the evolution of clinical gait analysis part ii kinematics

Gait and Posture 16 (2002) 159–179

The evolution of clinical gait analysisPart II Kinematics�

D.H. Sutherland *

Children’s Hospital San Diego, 3020 Children’s Way MC 5054, San Diego, CA, USA 92123-4282

Accepted 18 December 2001

Abstract

Kinematics is treated as a single topic in this manuscript and the emphasis is on early history, just as it was in Part I,Electromyography. Needless to say, neither kinematics nor electromyography, nor kinetics and energy (the latter to be includedin Part III) are stand-alone components of clinical gait analysis. The only reason for this selective format is that it lessens my taskto be able to write about one subject at a time. One of the consequences of this arbitrary separation is that some contributors,who have enriched more than one portion of clinical gait analysis, are highlighted only in the area in which they have contributedthe most. I began with Kinesiological Electromyography in Part I because the earliest stirrings of the dream of clinical gaitanalysis were expressed in the development of KEMG (kinesiological electromyography). The early investigators realized that verylittle could be said about the dynamic action of muscles without KEMG. Next, in chronological order, came kinematics. I havebeen an active participant and eyewitness, and take full responsibility for attempting to write an early history at a time when mostof the contributors are still alive. Ordinarily, history is written much later, in order to fully grasp the significance of individualcontributions in the tapestry of the whole. As stated in Part I, Electromyography, the emphasis has been placed on the earlyhistory. The application of motion analysis to sports medicine, and sports medicine functional analysis, is covered only lightlyhere, and this should not be interpreted as minimizing its importance. The literature on this subject is now quite voluminous andit would not be possible to cover it adequately in this manuscript. Later historical writings may differ significantly and willhopefully give more recognition to pioneers in later generations: those physicians, engineers, physical therapists and kinesiologistswho are lifting the level of clinical gait analysis and directing their energies in expanding clinical directions. It is hoped that thismanuscript will prompt additional manuscripts, as well as letters to the editor of Gait and Posture on the content of this reviewpaper. © 2002 Published by Elsevier Science B.V.

Keywords: History; Kinematics; Clinical gait analysis

www.elsevier.com/locate/gaitpost

1. Introduction

Accurate measurement of motion is central in anyscientific method of gait analysis. Measurements ofindividual joint angular rotations, as well as transla-tions of segments and of whole body mass, allow thecomparisons with normal that are necessary to distin-guish pathological from normal gait. Complex hard-ware and software are necessary to accomplish this taskwith accuracy and reliability. This component of clini-cal gait analysis has proven to be very challenging andthe evolutionary process continues to this day.

The individual joint angles and the displacements ofsegments and of the whole body mass were recognizedto be essential measurement requirements in the late1800s by Braun and Fischer [1–5]. Their clever ap-

� Note from re�iew editor: This article is the second in a series ofthree historical narratives that Dr Sutherland has very kindly agreedto author for Gait and Posture. As Dr Sutherland indicated in hisabstract for Part I, these are very personal accounts that focusprimarily, although not exclusively, on the early history of clinicalmotion analysis. He further acknowledged that not all importantcontributors or events may be chronicled or weighted in the samemanner as others might have done. Still, these accounts are extremelyvaluable because they provide a very alive ‘behind the scenes’ view ofhow our field has progressed over the years as told by one of its truepioneers, with a richness that could never be captured by a merelisting of names or documented events.

* Tel.: +1-858-966-5807; fax: +1-858-966-7494E-mail address: [email protected] (D.H. Sutherland).

0966-6362/02/$ - see front matter © 2002 Published by Elsevier Science B.V.PII: S0966 -6362 (02 )00004 -8

mailto:[email protected]

D.H. Sutherland / Gait and Posture 16 (2002) 159–179160

proach to kinematic analysis was to apply Geisslertubes to the limb segments, interrupt the illumination atregular intervals by a large tuning fork, and pho-tograph the subject walking in total darkness with fourcameras while the lenses were open. One camera waspositioned in front of the subject, one behind, and oneon each side, making their measurements tri-dimen-sional. The subjects were protected from electricalshock by wearing rubber suits resembling wet suits. Theprocess of collecting data required 8 or 10 hours persubject and then it involved months of work to reducethe data and calculate kinematic measurements. Thiswas a fantastic scientific achievement, however, becauseit was so time consuming, Braun and Fischer’s methodcould only be applied in gait research.

