output than specified in the system’s manual. It is alsopossible that the entire system behaved as a much more complex system ofmasses and springs, with several degrees of freedom, than the mathematicalmodel used.It was noticed that the wooden support platform (supported by the springs)deflects during vibration and that altering the body’s posture and rigidityalters the amount of electric current required by the electrodynamic shakerin order to maintain the designated vibration level. This strengthens theassumption that the vibration system is, in reality, a very complex systemof masses and springs.Another possibility is that the system simply did not produce the desiredaccelerations at frequencies in ranges other than the range for which theinstalled spring set was appropriate. The system’s frequency response curve,shown in Figure 3.29, shows that the system amplification ratio quickly dropsto very low levels and is effectively at zero for frequencies above 40 Hz.1173.5. Discussion00.511.522.533.544.550 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100Amplification Ratio Xk/F Frequency [Hz]  CWBV System Frequency Response  Figure 3.29: The CWBV system’s frequency response, with the singlespring set that was used during testing, throughout the entire 5–80 Hz fre-quency range.The system’s frequency response curve indicates that vibrations at frequen-cies above 15 Hz are simply to fast for the system to respond. When vi-bration is applied at frequencies that are too high for the system, the elec-trodynamic shaker will attempt to displace the mass but before the massis sufficiently displaced and the force required to generate the required dis-placement is increased accordingly (potentially increased above the electro-dynamic shaker’s capability), the direction of the required acceleration (andthus the direction of the force) is reversed. The result is vibrations withdisplacements that are much smaller than the displacements that wouldproduce the desired accelerations.This study’s EMG results showed that WBV can illicit involuntary muscleactivity in people with SCI. Applying this study’s spasticity criterion of apeak/baseline EMG activity ratio of at least 4, resulted in several of themuscle responses being classified as spasms. The average peak/baselineactivity ratio, of the results that were considered to be spasms, for theRectus Femoris muscle was 16.8. This is much higher than the RectusFemoris muscle activity ratios calculated from Sko¨ld’s results6. However,when the single extreme peak/baseline ratio value of 89.8 is excluded, theaverage Rectus Femoris muscle activity ratio becomes 4.6. The averageBiceps Femoris muscle peak/baseline activity ratio is also 4.6. This is veryclose to Sko¨ld’s average results of 4.76 for MAS scores under 2.The different patterns of muscle activity (continuous/discrete) produced sug-1183.5. Discussiongest not only that WBV can illicit an involuntary muscle response but thatdifferent frequencies of WBV may result in different patterns of muscle ac-tivity. This should be investigated further.It is interesting that subject 1 experienced spasms on day one at both ex-tremes of the the frequency range (10–20 Hz and 80 Hz) but had no spasmselicited by the frequencies in between those extremes. The 10 Hz and 80 Hzvibration sessions were not the first sessions of the day, so it is not likely thatthe shock of sudden exposure to vibration triggered those spasms, but theywere conducted consecutively so the previous exposures may have served asa primer for spasticity.It is possible that the spasms were spontaneous and unrelated to the WBVbut the EMG graph shows several bursts of activity near the end of thesessions and no activity near the start of the sessions for the Rectus Femorismuscle and only one burst near the start of the 10 Hz session for the BicepsFemoris muscle. If the spasms were unrelated to the WBV, one would expectthem to present randomly/evenly throughout the EMG recording and notclustered near the end of the exposure.Neither subject experienced any muscle activity that can be defined as aspasm when exposed to WBV in the 25–50 Hz frequency range. Both sub-jects experienced most of the muscle activity responses during WBV expo-sures at frequencies of 20 Hz and below.It is interesting that both subjects experienced more spasms on the firstday of testing than on the second day. All of the vibration frequencies thatproduced a spasm on the first day of testing, except for subject 1’s RectusFemoris muscle spasm during the 20 Hz vibration session, failed to produceanother spasm on the second day of testing. Subject 2, who experiencedseveral spasms on day 1, experienced no spasms at all on day 2.One possible explanation for the reduced number of spasms on the secondday of testing (relative to the first day of testing) is that the subjects, havingalready experienced WBV sessions on the first day of testing, knew what toexpect during the sessions and were already familiar with all the researchersin the testing room and were therefore more relaxed on the second day.Other naturally varying factors such as stress level, bladder fullness andamount of sleep may have also played a roll in the subjects’ susceptibilityto spasticity.1193.5. DiscussionAnother possible explanation is that the subjects’ level of fatigue, especiallymuscle fatigue, differed on the two days of testing. Both subjects partic-ipated in the study, during at least one day of testing, after engaging inphysical activities. Subject 1 arrived for testing on day 1 after swimmingin the testing facility’s swimming pool. It is unknown whether she engagedin any physical activities prior to testing on day 2. Subject 2 arrived fortesting on day 2 shortly after playing an intense game of wheelchair rugby.His fatigue, or relaxed state, could have played a role in his muscles’ lack ofresponse to vibration on day 2.The CWBV system worked well, considering that it was built with a veryminimal budget and included an old, underpowered, electrodynamic shakersystem. Dialing in the desired vibration frequency and acceleration magni-tude was a relatively straightforward process, recording muscle activity wassimple and the entire test ran smoothly. Conducting the test does require atleast 2 operators - one to operate the vibration generation unit and anotherto operate the data recording unit. It is not possible however to verify, inreal time, that the subject is being exposed to the desired accelerations asthe operator in charge of the vibration generation unit must keep track ofthe amplifier’s electric current consumption so that it does not overload,monitor the subject for signs of distress or requests to stop the exposureand monitor the duration of exposure. Using 3 operators would allow realtime monitoring of the vibration acceleration magnitudes.If applying vibrations throughout the entire 5–80 Hz frequency range with-out installing the appropriate spring sets at each range resulted in accelera-tions that were lower than desired (and designed for), then the subjects werenot exposed to the same weighted acceleration at every exposure. To explorethis, the system should be tested again with the same, single, spring set in-stalled. A load should be vibrated throughout the entire 5–80 Hz frequencyrange and the actual acceleration, as measured by the feedback accelerom-eter, should be compared to the desired acceleration at each frequency. Inany case, this is an error on the part of the system operator and not a faultin the system itself.If using a single spring set throughout the entire 5–80 Hz frequency rangedoes in fact result in lower than expected displacements and accelerations,then it will be necessary to use the appropriate spring set for each frequencyrange. This will make testing subjects with SCI extremely difficult. TheCWBV system’s seating station is supported by the springs and exchanging1203.5. Discussionspring sets requires that the