One of the methods used by Eberhardt and Inman [6]in the 1940s also included the use of interrupted light.A photograph was obtained with the subject walking infront of the open lens of a camera while carrying smalllight bulbs located at the hip, knee, ankle and foot. Aslotted disk was rotated in front of the camera, produc-ing a series of white dots at equal time intervals. Thesedots could be laboriously connected to provide jointangles that could be manually measured. Again, thiswas a slow and labor-intensive process, not suitable forclinical application. In order to examine transverseplane rotations, Vern Inman, MD, PhD, drilled pinsinto the pelvis, femur, and tibia, and recorded pinrotation with the aid of a movie camera located abovethe subject [7]. One of his subjects, David Chadwick,MD, then a student at the University of California,Berkeley, later became the Medical Director of Chil-dren’s Hospital of San Diego. He described his experi-ence as ‘very painful’, something he would not haveagreed to had he understood ‘what it would be like’.Needless to say, this technique gained very few follow-ers, although there has been some recent use of pinsinserted into bones in normal subjects for a differentpurpose, i.e. to determine the difference between move-ment of markers taped to the skin surface and thoseplaced into the skeleton.

2. Early pioneers and techniques (post Inman)

2.1. Strobe light, reflecti�e strips and manualgoniometer

Mary Pat Murray, PhD, working at the Veteran’sAdministration Hospital in Milwaukee, Wisconsin, de-vised a simple, effective, and low cost way to recordand measure movements. She and her team attachedreflective targets (including reflective strips in the lowerextremity) to specific anatomic landmarks and the sub-jects walked in the illumination of a strobe light. Theresultant photograph was used to make measurements

of the individual segments. Her method did includeupper extremity and trunk markers, as well as pelvisand lower extremity. She successfully used this methodto produce landmark articles in the 1960s, 70s and 80soutlining the walking patterns, first of normal men [8],then of normal women [9], and then patients withpathological conditions [10–12]. Although, viewed bytoday’s standards, this appeared to be a crude method,the sagittal plane joint angle measurements of normalsubjects in her publications are very similar to thoseobtained with current technology, (see Fig. 1) [9]. Theprimary problems with Dr Murray’s method were theneed for manual measurements of all the joint anglesand the inherent difficulty with the method in providinghip, knee, and ankle joint rotations in the transverseplane.

2.2. Electrogoniometry

There was a flurry of enthusiasm for recording jointangles with electrogoniometers. The Karpovich broth-ers were early contributors who used goniometers torecord joint angles. Their reason for using electrogo-

Fig. 1. Sagittal measurements of pelvis, hip, knee and ankle in normalwomen. Reprinted from the Archives of Physical Medicine andRehabilitation, with permission from W.B. Saunders Company [9].

D.H. Sutherland / Gait and Posture 16 (2002) 159–179 161

Fig. 2. Triaxial Goniometer as applied to a subject for bilateral hipand ankle joint motion analysis. Reprinted from the Journal ofBiomechanics, vol. 13, 1980, pp. 989–1006, Chao: ‘Justification oftriaxial goniometer for the measurement of joint rotation ’; with permis-sion from Elsevier Science [22].

the necessity for matching the size of the individualwith the appropriate goniometers, the offset of therecording device to the side of the limb segments, andthe inability to obtain simultaneous measurements ofall of the moving segments. Quoting from Dr Chao’sarticle entitled, Justification of Triaxial Goniometer forthe Measurement of Joint Rotation, ‘‘This paper at-tempts to provide the theoretical and experimental jus-tifications of the existing triaxial goniometer design sothat these potential problems can be resolved’’ [22]. Theexperimental device used for the justification was amechanical model, thus any problems with skin motionwere not considered. Quoting further, in the ‘Discus-sions’ section of the article, Dr Chao states, ‘‘Althoughthe triaxial goniometer is the only instrument that canprovide instantaneous angular motion of a joint inthree dimensions, its user must realize the potentialdrawbacks of the method in order to avoid unnecessarycomplications. First of all, the external attachment ofthe device could introduce error in the data due torelative movement of the underlying soft tissues’’. Chaogoes on to mention other critical points to consider inthe use of the triaxial goniometers relating to align-ment, lateral projection, and the weight of the measur-ing device. Although these difficulties have preventedwidespread adoption of electrogoniometers for routineclinical gait analysis, goniometers are effective whenmultiple recordings are required, when studies are beingcarried out outside of a motion analysis laboratory, andwhen sagittal movements are sufficient for data acquisi-tion. A final objection yet remains: moment studiescannot be made without the measurements of the posi-tion of joint centers in space, something that goniome-ters do not provide.

2.3. Cine film and passi�e marker systems with manualentry of marker positions

2.3.1. Vanguard Motion AnalyzerOther investigators concentrated on developing pho-

tographic techniques for gait analysis. Photographicmethods have a key advantage over other techniques inthat the whole body can be included and the relation-ship of each extremity and the trunk can be simulta-neously viewed. The opportunities for measurement arethus greatly expanded over prior techniques. A furtheradvantage is that individuals of all sizes are suitable forclinical analysis. Initially, however, there wereformidable obstacles, including the need for excessivetime spent in reducing the data and the absence ofcomputer availability for storing data and performingvoluminous mathematical computations.

While casually scanning a technical journal, my eyesfocused on an advertisement for a Vanguard MotionAnalyzer. The very name was intriguing and its capabil-ity for projection of movie film on a backlit screen for

niometers was that many gait cycles could be collectedquickly, and analog graphs of motion could be dis-played, without the need for data reduction by hand[13]. In 1976, Bajd et al. [14] published an articledescribing online electrogoniometric gait analysis usingsix precision potentiometers, giving time-dependent an-gles in hip, knee, and ankle of both legs in the sagittalplane. Their reasons for choosing this method of instru-mentation were that it was suitable for online process-ing of measured data, and was simple, reliable andinexpensive. Other important contributors to electrogo-niometry are McLeod [15], Tata [16], Johnston andSmidt [17], Lamoreux [18], Kinzel et al. [19] andTownsend et al. [20]. Foort presented the electronicrecording of joint function with analog recordings ofthree-dimensional hip, knee, and ankle joint motion[21] at a workshop on Human Locomotion and ClinicalAnalysis of Gait in Philadelphia, in 1976. Edmund Y.S.Chao, PhD published a report in 1980 on the design ofa triaxial goniometer, based on the gyroscope conceptutilizing Eulerian angles in the computation of themeasurements (see Fig. 2) [22]. Again, this did not gainwide acceptance, arguably because of the difficulty inpreventing cross talk from the three motion axes. Ananecdotal description by Jacquelin Perry (unpublished)of significant motion recorded in the hip joint of apatient with a solid hip fusion did not help promoteadoption of this method. At first glance, the goniomet-ric method holds great appeal. However, with thetremendous range of height and weight of subjects, andthe difficulty for small subjects to walk comfortablywith this amount of hardware, widespread adoption ofthis technique never occurred. The difficulties encoun-tered with the use of goniometers, then and now, are