support platform be raised. Raising the supportplatform requires that the connection between the electrodynamic shakerand the support platform be loosened and that the subject transfer out ofthe test seat. The subject would then have to transfer back on once the newspring set has been installed and the shaker-platform connection tightened.Exchanging several spring sets during a test session would make participa-tion in the study very difficult for subjects with SCI as transferring in to andout of the test system’s seat is more difficult than in to or out of a wheelchairdue to the added height of the test seat, obstructions from the corners ofthe support platform and frame and the various cables that surround thesystem. It is also likely that the EMG sensors will either be displaced ordetached from the subject’s skin and that the EMG data wires will becometangled during the transfers.Having to reposition/reattach the EMG sensors after every transfer willlikely affect the EMG signal quality. It is also likely that the transfersthemselves (between the subject’s wheelchair and the test seat) would triggerincreased muscle activity, due to the imposed movements of the subject’slegs, as several previous studies have successfully triggered muscle spasmsin SCI subjects by imposing movement of their limbs7;6;5;26;27;29. It wouldbe impossible to determine whether the increased muscle activity, recordedduring exposure to vibration, is a result of the WBV or of the many transfersprior to the exposure.If the CWBV system, in its current configuration, cannot expose human sub-jects to the desired accelerations at every frequency in the 5–80 Hz rangewithout the need to replace the spring sets, it would not be advisable to usethe same system as used in this study if a similar study is to be conductedwith SCI subjects. A more powerful electrodynamic shaker, which can pro-vide a driving force that is large enough to directly vibrate human subjects,should be used. This would eliminate the need to incorporate supportingsprings into the system.On the other hand, if the apparent mass that is being vibrated is lowerthan the static mass of the load (due to the human body being made up ofmany individual systems with different resonant frequencies) and the forcerequired to produce the desired accelerations is within the current CWBVsystem’s capability, then there is no need to replace the spring sets duringtesting. The CWBV system, in its current configuration, could then be used1213.6. Conclusionwith a single spring set installed and the subjects would not have to transferin and out during testing. The system would then be appropriate for futurestudies that include subjects with SCI.3.5.1 LimitationsThe CWBV study was limited by several factors. As this study was a pilotstudy, intended to prove the feasibility of controlled WBV application andmuscle activity recording, only 2 subjects participated in the study. Nodecisive conclusions about WBV as a trigger of spasticity in populationswith SCI can be made from such a small sample group. It is also possiblethat a consistent frequency weighted acceleration was not maintained atevery vibration exposure due to the use of a single spring set throughoutthe entire 5–80 Hz frequency range, rather than installing the appropriatespring set for each range as calculated during the design of the system.3.6 ConclusionNo conclusions on which whole body vibration frequencies trigger spasms inpeople with SCI can be reached at this point as only 2 subjects participatedin the CWBV pilot study. However, results from this pilot study seem tosuggest that the link between whole body vibrations and involuntary muscleactivity in people with SCI is real. An additional study, with a larger groupof subjects, should be conducted in order to reach conclusions on this issue.There is also a need for a standard definition of muscle spasticity, in termsof muscle activity level and duration. A larger study, that perhaps combinesEMG signal recording with subject feedback, could help determine not onlywhat level and duration of muscle activity should be considered a spasm,but also define spasticity level and/or comfort categories.Both subjects experienced most of their muscle activity during WBV ex-posures at frequencies of 20 Hz and below, and neither subject experiencedany muscle activity that can be defined as a spasm when exposed to WBVin the 25–50 Hz frequency range. If additional studies, with a larger num-ber of subjects, repeat this result and thus show that WBV at 25–50 Hzis unlikely to trigger spasticity, a wheelchair suspension system that shifts1223.6. Conclusionthe wheelchair’s vibration frequencies into this range could be beneficial inpreventing muscle spasticity triggered by wheelchair vibration.3.6.1 Recommendations for Future WorkFuture research should take into consideration the type of physical activitiesthat the subjects were engaged in prior to being exposed to WBV and theirlevel of fatigue during the sessions. It could perhaps be a good idea tostandardize the subjects’ level of muscle fatigue by requesting that theyavoid any strenuous physical activities on the day of testing, thus eliminatingmuscle fatigue as a confounding variable. It might even be desirable tohave subjects repeat the test again after engaging in a controlled physicalactivity, allowing the effect of muscle fatigue on muscle response to WBVto be examined.The effect of vibration acceleration magnitudes on spasticity should also bestudied by exposing subjects with SCI to varying magnitudes of accelera-tion at specific frequencies and measuring the subjects’ muscle responses.Consideration should also be given to the effect of frequency order as thespasticity may be triggered not only by the specific exposure in which it wasobserved, but also by a cumulative effect of previous exposures.The values of muscle activity amplitude and duration that constitute aspasm are not well defined and further research on this is needed. If astandard is defined in the future, the peak/baseline ratio criterion used inthis study can easily be modified accordingly and the study’s EMG data canbe reanalyzed without the need to conduct additional testing.1233.7. Bibliography3.7 Bibliography[1] R. W. Bohannon and M. B. Smith. Interrater reliability of a modifiedashworth scale of muscle spasticity. Phys Ther, 67(2):206–207, Feb1987.[2] F. Biering-Srensen, J. B. Nielsen, and K. Klinge. Spasticity-assessment:a review. Spinal Cord, 44(12):708–722, Dec 2006.[3] M. P. Kadaba, M. E. Wootten, J. Gainey, and G. V. Cochran. Re-peatability of phasic muscle activity: performance of surface and intra-muscular wire electrodes in gait analysis. J Orthop Res, 3(3):350–359,1985.[4] A. Gollhofer, G. A. Horstmann, D. Schmidtbleicher, and D. Schnthal.Reproducibility of electromyographic patterns in stretch-shorteningtype contractions. Eur J Appl Physiol Occup Physiol, 60(1):7–14, 1990.[5] A. M. Sherwood, W. B. McKay, and M. R. Dimitrijevi. Motor controlafter spinal cord injury: assessment using surface emg. Muscle Nerve,19(8):966–979, Aug 1996.[6] C. Skld, K. Harms-Ringdahl, C. Hultling, R. Levi, and A. Seiger. Si-multaneous ashworth measurements and electromyographic recordingsin tetraplegic patients. Arch Phys Med Rehabil, 79(8):959–965, Aug1998.[7] B. Zupan, D. S. Stoki, M. Bohanec, M. M. Priebe, and A. M. Sher-wood. Relating clinical and neurophysiological assessment of spasticityby machine learning. Int J Med Inform, 49(2):243–251, Apr 1998.[8] Carlo J. De Luca. The use of surface electromyography in biomechanics.Journal of Applied Biomechanics, 13 (2):135–163, 1997.[9] International Organization for Standardization. Mechanical vibrationand shock – Evaluation of human exposure to whole-body vibration –Part 1: General requirements (ISO 2631-1), 1997.