easy frame-by-frame viewing, and measurement of se-lected points with x and y coordinates, was most ap-pealing. Dick Freeborg, later President ofInstrumentation Marketing Corporation and now aVice-President in the Kodak Corporation, suggestedthat I contact Ray Linder at Lockheed Aircraft Corpo-ration Missiles and Space Company, Sunnyvale, Cali-fornia. Ray Linder was the leader of a section chargedwith making measurements of machines, rocket trajec-tories, etc. In 1965, he published a description of themethods that he and his team had developed to mea-sure pitch, yaw, and roll, using mathematical formulae,two or more cameras, and a two-dimensional coordi-nate system of measurements [23]. After a telephonecall and a letter from me, Ray Linder invited me tocome to his section during their lunch hour and explainthe need for human gait measurements. Roger Mann,MD, accompanied me. Ray Linder made a prescientcomment after our presentation to the group, ‘‘Youmean you would like to measure the movements of theskeleton from surface markers with skin movementsconfounding the interpretation. Is not that like tryingto measure the movements of a broomstick within agunny sack?’’ Nothing daunted, we were fascinated tosee about 17 Vanguard Motion Analyzers in one room.We were introduced to an interested and bright groupof people employed at the task of making measure-ments of pitch, yaw and roll. Their level of interest inour project was very exciting. Out of this contact, JohnHagy and Richard Oyama came forward as volunteersto the Shriners Hospital in San Francisco, bringingwith them high-speed movie cameras, generouslyloaned by Lockheed Missiles and Space Company. In arelatively short time, Hagy, Oyama, and Keller helpedus establish a system to add kinematics to the elec-tromyography already in clinical use in our laboratory.Hagy assisted during evenings and weekends until wewere able to persuade him to come full-time in April of1971. John Hagy and Cecil Keller, also a Lockheedemployee, assisted in the development of our movementmeasurement system, first reported in 1967 in the arti-cle, ‘Measurement of Movements and Timing of Mus-cle Contraction from Movie Film’ [24]. Initially, ourmethods of computation were very time consuming.After recording x and y coordinate measurements fromthe cine film displayed on the Vanguard Motion Ana-lyzer, we used a slide rule to perform the trigonometriccomputations. Later, an optical encoder was added toreplace the necessity of manually recording the x and ycoordinates. This task was further simplified by utiliz-ing a sonic digitizer to input the data (see Fig. 3) [25].

Significant progress in time reduction came when acomputer was added to store and perform the mathe-matical calculations, (see description of our first use ofa dedicated computer by Electronic Processors, Inc. inAppendix C). We used this method for many years, but

hand digitizing continued to be an obstacle. Individualswere trained and employed to do the digitizing. Theywere enthusiastic, but in time became bored with therepetitious nature of their task. The average length oftime for a technician to be employed in our laboratorywas 2–4 years; many used this opportunity as a step-ping stone in their career.

We made use of these gait analysis data to providetreatment recommendations and to study the outcomeof treatment intervention. Some of the papers from ourlaboratory were on the subjects of crouch gait [26], gaitanalysis in cerebral palsy [27], the role of the ankleplantar flexors [28], and the pathomechanics of gait inDuchenne muscular dystrophy [29]. When our methodsand results were given, many were enthusiastic aboutthe possibilities for further development of this emerg-ing application of science, but very few were willing toundertake such a labor-intensive effort. The real break-through, needed to bring about widespread adoption ofthree-dimensional movement measurements for clinicalgait analysis, was yet to come when the measurementscould be automated. In fact, had there been no furtherimprovements in technology, clinical motion analysismight have continued in only a few locations. The timeand energy required to digitize film prevented manycenters from using gait analysis as a clinical tool. Inspite of the difficulties, the stage was set forautomation.

2.4. A turning point! Fully automated mo�ementmeasurements

2.4.1. European contributors and techniquesEngineers and physicists deserve full credit for devel-

oping methods to automate human movement measure-

Fig. 3. Vanguard motion analyzer and Graf-Pen sonic digitizer usedto determine the X and Y coordinates of the markers shown on theviewing screen. Reprinted with permission from the Journal of Boneand Joint Surgery [25].


Fig. 4. Cat on a treadmill with reflective markers. Reprinted withpermission from Dr Hans Furnee [30].

Measurement Systems: Aspects of Data Acquisition,Signal Processing and Performance’ [36].

J. Paul, PhD, who started his kinematic measurementswith two orthogonal Bolex cine film cameras driven bysynchronous electric motors at 50 Hz and a homemadeground-to-foot force platform, wrote this personal letterto me in response to some questions I had posed:

‘‘As you may imagine, the processing of the cine filmand hand digitizing was an onerous procedure, and Iwas therefore very excited when, in 1967, I saw thepresentation of Furnee, hybrid instrumentation inProsthetics Research Proceedings in the 7th Interna-tional Conference on Medical and Biomedical Engi-neering in Stockholm. Furnee presented his inventionof a single camera television system for 2-D move-ment analysis. As soon as possible thereafter, I gottwo of our research students to develop a 2-camerasystem for 3-D video/computer movement analysis.In 1972, M.O. Jarrett, and B.J. Andrews started theirPhD studies with the remit of developing the system.Between them, they got our system up and runningbased on a PDP-12 computer, which allowed us atotal of 36 analog inputs. Brian Andrews did notcomplete his PhD at that time, but came back to usin 1980 as a member of the staff. At that time, he didfurther development on the system to allow it towork with a PDP-11 computer and took the opportu-nity to try the use of shutters on the camera toimprove definition, but then implemented what Ibelieve was the first application of strobe infraredlighting.’’