[10] Philip S Requejo, Grigor Kerdanyan, Jean Minkel, Rodney Adkins, andRobert Waters. Effect of rear suspension and speed on seat forces andhead accelerations experienced by manual wheelchair riders with spinalcord injury. J Rehabil Res Dev, 45(7):985–996, 2008.1243.7. Bibliography[11] D. P. VanSickle, R. A. Cooper, and M. L. Boninger. Road loads actingon manual wheelchairs. IEEE Transactions on Rehabilitation Engineer-ing, 8(3):371–384, Sept. 2000.[12] Wolf EJ, Cooper RA, DiGiovine CP, and Ammer ML. Analysis ofwhole-body vibrations on manual wheelchairs using a hybrid iii testdummy. In Proceedings of the Annual RESNA Conference. RESNAPress, 2001.[13] Rory A Cooper, Erik Wolf, Shirley G Fitzgerald, Michael L Boninger,Rhys Ulerich, and William A Ammer. Seat and footrest shocks andvibrations in manual wheelchairs with and without suspension. ArchPhys Med Rehabil, 84(1):96–102, Jan 2003.[14] Setsuo Maeda, Makoto Futatsuka, Jiro Yonesaki, and Maki Ikeda. Re-lationship between survey results of vibration complaints of wheelchairusers and vibration transmissibility of manual wheelchair. Environmen-tal Health and Preventative Medicine, 8:82–89, July 2003.[15] Carmen P DiGiovine, Rory A Cooper, Shirley G Fitzgerald, Michael LBoninger, Erik J Wolf, and Songfeng Guo. Whole-body vibration dur-ing manual wheelchair propulsion with selected seat cushions and backsupports. IEEE Trans Neural Syst Rehabil Eng, 11(3):311–322, Sep2003.[16] Wolf E, Cooper R, Kelleher A, Fitzgerald S, and Ammer W. Assess-ment of vibrations during manual wheelchair propulsion over selectedsidewalk surfaces. In Proceedings of the annual RESNA conference,2003.[17] Rory A Cooper, Erik Wolf, Shirley G Fitzgerald, Annmarie Kellerher,William Ammer, Michael L Boninger, and Rosemarie Cooper. Evalua-tion of selected sidewalk pavement surfaces for vibration experiencedby users of manual and powered wheelchairs. J Spinal Cord Med,27(5):468–475, 2004.[18] Pearlman P, Cooper R, Wolf E, Kelleher A, Fitzgerald S, and Am-mer W. Vibrations during manual wheelchair propulsion over selectedsidewalk surfaces are sensitive to weather-related surface wear. In Pro-ceedings of the annual RESNA conference, 2004.1253.7. Bibliography[19] Erik Wolf, Jonathan Pearlman, Rory A Cooper, Shirley G Fitzgerald,Annmarie Kelleher, Diane M Collins, Michael L Boninger, and Rose-marie Cooper. Vibration exposure of individuals using wheelchairs oversidewalk surfaces. Disabil Rehabil, 27(23):1443–1449, Dec 2005.[20] L. C. Fothergill and M. J. Griffin. The subjective magnitude of whole-body vibration. Ergonomics, 20(5):521–533, Sep 1977.[21] M. J. Griffin E. M. Whitham. The effects of vibration frequency anddirection on the location of areas of discomfort caused by whole-bodyvibration. Applied Ergonomics, 9:231–239, December 1978.[22] Chikako Kaneko, Takahide Hagiwara, and Setsuo Maeda. Evaluationof whole-body vibration by the category judgment method. Ind Health,43(1):221–232, Jan 2005.[23] Juha Kiiski, Ari Heinonen, Teppo L Jrvinen, Pekka Kannus, and HarriSievnen. Transmission of vertical whole body vibration to the humanbody. J Bone Miner Res, 23(8):1318–1325, Aug 2008.[24] W. B. McKay, H. K. Lim, M. M. Priebe, D. S. Stokic, and A. M.Sherwood. Clinical neurophysiological assessment of residual motorcontrol in post-spinal cord injury paralysis. Neurorehabil Neural Repair,18(3):144–153, Sep 2004.[25] Milan R. Dimitrijevic. Evaluation and treatment of spasticity. Neu-rorehabilitation and Neural Repair, 9:97 – 110, 1995.[26] Brian D Schmit and Ela N Benz. Extensor reflexes in human spinal cordinjury: activation by hip proprioceptors. Exp Brain Res, 145(4):520–527, Aug 2002.[27] Ming Wu, T. George Hornby, Jennifer Hilb, and Brian D Schmit. Ex-tensor spasms triggered by imposed knee extension in chronic humanspinal cord injury. Exp Brain Res, 162(2):239–249, Apr 2005.[28] Line Laesse, Jens Bo Nielsen, Fin Biering-Srensen, and Jens Snksen.Antispastic effect of penile vibration in men with spinal cord lesion.Arch Phys Med Rehabil, 85(6):919–924, Jun 2004.[29] Ming Wu, T. George Hornby, Jennifer H Kahn, and Brian D Schmit.Flexor reflex responses triggered by imposed knee extension in chronichuman spinal cord injury. Exp Brain Res, 168(4):566–576, Jan 2006.1264. DiscussionThis research had two main goals. The first goal was to determine whetherthe wheelchair wheel design influences the whole body vibration (WBV)that is experienced by the wheelchair user.The second goal was to design and build a system that would allow humansubjects with spinal cord injury (SCI) to be exposed to controlled WBV andalso allow their muscle activity in response to the applied vibration to berecorded and analyzed, thus allowing the relationship between WBV andmuscle spasticity in populations with SCI to be examined in future studies.Previous studies have examined the influence of several different wheelchaircomponents, such as suspension systems, cushions and ride surface, onwheelchair vibration. No studies however, have investigated how wheel de-sign affects wheelchair vibration. Therefore, the wheelchair wheel compari-son study adds new knowledge to the field of wheelchair vibration.The controlled whole body vibration (CWBV) study is innovative as it hasshown that humans with SCI can be safely exposed to controlled WBV andmuscle activity responses can be recorded and analyzed. Muscle spastic-ity may also be identified, assuming a standard criteria for spastic muscleactivity is (or will be) defined. This has not been attempted by previousstudies.The Wheel Comparison StudiesIn order to examine how wheel design influences wheelchair vibration, twowheel comparison studies were conducted: A daily activities simulationcourse (DASC) study and a controlled speed study.127Chapter 4. DiscussionThe Daily Activities Simulation Course StudyThe DASC study utilized a series of obstacles that have been used in previousstudies. By using individuals with SCI wheeling over several different vi-bration inducing obstacles, the DASC study allowed collection of wheelchairvibration data under conditions that closely resemble real world conditions.The feedback received from the participating individuals (via questionnairesand visual analog scales) suggested that experienced wheelchair users couldnot detect a difference, in their ride comfort or muscle spasticity, betweenthe Spinergy R© and steel-spoked wheels.This result, although highly subjective, puts the practical significance of anyobjective mechanical differences that might exist between the wheel designsinto question. If the individuals who use manual wheelchairs do not “feel” amore comfortable ride with less vibration or a ride that triggers less musclespasticity, what real benefit does a different wheel design offer?An acceleration analysis showed no statistically significant difference be-tween the tested Spinergy R© and steel-spoked wheels. The vibration induc-ing obstacles that were more “aggressive” (the ramp and three sinusoidalbumps) did produce higher acceleration values than the other obstacles.This is not surprising. However, there was however no significant acceler-ation difference between the two wheel designs and no consistent trend inregards to which wheel produced higher accelerations.A frequency content analysis showed that the wheelchair vibration, witheither wheel design, consists of frequencies clustered mainly around the 4–6Hz and 30–36 Hz ranges, regardless of which obstacle induced the vibration.The power of vibration, with either wheel design and with all obstacles,drops to negligible levels above 60 Hz.The only statistically significant frequency content differences between thetwo wheel designs were detected in the 52–60 Hz range, where the powerof vibration is very low in comparison with the vibration power at lowerfrequencies.Having individuals with SCI who use a manual wheelchair on a regular basisparticipate in the DASC study, use their own wheelchairs and their ownwheeling style as they negotiated the various obstacles, produced wheeling128Chapter 4. Discussionconditions that simulated “real world” conditions relatively well. On theother hand, allowing the subjects to use their own wheelchair resulted ina variety of wheelchairs being used in the DASC study. Each chair wasdifferent from the other in frame design and materials, suspension, weight,etc. Every user had his/her own seat cushion as well. This undoubtedlyintroduced many variables that affected the vibration of each wheelchair.Subjects also exhibited a variety of wheeling styles, speeds and obstaclenegotiation techniques. Some subjects would wheel slowly and cautiouslyover the obstacles while others would quickly wheel over them. Severalsubjects raised their front casters as they encountered an obstacle, absorbingthe impact with their main wheels only. Obstacle 4 (the ramp) seemed toencourage the most variability in technique, with some subjects coming to acomplete stop on the obstacle’s edge before continuing forward. One subjectleaped right over the ramp while another subject refused to wheel over it atall.While every subject used a different wheelchair and had a different wheelingstyle, every subject was also his own control as each subject used the samewheelchair with both wheel sets. Having each subject serve as his/her owncontrol should have limited the differences between trials (each trial is a runthrough the obstacle course with one set of wheels) to differences due towheel design. It is assumed that each subject maintained a similar wheelingstyle and speed while using each wheel set but there was no way for this tobe verified or controlled.In order to eliminate the variability that the use of many different wheelchairsand uncontrolled wheeling styles and speeds introduced, it was necessary toconduct another wheel comparison study where a single wheelchair would beused and the wheeling style and speed could be controlled. The controlledspeed wheel comparison study attempted to eliminate those variables.The Controlled Speed Wheel Comparison StudyThe multiple subject controlled speed (MSCS) wheel comparison study wasdesigned to eliminate many of the unknown variables present in DASC study.The use of a single wheelchair was designed to eliminate the variabilitythat the use of many different wheelchairs introduced into the DASC study.The controlled acceleration was designed to eliminate the variability due to129Chapter 4. Discussionpersonal wheeling speeds and the single obstacle combined with a passiverider (no subject action required) was designed to eliminate variability dueto obstacle negotiation techniques.The MSCS study results, which again showed no significant accelerationdifferences , no pattern as to which wheel type produced larger accelerationsand no significant frequency content differences between the two wheel types,further indicated that the two tested wheel designs do not have a differenteffect on wheelchair vibration.The MSCS study does not simulate “real-world” conditions as there is awide variety of wheelchair systems in use (each with its own structuraldesign, suspension systems and cushions) and countless different obstaclesthat wheelchair users encounter at different speeds on a daily basis. If nodifference between the two wheel designs could be revealed when a singlewheelchair, single obstacle and controlled speed were used, it is unlikely thatany difference would be seen under “real-world” conditions.Despite it becoming clear that neither of the two tested wheel designs offeredany advantage over the other in practical settings, in an effort to detect evenmechanically insignificant differences between the two wheel designs, it wasdecided repeat the MSCS study but to eliminate all remaining variables thatcould not be controlled.The different physical characteristics of the subjects who participated inthe MSCS study, such as different body weight or the different body fatdistribution between males and females, may have affected the results. It ispossible that some subjects braced themselves and shifted their position asthey were impacting the obstacle. Some subjects may have leaned one wayor the other as the wheelchair was accelerating and caused it to impact theobstacle at an angle.The single subject controlled speed (SSCS) wheel comparison study elim-inated these variables by using a single subject who maintained the sameposture throughout the study. No significant acceleration differences weredetected during the SSCS, despite all variability being controlled. However,quite a few significant frequency content differences were detected. Thesedifferences in vibration frequency were in the ranges of the most powerfulfrequencies.The vibration frequency content differences that were detected in the SSCS130Chapter 4. Discussionshow that different mechanical properties do exist between wheelchairs mountedwith different wheel designs but the testing conditions that were set in or-der to detect these differences are completely unrealistic. The differencesare therefore practically insignificant.The Controlled Whole Body Vibration StudyThe controlled whole body vibration (CWBV) study showed that safelyexposing individuals with SCI to controlled WBV and analyzing their mus-cle activity response is indeed feasible. The results of the CWBV studystrongly indicate that involuntary muscle activity, in individuals with SCI,can be triggered via WBV. The different patterns of muscle activity thatwere produced suggest that different frequencies of WBV may trigger dif-ferent patterns of muscle activity.It is possible that the pattern of muscle activation that begins as the vibra-tion begins and stops as the vibration stops is actually an artifact causedby an EMG cable that was not properly isolated from the vibration system.It is also possible that this activation pattern is actually a tonic stretch re-flex as the tonic stretch reflex can be triggered through the application of aconstant rate of stretch or by continuous activation of muscle receptors andis characterized by a sudden onset of muscle activity that may decrease inamplitude or be sustained while the vibration is applied1.A frequency power spectrum (FPS) analysis of the EMG signals that aresuspected as being either an artifact or an actual tonic stretch reflex wouldprobably make it possible to distinguish between real muscle activation andcable vibration. A signal with a large power value at the same frequencyas the applied vibration is likely to be the result of cable vibration. Dis-tinguishing between muscle activity and cable vibration when the appliedvibration is at frequencies similar to natural muscle activation frequencieswould probably require a close examination of the signal by a very experi-enced researcher.The fact that some vibration sessions produced several bursts of relativelyhigh level muscle activity that appeared several seconds into the vibrationsession and continued even after the WBV was stopped suggest that thesemuscle responses are in fact spasms and not an artifact caused by a vibratingEMG cable.131Chapter 4. DiscussionOnly one WBV session that produced a muscle spasm on the first day oftesting, produced a muscle spasm at the same frequency on the second dayof testing. One possible explanation for this is that if muscle spasticity isindeed triggered by WBV, it is not always triggered by the same frequencies.It is also possible that spasticity is not only frequency dependent but isalso influenced by many additional mechanical elements, such as order ofWBV frequency application. The fact that subject 2 experienced no musclespasticity when exposed to WBV after engaging in intense physical activitysuggests that an individuals physical condition at the time of WBV exposurecould also influence the triggering of spasticity.This study’s criterion for spasticity was a peak/baseline EMG activity ratioof at least 4. This criterion was based on the results of a previous studythat triggered spasms, via imposed limb movements, while conducting si-multaneous EMG recordings and MAS