‘‘When Jarrett finished his studentship, he was em-ployed jointly by ourselves and George Murdoch atDundee to develop a system for implementation atDundee. He did this, but to my great astonishment,implemented a two-dimensional system there, andJulian Morris became aware of this, and went on todevelop the Oxford Metrics System, which was devel-oped to be three-dimensional. Julian was very sur-prised, at a later date, to find that our first and onlycomputer television movement analysis system hadbeen three-dimensional. Apparently Jarrett, withwhom he was interacting had not told him!’’

‘‘At that time, we were not very assiduous at publica-tion of our work, and the only one which I can citeis, Jarrett, Andrews and Paul, 1976, which is the textof a conference presentation as you will see on mypublication list [37–39]’’.

From the letters of Hans Furnee and J. Paul, andtheir publications, there is an unmistakable sequenceof interweaving paths and shared enthusiasm for au-tomating gait analysis. Hans Furnee led the effort; J. Paul

ments. E.H. Furnee, PhD, Faculty of Applied Physics,Technical University, Delft, The Netherlands, beganaround 1967 to develop TV/motion analysis systems withautomated recording of reflective marker positions. Hisexperiment, capturing multiple joint angles of a catrunning on a treadmill, was a strong portent of the futurewidespread adoption of the photo-electronic method tomeasure the kinematics of human walking. The markersused by Furnee were passive paper disks that werebrightly visible in ultraviolet light, and the ultravioletlight was pulsed to prevent ‘smearing’, (see Figs. 4 and5) [30].

Prior to the first publication directly authored byFurnee [31,32], a student of Furnee, P.C. Steilberg,reported the method more completely as BS and MSthesis work in 1967 and 1968 [33,34]. Furnee is theoriginator of the Primas System, from the MotionStudies Lab, Delft University of Technology. The currentPrimas system emphasizes real-time marker identifica-tion and real-time marker centers by circle fitting (pilot),developed, respectively at Delft University in 1990 and1992. Additional publications authored by Dr Furneeinclude: ‘TV/Computer Motion Analysis Systems: TheFirst Two Decades’ [35] and ‘Opto-Electronic Movement

Fig. 5. Variation of angles at shoulder, elbow, hip and knee of limbsof same side during stepping on treadmill at 2.0 m s−1. Upwardexcursion indicates extension, downward flexion. Reprinted with per-mission from Dr Hans Furnee [30].


was excited by his work and set graduate students,Andrews and Jarrett, to work, with the resultant pro-duction of a 3-D automated video camera system formovement measurement. Prior to his entry into auto-mated movement measurements, J. Paul had alreadybecome an expert in gait analysis, using a digitizedcine-film system for movement measurements. His pub-lications cover a wide range of clinical subjects, with aspecial emphasis on the gait of amputees. A largenumber of students of Dr Paul are spread worldwideand it is fair to say that he has been a central figure inthe development and expansion of clinical motionanalysis.

In response to questions posed to Julian Morris,DPhil, I received a personal correspondence datedApril 13, 1996. The following correspondence is printedin its entirety.