spasticity assessment2. The averagepeak/baseline activity ratio results of this study were much higher thanthose of the previous study (the study on which the spasticity criterion isbased). However when one extreme peak/baseline ratio result is discarded,the average muscle activity ratio from this study becomes very close to theprevious study’s average results for a MAS spasticity score under 2.The CWBV system used an electrodynamic shaker that was not powerfulenough to directly vibrate a human subject throughout the entire desiredfrequency range of 5–80 Hz. To overcome the limitation of an insufficientdriving force, supporting springs were incorporated into the CWBV sys-tem design, with a different set of springs intended to be used at differentfrequency ranges. During testing, a single spring set, giving the systeman approximate natural frequency of 8 Hz, was used throughout the entire5–80 Hz frequency range of vibration. The CWBV system operated success-fully throughout the entire 5–80 Hz frequency range without overloading theelectrodynamic shaker.It is possible that the CWBV system’s electrodynamic shaker was actuallycapable of a higher force output than specified in the system manual, how-ever this is unlikely. The force required to vibrate even an 80 kg load (whichis lower than the 100 kg load estimation used during the design of the sys-tem) at 80 Hz is still 3 times the maximum force output of the shaker (Threetimes for the required effective force. Four times for the required maximumforce). It is unlikely that the electrodynamic shaker can produce a forcethat is 3–4 times greater than its specified maximum force.132Chapter 4. DiscussionAnother possibility is that the CWBV system’s controller could not adjustthe driving force fast enough at vibration frequencies that were much higherthan the system’s natural frequency. The system’s natural frequency wasapproximately 8 Hz. When applying vibration at higher frequencies, a largeforce is required in order to achieve the displacements that would result inthe desired accelerations. The system’s controller must process the actualacceleration data, received via the feedback accelerometer, and adjust thedriving force appropriately . It is possible that the controller could notprocess the feedback data and adjust the output force before the directionof force was reversed. This could have been the result of improperly settingthe controller’s vibration adjustment rate (the “Compression dB” dial onthe controller’s control panel). The result may have been vibrations withdisplacements that are much smaller than the displacements that wouldproduce the desired accelerations.If using a single spring set throughout the entire 5–80 Hz frequency rangeinstead of installing an appropriate spring set for each frequency range re-sulted in lower than expected displacements, a lower than expected drivingforce would have been required. This would explain why the electrodynamicshaker did not overload (attempting to produce a force larger than its max-imum force output capability) even though it operated at frequencies thatrequired a different spring set than was installed. This would also mean thatthe subjects were not exposed to the desired accelerations at many vibrationfrequencies.It is also possible that the entire system, being made up of several non stiffelements, behaved as a much more complex system of masses and springswith several degrees of freedom. The bodies of the subjects, consisting ofseveral different systems with different natural frequencies, may have alsoaffected the system’s behavior. It was noticed that the wooden supportplatform (supported by the springs) deflects during vibration and that al-tering the body’s posture and rigidity alters the amount of electric currentrequired by the electrodynamic shaker in order to maintain the designatedvibration level. This strengthens the assumption that the vibration systemis, in reality, a very complex system of masses and springs.Calculating the force required to vibrate a human body, which is not a singlerigid mass, may not be as simple as using the subject’s mass in the governingequations of motion as the apparent mass that is vibrated may be lower thanthe static mass of the subject. If the apparent mass of the vibrated load133Chapter 4. Discussionwas in fact low enough so that the force required to vibrate the load, ateach frequency, was lower than the electrodynamic shaker’s maximum forceoutput capability then the subjects were actually exposed to the correctaccelerations with a single spring set being installed throughout the entire5–80 Hz frequency range.If the desired accelerations were experienced by the subjects throughoutthe entire 5–80 Hz range then that would mean that there is no need toexchange spring sets during testing. The CWBV system could then be usedfor future studies that include subjects with SCI as the subjects would nothave to transfer in to and out of the test system and all the drawbacks ofthese multiple transfers would be eliminated.Most of the spasms that were triggered during the CWBV study, were trig-gered in response to vibrations at the lower end of the frequency range (5–20Hz). When these results are combined with the results of the wheel compar-ison studies, it is evident that the frequency power spectrum of wheelchairvibration has the most power at the lower end of the frequency range (2–10Hz). Therefore, most of the spasms were triggered at frequencies that are themost powerful during wheelchair vibration. Figure 4.1 shows total numberof spasms triggered per muscle at each frequency during the CWBV studyand the peak power of the Spinergy and steel-spoked wheelchair wheels’frequency power spectrum during the DASC study.1344.1. Conclusion012345 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80Number of Events Frequency [Hz] CWBV Total number of exposures & spasms  with DASC peak frequency power overlay  E xpos uresR ectus FemorisBic eps FemorisAdducto rsGa stro c nemiusDASC  FPS (Steel)DASC  FPS (Spinergy)Figure 4.1: Combined research results. CWBV study: Total number ofapplied vibration exposures at each frequency and total number of spasmstriggered. DASC study: FPS peak power results.4.1 ConclusionThis research contributed to the body of knowledge on wheelchair vibrationand on whole body vibration (WBV) as a trigger of muscle spasticity inindividuals with SCI. The influence of wheelchair wheel design on wheelchairvibration has not been studied before and exposing individuals with SCIto controlled whole body vibrations while simultaneously recording theirmuscle response, using surface EMG, for detection of spasticity has alsonever been done before.The relatively new composite material spoked wheelchair wheel design thatwas tested (Spinergy R©) does not significantly differ from the commonly usedsteel spoked wheel design, in terms of vibration frequency content and ac-celeration magnitudes. When weighing the advantages/disadvantages thatSpinergy R© wheels offer over standard wheels, vibration should not be theonly consideration. Ride comfort, cost (purchase and maintenance) andweight of Spinergy R© wheels versus standard wheels should also be consid-ered. However, in terms of vibration frequency content, acceleration magni-1354.1. Conclusiontudes and spasticity reduction (in individuals with SCI) Spinergy R© wheelsseem to offer no practical benefit to manual wheelchair users. Additionalresearch on the influence of wheel design on wheelchair vibration should notinclude the Spinergy R© wheel design.Controlled WBV, in terms of vibration frequency and acceleration mag-nitude, can be safely applied to individuals with SCI and muscle activityresponse to WBV can be recorded for analysis and detection of spasticityusing surface EMG. Although a standard definition of muscle activity thatconstitutes a spasm is not currently available, the results of this researchsuggest that WBV can illicit involuntary muscle activity in people with SCIand that some muscle responses to WBV may be considered as spasms.Additional research on WBV as a possible trigger of muscle spasticity inpopulations with SCI is recommended. A repeat of the controlled WBVstudy conducted in this research, with a much larger number of participatingsubjects, is highly recommended so that more definitive conclusions aboutthe relationship between WBV frequencies and accelerations and the onsetof spasticity in people with SCI can be made.An understanding of the vibration frequencies and accelerations that triggerspasticity in populations with SCI will help define new standards on exposureto WBV that are dedicated to people with SCI. Defining such standards isnecessary as currently only standards relevant to the exposure of uninjuredhumans to WBV exist. The current standards do not take into considerationthe increased susceptibility of individuals with SCI to injury and musclespasticity.The results of this research suggest that current wheelchair systems vibrateat frequencies that are most likely to trigger spasticity in individuals withSCI. If a larger study confirms these results, the design of wheelchairs andother vehicles for individuals with SCI should be adapted to avoid thosevibration frequencies that are most likely to trigger spasms. Wheelchairsuspension systems should be designed to not only reduce acceleration am-plitudes, but also shift the frequencies of vibration.Helping individuals with SCI avoid exposure to WBV that may trigger mus-cle spasticity will hopefully reduce the number and intensity of spasms ex-perienced by those individuals. This will undoubtedly improve the qualityof life of individuals with SCI.1364.2. Bibliography4.2 Bibliography[1] Milan R. Dimitrijevic. Evaluation and treatment of spasticity. Neurore-habilitation and Neural Repair, 9:97 – 110, 1995.[2] C. Skld, K. Harms-Ringdahl, C. Hultling, R. Levi, and A. Seiger. Si-multaneous ashworth measurements and electromyographic recordings intetraplegic patients. Arch Phys Med Rehabil, 79(8):959–965, Aug 1998.137A. CWBV System Equationof Motion: CompleteSolutionFigure A.1: Mathematical model of the vibration system.Summing the forces and applying the relationship:∑F = mx¨We get:Fd − cx˙− kx = mx¨ (A.1)where Fd is the external driving force, c is the viscous damping coefficient,m is the mass being vibrated and t is time. Assuming a sinusoidal drivingforce, the equation of motion becomes:mx¨+ cx˙+ kx = F sin(ωt) (A.2)138Appendix A. CWBV System Equation of Motion: Complete SolutionThe natural frequency and viscous damping coefficient of a mechanical sys-tem consisting of a mass and spring are:ωn =√km⇒ m =kω2nc = 2mωnζ ⇒ ζ =c2mωnEquation A.2 can therefore be rewritten as:x¨+ 2ζωnx˙+ ω2nx =Fkω2n sin(ωt) (A.3)The general solution of equation A.3 is:x(t) = xc(t) + xp(t) (A.4)where the complimentary solution, xc(t), is the solution of the free (unforced)vibration:x¨+ 2ζωnx˙+ ω2nx = 0 (A.5)The solution of equation A.5 is in the form:xc(t) = AeλtPlugging xc(t) into equation A.5 produces:λ2 + 2ζωnλ+ ω2n = 0 ⇒ λ1,2 = ωn(−ζ ±√ζ2 − 1) (A.6)and the solution to equation A.5 becomes:xc(t) = A1e(−ζ+√ζ2−1)ωnt +A2e(−ζ−√ζ2−1)ωnt (A.7)where the constants A1 and A2 are determined from the initial displacementand velocity conditions (xc(0) and x˙c(0) respectively).Since:eiθ = cos θ + i sin θ⇒ cos θ = Re(eiθ)⇒ sin θ = Im(eiθ)139Appendix A. CWBV System Equation of Motion: Complete Solutionthe sinusoidal driving force is:Fkω2n sin(ωt) = Im(Fkω2neiωt)(A.8)and therefore equation A.3 can be written as:x¨+ 2ζωnx˙+ ω2nx = Im(Fkω2neiωt)(A.9)The particular solution of equation A.9, xp, is in the form:xp(t) = XeiωtPlugging xp(t) into equation A.9 produces:−Xω2eiωt + 2ζωnXiωeiωt + ω2nXeiωt =Fkω2neiωt⇒ Xeiωt(−ω2 + 2iζωωn + ω2n) =Fkω2neiωt⇒ X =Fk ω2nω2n − ω2 + 2iζωnω(A.10)Multiplying and dividing the right side of equation A.10 by the denomena-tor’s conjugate, ω2n − ω2 − 2iζωnω, produces:X =Fk ω2n(ω2n − ω2 − 2iζωnω)(ω2n − ω2)2 + (2ζωnω)2(A.11)Using Euler’s formula for complex numbers:z = x± iy = |z|(cosφ± i sinφ) = (√x2 + y2)e±iφφ = tan−1(yx)equation A.11 becomes:X =Fk ω2n√(ω2n − ω2)2 + (2ζωnω)2(ω2n − ω2)2 + (2ζωnω)2e−iφ=Fk·ω2n√(ω2n − ω2)2 + (2ζωnω)2e−iφ=Fk·1√(1− ( ωωn )2)2 + (2ζ ωωn )2e−iφ(A.12)140Appendix A. CWBV System Equation of Motion: Complete Solutionandφ = tan−1( 2ζωnωω2n − ω2)= tan−1(2ζ( ωωn )1− ( ωωn )2) (A.13)where X is the displacement amplitude and φ is the phase shift between thedriving force and the displacement.Plugging X from equation A.12 back into xp(t):xp(t) =Fk·1√(1− ( ωωn )2)2 + (2ζ ωωn )2e−iφeiωt=Fk·1√(1− ( ωωn )2)2 + (2ζ ωωn )2ei(ωt−φ)(A.14)Since only the imaginary portion of Fk ω2neiωt is applied in equation A.9, onlythe imaginary portion of xp(t) is applicable and xp(t) becomes:xp(t) = Im(Fk·1√(1− ( ωωn )2)2 + (2ζ ωωn )2ei(ωt−φ))⇒ xp(t) =Fk·1√(1− ( ωωn )2)2 + (2ζ ωωn )2sin(ωt− φ)(A.15)141B. CWBV System Manual Sig nal G enerat or  C ont rol ler  Amplif ier  Elect ro dy namic S hak er  Accelerat io n Fe edback  DISCLAIMERThis manual comes with absolutely no guarantee of any kind. Ex-posing humans to whole body vibration may be harmful. Do notoperate the system without first understanding the risks involved(refer to ISO 2631-1). The manual refers to equipment that wasavailable during original design and testing. Using different equip-ment may or may not work and may or may not require adjust-ment of settings.142Appendix B. CWBV System ManualASSEMBLYPosition shaker and support frame so that they are as concentric as possi-ble Place either coil springs or disc springs (disc type depends on intendedvibration frequency range) in support frame caps (Figure B.1).(a) Compression Spring (b) Disc SpringFigure B.1: Supporting SpringsPlace disc/washer on upward facing winged-nut that is attached to the boltconnected to the shaker’s armature. Place wooden support plate (with at-tached seat) on springs. The bolt from the shaker’s armature must runthrough the central hole in the wooden plate without applying pressure tothe bolt (Figure B.2). Carefully reposition shaker/frame to verify.Figure B.2: Support platform resting on the support springs with the boltrunning from the shaker’s armature through the central hole of the platform.143Appendix B. CWBV System ManualThe wooden plate should be supported only by the springs and not by thebottom winged-nut. Space should be left between the wooden plate and thebottom disc/washer so that the plate can move down without moving thebolt with it. This will allow the armature to remain in the neutral positionuntil actual operation. Place washer/disc on bolt on top of wooden plateand fasten with winged-nut (Figure B.3). After the subject/load hasbeen mounted on the seat: Tighten both top and bottom winged-nuts.Figure B.3: Fastened wooden support platform - Top view.Bolt feedback accelerometer to seat-plate (Figure B.4).Figure B.4: Feedback accelerometer bolted to seat-plate.CONNECTIONS• Feedback accelerometer connects, through hardware filter (2KHz), toSignal Conditioner “INPUT”.144Appendix B. CWBV System Manual• Signal conditioner “OUTPUT” connects to Controller “NORMAL-IZED INPUT”.• Signal conditioner “OUTPUT” may also be connected (via BNC split-ter) to oscilloscope for visual inspection of actual vibration.• “ACCELERATION OUTPUT” on Controller should remain discon-nected.• Signal Generator “OUTPUT” connects to Controller “SERVO IN-PUT”.• Signal Generator “OUTPUT” may also be connected (via BNC split-ter) to oscilloscope for visual inspection of generated vibration signal(frequency).• Controller “SERVO OUTPUT” connects to Amplifier “INPUT”.• Amplifier “OUTPUT” cable connects to Electrodynamic Shaker.