The history of the development of Oxford Metricsand VICON is as follows:

‘‘I first met Mick Jarrett at an ISPO1973conference in Montreux,Switzerland, while he and I wereboth graduate students. He wasworking (together with BrianAndrews, now in Edmonton,Alberta) on a 2-D TV computersystem and I was usingaccelerometers for measuring gait.Mick and I both presented ourdoctoral theses later the same year,his to Strathclyde University andmine to Oxford.Mick was by now working with1974–75David Condie in University ofDundee, and I was working at theNuffield Orthopaedic Centre inUniversity of Oxford. We bothwanted to set up automated,non-cine gait analysis facilities inour respective hospitals.I arranged for an Oxfordengineering friend of mine, MalcolmHerring, who was a far moreexperienced electronics designer thaneither of us, to redesign Mick’sprototype from Strathclyde. Three ofthese (multi-camera but still 2-D)systems were built—one forStrathclyde, one for Dundee, andone for Oxford.Having left the Nuffield Orthopaedic1979in 1977 (Mike Whittle took over myjob), I was now Technical Director

of Oxford Medical Systems, part ofthe Oxford instruments group.Although our main product rangewas for cardiology (Holtermonitors), I believed that there wasa commercial market for anautomatic 3-D gait analysis system.I hired Malcolm Herring (seeabove), Graham Klyne, and AnnabelMacLeod as the development team,and made a licensing deal withUniversities of Strathclyde, Dundee,and Mick Jarrett personally, for theexisting technology. Over the next 2years, the VICON (the name derivesfrom video-converter) developmentteam redesigned the hardware andwrote the first 3-D photogrammetrysoftware applied to this field. Thefirst system was shipped to EricRadin, MD, in West Virginia in1980, and the second to SheldonSimon, MD, Boston Children’sHospital, a few months later.The Oxford Instruments Group1983–84decided to float on the LondonStock Market and focus on theircore businesses, primarily buildingcryogenic magnets for MRI. They‘spun-off’ the biomechanics business,by then called Oxford Dynamics,which was sold to myself and othermembers of the VICONdevelopment team. We renamed thecompany, Oxford Metrics Ltd.(Graham Klyne won a bottle ofwine for the name).During the 70s I was aware, throughpublications, of the work of HansFurnee and David Winter. However,I do not think I first met either ofthem until after VICON waslaunched. To my knowledge, Hansdid not ever spend time working inOxford.The connection between Hans andVICON is largely technical, ratherthan historical. The hardware tech-nology used by VICON is basicallythe same as that developed by Hans.Whether either one derived from theother is hard to say. Certainly, Ibelieve that Hans published his earlywork and may have been visited inHolland by Mick Jarrett while thelatter was a graduate student.’’


Author’s comment: In a letter to me, Hans Furneeconfirmed the visit of M.O. Jarrett to his laboratoryand stated that he had freely shared information withJarrett.

Again, back to Julian Morris’ letter:

‘‘Although they both use video, there is minimaltechnical connection between what David Winterpublished and VICON. I certainly believe that Hansand David developed video-based systems a year ortwo ahead of Mick, but as you imply, this is, forsome, a sensitive area! Mick Jarrett and Brian An-drews, I believe, wrote most of the software for the1973 Strathclyde TV system jointly. However, theVICON development team never saw or used any ofit. The first true VICON 3-D software was designedby Graham Klyne and myself, and written entirelyby Graham. We drew on the published ideas ofmany others, including Herman Woltring.’’

Eric Radin, MD, used the technology supplied byVICON to study the running and walking gait ofsheep on a treadmill. In 1983, along with gait col-leagues from San Diego, Ed Biden, and Marilynn Wy-att, I visited Eric Radin’s Laboratory in Morgantown,West Virginia. The Laboratory provided an extraordi-nary scene of stacked bales of straw, sheep in a pen atthe side of the room, and the odor of sheep lying likea pall over all. I asked Dr Radin, ‘Eric, do you doany studies of human patients?’ Eric, in his inimitablestyle, said, ‘‘Well, yes, of course, but we only letpeople in on Fridays’’. I silently wondered how wellhuman subjects responded to being studied in thisenvironment! Following this visit, we purchased VI-CON hardware and Ed Biden, DPhil, wrote customsoftware in 1984 for clinical application in our SanDiego Gait Laboratory. After a period of comparisonof studies on normal individuals with cine-film digi-tization, and data collected on the same individualswith the VICON hardware and Ed Biden’s software,we made the transition from film digitization to auto-mated data capture. The initial problems were associ-ated with the difficulty of identifying and trackingmarkers; this was initially done in two dimensions. Atremendous move forward occurred with the contribu-tion of Andrew Dainis, who wrote three-dimensionaltracking software (details to follow later).

Michael Whittle, MD, PhD, spent 2 years doingsurgical research after internship, which led to a mas-ter’s degree in biomechanics. As a research medicalofficer in the Royal Air Force, he was loaned for 3years to NASA in Houston to supervise the muscu-loskeletal experiments on the Skylab Space Station.One of these experiments was on the 3-D measure-ments of the astronaut’s body form [40], and this

became the subject of his PhD dissertation, which heobtained after he returned to Great Britain. He tookover the directorship of the Motion Laboratory atOxford after Julian Morris left to found Oxford Met-rics. The only software available for the new OxfordMetrics system was for data capture, so Dr Whittlewrote full 3-D motion capture software, ‘‘So, in effect,we had the first 3-D TV computer system in theworld’’. (Author’s comment: Communication from J.Paul indicates earlier 3-D development in Strathclyde.)Michael Whittle now holds the Cline Chair of Reha-bilitation Technology at the University of Tennessee atChattanooga. He is author of a book entitled, ‘GaitAnalysis: An Introduction’, which is now in its 2ndedition [41].

Another important player in the exciting world ofmotion capture is the Bioengineering Technology Sys-tems, or BTS, which is home-based in Milan, Italy.BTS traces its origins to the contribution of the bio-engineering center of the Pro Juventute Foundationand the Politecnico di Milano. The company wasformed in 1986. The engineering contributions of Fer-rigno, Pedotti, and Cappozzo were key in the develop-ment of the ELITE System [42–44]. BTS rapidlyexpanded into the complete world of clinical gait anal-ysis, combining kinematics, kinetics, and electromyo-graphy in a robust, all-inclusive approach to clinicalgait analysis and research in motor skeletal function.In point of time, BTS entered the field of clinicalmotion analysis after Oxford Metrics, Inc. and MotionAnalysis Corporation.

The story does not stop here. The entry of manynew companies and new systems of motion captureattests to the enduring fascination with movementanalysis. The competition between the companies nowmarketing motion capture systems has resulted inmore rapid processing of information, new methods ofdisplaying the data, and a surging interest in clinicalgait analysis. Laboratories are now available in all ofthe developed countries and in many of the developingcountries.

2.5. Automated mo�ement measurements

2.5.1. North American early contributors andtechniques

While the early advances in gait analysis and auto-mated movement measurements were occurring in theNetherlands, England, and Scotland, exciting activitieswere also taking place in Canada. Robert K. Green-law, MD, then the Chief Surgeon of the Shriners Hos-pital, Winnipeg, Canada, and a former resident in theShriners Hospital in San Francisco, was aware of DrInman’s work and of my work in the San FranciscoShriners Hospital Gait Laboratory. It was not obvious


Fig. 6. Gait studies data acquisition system. ‘Locomotion studies as an aid in clinical assessment of childhood gait ’—reprinted from, by permissionof the publisher, CMAJ, 1975; 112 (9), pp. 1091–www.cma.ca [46].

to me then that orthopedic resident Greenlaw had aninterest in motion analysis, but the events that followedproved otherwise. He had the good judgment to askDavid Winter, PhD, to design the first gait analysislaboratory in the Shriners system in Canada.

David Winter, PhD, then with the Department ofElectrical Engineering and the Department of Surgery,University of Manitoba, Winnipeg, designed a GaitLaboratory for the Winnipeg Shriners Hospital. Dinnet al. submitted a manuscript in 1969 entitled, ‘Com-puter Interface for Television’, which was ultimatelypublished in IEEE Transaction of Computers in 1970[45]. In a letter to me, Dr Winter states, ‘‘Our firstroutine use of the interface began early in 1970 at theShriners Hospital in Winnipeg. For the 5 years I wasthere it was the backbone of all our kinematic datacollection. Actually, it goes back to the time of Furnee.Our first operational TV interface was reported atbiomedical engineering conferences on this side of theAtlantic at the same time (unknown to us) as he wasreporting in The Netherlands’’, (see Fig. 6), [46]. Winteret al. subsequently published an article entitled: ‘Televi-sion–Computer Analysis of Kinematics of HumanGait’ in Computer and Biomedical Research in 1972 [47].Winter et al. published an important study of thekinematics of normal locomotion in the Journal ofBiomechanics in 1974 [48]. Doctor Winter’s publicationssince that time are legendary [49]. He has profoundlyinfluenced the course of clinical gait analysis throughhis scientific studies, teaching, mentoring of many grad-uate students and his many publications. His bookentitled, ‘Biomechanics and Motor Control of HumanMovement’, now in its 2nd edition, is a ‘must read’ forall who are interested in gait analysis [50].