• The Amplifier, Controller, Signal Generator, Signal Conditioner andOscilloscope all need to be connected to a standard electric outlet (110VAC).SYSTEM CONTROLS        Controller  Amp lifier  Signal Ge ne rator  Sig na l  Condi t i oner  Oscil l oscope  Figure B.5: Complete control system.145Appendix B. CWBV System ManualFigure B.6: Amplifier - LDS P1000L.Figure B.7: Controller - LDS VLL1.Amplifier:• “Overcurrent”: Displays the amount of electric current being outputby the amplifier. Press button to view the amount of electric currentat which the amplifier will cut off (stop output). To adjust currentcut-off level, press and hold “Overcurrent” button and turn adjacentdial (using screwdriver).Controller:• “Sine/Random”: indicates vibration signal form.• “CHARGE AMPLIFIER”: Unknown use. Possibly adjusts for dif-ferent feedback accelerometers that produce charge (not voltage) inresponse to acceleration. The use of the 100 mV/g feedback accelerom-eter through the signal conditioner bypassed this option.146Appendix B. CWBV System Manual• “Trip Reset”: Safety feature that prevents system from vibratingabove a set level. Lights up when vibration level exceeds set cut-off level. System operation will cease. Press to reset. System will notautomatically restart.• “Set Trip Level”: Press to view current trip level. Level will display on“VIBRATION LEVEL” display. To set “Trip Level”, press the “SetTrip Level” button and adjust appropriate dial (using screwdriver).• “Metering Selector”: Selects how the system will be controlled:– “Disp”: Displacement.– “Vel”: Velocity.– “Acc”: Acceleration.• “Mode”: Controls current state of system - should be initially set to“Inhibit”.• “PROGRAM LEVEL”: Controls the vibration level.• “Rate”: Controls the rate of error correction (if the system vibrates ata level that is higher/lower than set, a high rate will correct the errorin larger steps).• “Compression dB”: Controls the rate of error correction when “Rate”is set to “Manual”.OPERATION1. Power on all equipment (Controller power switch located on backpanel).2. Set Signal Generator to desired vibration frequency.Amplifier: All indicator lights on Amplifier should be on (steady, notblinking) Press “Overcurrent” button to view level of output current atwhich the amplifier will cut off (stop output). To adjust current cut-offlevel, press and hold “Overcurrent” button and turn adjacent dial (usingscrewdriver).147Appendix B. CWBV System ManualController:3. “Sine/Random”should be set to “Sine” (press to set) for a sinusoidalinput signal.4. Set vibration level by adjusting the “PROGRAM LEVEL” dials. Nofeedback will display.5. Press “Select Mode” once to select “Program” mode• After time delay, Controller will adjust and “VIBRATION LEVEL”display will indicate the corresponding vibration level in Volts/g(Peak, not Peak-Peak).• System does not vibrate while in “Program” mode.6. Adjust “VIBRATION LEVEL” dials until desired vibration level isindicated.7. Press “Select Mode” once to select “Standby” mode.• Controller will adjust and “VIBRATION LEVEL” display willrevert back to (almost) zero.• Wait for display to stabilize (never goes back to zero).8. Press “Select Mode” once to select “Operate” mode.• After time delay, shaker will begin to vibrate and “VIBRATIONLEVEL” display will indicate the corresponding vibration levelin Volts/g (Peak, not Peak-Peak).9. To stop, Press “Select Mode” once to select “Inhibit” mode — Systemwill stop immediately148C. Ethics ApprovalCertificate149Appendix C. Ethics Approval Certificate  The University of British ColumbiaOffice of Research ServicesClinical Research Ethics Board – Room 210, 828 West 10th Avenue,Vancouver, BC V5Z 1L8  P RI NCI PAL  I NVE ST I G AT OR: I NS T I T UT I ON /  DE PART M ENT : UBC CREB NUM BE R:Bonita Sawatzky UBC/Medicine, Facultyof/Orthopaedics H07-01385I NS T I T UT I ON(S ) W HERE RE SEARCH W I L L  BE  CARRI ED OUT :I n st i t u t i o n S i t eVancouver Coastal Health (VCHRI/VCHA) GF Strong Rehabilitation CentreOt h er l o cati o n s w h ere th e research  w i l l  b e co n d u ct ed :N/A CO- I NVES T I G AT OR( S) :Peter CriptonHeather Finlayson  S PONSORI NG  AG E NCI E S:- Natural Sciences and Engineering Research Council of Canada (NSERC) - "Wheelchair vibration:origin, implications, and reduction for manual wheelchairs" P ROJECT  T I T L E :Whole Body Vibration Induced Lower Limb Muscle ActivityT HE CURRENT  UBC CREB APP ROVAL  F OR T H I S ST UDY EXP I RES:  S ep t emb er 23,  2009Th e  UBC  Cl i n i cal  Research  E th i cs  Bo ard  Ch ai r  o r  Asso ci at e  Ch ai r,  has reviewed  the abovedescribed research project, including associated documentation noted below, and finds the researchproject  acceptable on  ethical  grounds for  research  involving  human subjects and  hereby grantsapproval. DOCUM E NT S  I NCL UDED I N T HI S APP ROVAL : AP PROVAL  DAT E:Do cu men t  Name Versi o n DateP ro to co l :WBV-Spasticity Pilot study protocol v.1 August 7,2008Co n sen t F o rms:consent form v.2 v2 September11, 2008Sep t emb er 23,  20081 of 2 30/Mar/10 12:02150Appendix C. Ethics Approval CertificateAd verti semen ts:WBV-flyer v.1 August 7,2008Qu est i o n n ai re,  Qu est i o n n ai re Co ver L ett er,  Test s:General form v2 September11, 2008Spasticity Scale Form 1 August 1,2008MAS v2 September11, 2008  CERTIFICATION:I n  resp ect  o f  cl i n i cal  t ri al s:1. The membership of this Research Ethics Board complies with the membership requirements forResearch Ethics Boards defined in Division 5 of the Food and Drug Regulations.2. The Research Ethics Board carries out its functions in a manner consistent with Good ClinicalPractices.3. This Research Ethics Board has reviewed and approved the clinical trial protocol and informedconsent form for the trial which is to be conducted by the qualified investigator named above at thespecified clinical trial site. This approval and the views of this Research Ethics Board have beendocumented in writing. The documentation included for the above-named project has been reviewed by the UBC CREB,and the research study, as presented in the documentation, was found to be acceptable on ethicalgrounds for research involving human subjects and was approved by the UBC CREB. Approval of the Clinical Research Ethics Board by one of:            Dr.  C aro n  St rah l en d o rf ,Asso ci ate Ch ai r  2 of 2 30/Mar/10 12:02151D. CWBV Study ConsentForm152Appendix D. CWBV Study Consent FormT H E  U N I V E R S I T Y  O F  B R I T I S H  C O L U M B I A   Second version: 11 September 2008 Page 1 of 3  11 Septem ber  2008       Principal Investigator:   ƌ.ŽŶŶŝĞ^ĂǁĂƚnjŬLJ͕ĞƉĂƌƚŵĞŶƚŽĨKƌƚŚŽƉĂĞĚŝĐƐ͕ŚŝůĚƌĞŶ͛Ɛ,ŽƐƉŝƚĂů  Co -Investigators :          Dr. Peter Cripto n,  De partme nts o f Mec hanic al E ng ine eri ng an d O rtho paed ic s,  UBC             2054  ʹ 6250 App lied S cience Lane,  Va ncouver,  BC ,  V6T  1Z4, Tel: 604- 822 -6629          Dr. H eather Fi nlayso n,  GF  Str ong  Rehabilita tion C entr e,  4255 Laurel St. , (604) 734-1313 ext. 2304   Whol e B ody Vibration In duce d Low er Limb Muscle Activity - A Pilo t Study   A pilo t study  o f t he relatio nship betw een w ho le bo dy vibratio ns and lo wer lim b muscle activ ity  in po pulatio ns with spinal co rd injury  Consent Form – Subject ( Age: 19 years  and older )  Background:   65 - 78% o f individuals with chro nic Spinal co rd injury have sy m pto ms o f spasticity  which c an im pair functio n and quality o f life. It has been repo rted t hat whee lchair v ibratio n m ay  trigger lowe r lim b m uscle spasm s in po pulatio ns with spinal co rd injury.  Our m ain go al in this st udy is to underst and whet her the who le bo dy  v i bratio ns experienced by  m anual wheelchair users w ith spinal co rd injury  c an t rigger invo luntary  m uscle activ ity  in the lowe r lim bs.   We  will be using a v ibrating seat  (a re gular w heelchair seat  sitt ing ato p a v ibrating dev ice) , show n in Figure 1, to apply  co ntro lled v ibratio ns to  o ur participants.     The seat  will v ibrate fo r 10 perio ds o f o ne m inute, with rest  periods betw een each v ibratio n perio d.  We will ƉůĂĐĞƌĞĐĞƉƚŽƌƐŽŶƚŚĞƉĂƌƚŝĐŝƉĂŶƚ͛ƐůĞŐŵƵƐĐůĞƐƚŚĂƚǁŝůůƌĞĐŽƌĚŚŝƐͬŚĞƌŵƵƐĐůĞactiv ity. We are  ho ping to  det erm ine whet her who le bo dy  v ibratio ns can trigger m uscle activ ity  in peo ple w ith chronic spinal co rd injury.  Injury Biomechanics Laboratory Department of Mechanical Engineering 2054-6250 Applied Science Lane Vancouver, BC, V6T 1Z4  Tel: 604.822.3131  Fax: 604.822.2684 Tel: (604) 822-2781 Fax: (604) 822-2403 www.mech.ubc.ca/~injury/index.html  Figure 1 – Vibrating Seat 153Appendix D. CWBV Study Consent FormT H E  U N I V E R S I T Y  O F  B R I T I S H  C O L U M B I A   Second version: 11 September 2008 Page 2 of 3  Invitation to Participate:   Yo u are being invited to  participate  in this st udy because yo u are o ver the age o f 19 years, hav e  ha d a  spinal co rd injury fo r ov er 1 year , use a m anual whee lchair and e xperience m uscle spasm s in y o ur legs. If y o u have an existing pressure so re o r hav e had o ne within t he past  year, a histo ry  o f cardio-vascular o r pulm o nary  disease , y o u will be ineligible fo r the st udy.    Purpose:  The purpo se o f this st udy  is to  underst and  whet her who le bo dy  v ibratio ns, like tho se ex perienced by m anual wheelchair users, can trigger inv o luntary  m uscle activ ity  in peo ple w ith chro nic spinal co rd injury.  This knowledge will hel p us to  understand the relatio nship bet ween who le bo dy  v ibratio n and m uscle spasticity , as we ll as design whee lchair sy stem s that avo id tho se v ibratio ns that m ay trigger spasms.   Study Procedure:  The time required fo r participatio n is 1 ho ur  o n tw o  separ ate  days  (2 ho urs to tal).   This includes a prelim inary m eet ing to  discuss the pro cedure and m ake sure y o u fully  underst and the study  (15 minutes), transfer to test seat and sy stem preparatio n (10 -15 minute s) and t he te st itself, w hich w ill tak e appro xim ate ly 20 minutes.    Yo u w ill be able to  ask quest io ns freely  abo ut any aspect  o f the study  at any tim e . Y o u will be able to  take breaks o r halt participatio n at  any time .    The initial scree ning will inv o lv e a m edical do cto r  asking y o u abo ut yo ur medical hi sto ry  and ex am ining y o ur lowe r lim b reflexes .  We will then measure yo ur, we ight.    The te st will inv o lv e transferring to  a wheelchair seat that is sits upo n a dev ice that will cause the seat  to v ibrate. Recepto rs that m easure m uscle activ ity  will be  attac hed  to y o ur legs  with special stickers  designed for att aching to  skin .  The recepto rs will also  be co nnected by  wire to  a dev ice that reco rds m uscle activ ity. We will v ibrate the seat  at a co nstant frequency  fo r 1 minute  and reco rd the respo nse o f yo ur m us cles to  this v ibratio n. There will be at least a 1 minute  rest  perio d with no v ibratio n. We will then v ibrate the seat at a different frequency  fo r 1 minute and reco rd the respo nse o f y o ur m uscles to  this v ibratio n. This will be repeat ed fo r 10 different f requencies.  If, at any  tim e, y o u are uncom fo rtable with the procedure, o r the test, we will sto p the te st im mediate ly.  We  will need to  att ach the sticke rs direct ly to  yo ur skin, and will  nee d y o u to we ar sho rt pants .   Risks:  There are no  risks asso ciate d with participatio n in this study. The v ibratio n that y ou will be ex po sed to  during this study  is acco rdance with inte rnatio nal standards on who le bo dy v ibratio n ( ISO 2631-1) and is belo w the le vels o f v ibratio n t hat y o u e xperience while using y o ur m anual whee lchair in yo ur daily  act iv ities.  Benefits  and Remuneration :  There are no  direct o r im mediate benefits  to yo u for participating in this study.    D ata co llecte d from  y o ur participatio n in this study  wil l be analyz ed to  determ ine whet her who le bo dy v i bratio n can trigger inv o luntary  m uscle activ ity  in peo ple with spinal co rd injury.  This knowledge will be used to  understand the relatio nship bet ween wheelchair v ibratio n and m uscle spasticity , and help define  wheelchair design  standards in an effo rt to  a vo id spasm  inducing v ibratio ns.   Yo u will rece iv e $25 as com pensatio n fo r yo ur time and t rav el ex penses to  participate  in this st udy.   154Appendix D. CWBV Study Consent FormT H E  U N I V E R S I T Y  O F  B R I T I S H  C O L U M B I A   Second version: 11 September 2008 Page 3 of 3   Withdrawal from Study: Sho uld yo u decline c o nsent, we  will immediate ly  w ithdraw yo ur inv o lv em ent from  the study  and yo u do  not have to  pro v ide any reaso ns fo r y o ur decisio n .  This will no t affect any future medical treatm ent o r care.    Confidentiality: Yo ur co nfidentiality  will be respected.  T here is no  way  to  identify  y o u based o n the digitally  reco rded data.   No  info rm atio n that disclo ses y o ur identity  will be released o r published witho ut y o ur specific co nsent to  the disclo sure.  Ho wev er, research reco rds and m edical reco rds identify ing y o u m ay  be inspected in the prese nce o f the Inv estigato rs o r their designate  by  representatives o f Health Canada, and the UBC Rese arch Et hics Bo ard fo r the purpo se o f m o nito ring the research.  Ho wever, no reco rds which identify you by  name o r initials w ill be allowe d to  leav e the Invest igato rs' o ffices.   Legal Rights:  Signing this co nse nt fo rm  in no  way  lim its yo ur legal rights against the sponso r, inv estigato rs, o r  anyo ne e lse.   Contact:  If y o u have any quest io ns o r desire further info rm atio n with respect to t his st udy, please co ntact the study P rincipal Inv est igato r, D r. Bonita Sawat zky  at 604 .875 .2345 ex t.  7274   If y o u hav e any  co ncerns o r quest io ns abo ut y o ur treatm ent o r rights as a research subject, y o u m ay co ntact Research Subjec t Info rm atio n L ine at t he Unive rsity  Of British Co lum bia Office O f  Research Serv ices at 604 -822 -8598.   Consent: I underst and that participatio n in this study  is entirely vo luntary  and that I m ay refuse to  participate at any  tim e witho ut any co nsequences to  my  co ntinuing m edical care. I hav e received a signed and date d co py  o f this co nsent fo rm  fo r my  o wn m edical reco rds. I co nsent to  participate  in this st udy.    Subject Consent:  __________________        _______________________  ________________    (Subject Signature)                 (Printed Name)             (Date)   Witness:  __________________          _______________________              ________________    (Witness Signature)                     (Printed Name)                                                (Date)         Investigator:  ___________________         _______________________              ________________ (Investigator’s signature)      (Printed Name)              (Date) 155E. DASC StudyQuestionnaire156Appendix E. DASC Study Questionnaire   157Appendix E. DASC Study Questionnaire    158F. CWBV StudyRecruitment Flyer159Appendix F. CWBV Study Recruitment Flyer   PARTIC I PANTS   REQ U IRE D  For  a study on the relationship between Whole Body Vibration and muscle activity in people with Spinal Cord Injury.  Participants will be asked to sit in a vibrating seat while muscle activity is recorded.  Participation will req uire approx . 2 hours.  If you :   H ave ha d a spinal cord injury for over 1 year   E xperience spasticity   D o not have pressure sores  (on buttocks)  or an active  cardiovas cular or pulmonary disease   And would li ke to participate in this study, p lease contact:  Dr. B onnie Sawatzky:   (604) 875 - 2345 ext. 7274 or Dr. Heather Finlayson: (604) 734 - 1313 ext. 2304                                                    heather.finlayson@vch.ca   YOUR PARTICIPATION W ILL HELP US UNDER STAND  THE RELATIONSHIP BET WEEN   WHEEL CHAIR VIBRATION  &  MUSCLE SPASTICITY     160

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0070931/manifest

Comment

Related Items