Sheldon Simon, MD, obtained his medical degreefrom Harvard and his residency training was completedin the Harvard Training Program. During his residencytraining, he worked in the laboratory of Robert Mann,PhD, at the Massachusetts Institute of Technology. Hereceived a Cave Traveling Fellowship and, in Januarythrough June 1974, divided his time between the SanFrancisco Shriners Gait Analysis Laboratory, then di-rected by Roger Mann, MD, and the PathokinesiologyLaboratory at Rancho Los Amigos, directed byJacquelin Perry, MD. He says that he was very im-pressed by the electromyographic studies at Ranchoand the force and kinematic studies in San Francisco.By the time of his return, the Kistler force plate wasavailable as a commercial product, but the price wassteep for his budget in Boston. He persuaded WaltSynutis, President and CEO of AMTI, to design a newstrain-gauge force plate at a price within his budget.The space available for him in Boston was small and herealized that a three-dimensional coordinate softwaresystem for kinematic measurements should be devel-oped. The laboratory opened in Boston Children’s Hos-pital in September 1974, with two AMTI strain gaugeforce plates, a three-camera 16 mm high speed camerasystem, a Vanguard Motion Analyzer and a computerwhich stored the measurements from force plates andelectromyograph simultaneously. The software for kine-matic measurements utilizing a three-dimensional coor-dinate system from film was conceived by Dr Simonand written by Roy Nuzzo, MD, an orthopaedic resi-dent at the time [51]. Mick Jarrett, PhD spent a gooddeal of time at Boston Children’s Hospital in 1982perfecting software for clinical application of the VI-CON system. According to Dr Simon, initially it took

www.cma.ca


just as long to process 3-D VICON measurements as3-D film measurements. Nonetheless, it was clear thatfurther developments would establish automated videomeasurements and there would be no turning back. In1986, Dr Simon moved to Ohio State University, wherehe held the positions of Chairman of the Department ofOrthopedic Surgery and Medical Director of the newMotion Analysis Laboratory. Numerous clinical andresearch publications have followed [52]. One of DrSimon’s many interests has been the application ofartificial intelligence to gait data. He says that herealized that interpretation would continue to offer thegreatest challenge and, in 1984, began work on anartificial intelligence application. This work has contin-ued to this time and a system, which can be separatelyused as a decision helper and as a trainer, is currentlybeing tested. Dr Simon is the editor of a book entitled,‘Orthopaedic Basic Science’, published by the AmericanAcademy of Orthopaedic Surgeons [53]. He now residesin New York City and continues his career-long interestin clinical gait analysis.

In 1978, James R. Gage, MD, visited Eugene Bleck’sGait Laboratory at Stanford Children’s Hospital,Jacquelin Perry’s laboratory at Rancho Los AmigosHospital, and my laboratory at San Diego Children’sHospital, in preparation for beginning his first labora-tory at Newington Children’s Hospital. United Tech-nology Research Corporation, located in Newington,Connecticut, had offered extraordinary engineering andfinancial support for the establishment of a Gait Labo-ratory. In early 1980, Gage returned to the San DiegoLaboratory for an in-depth look, bringing with himKen Taylor, United Technology Project Engineer, andJim Clark, Manager of the Newington Gait Laboratoryproject. The three men asked many questions, includingwhat we would do if we were starting another labora-tory with optimal funding and full technical assistance.We answered openly, even with ideas that were not yetfully realized in our own laboratory. This cooperationand sharing of information continued during the devel-opment of the Newington Laboratory. The NewingtonGait Laboratory opened in July 1981. Special featuresof this laboratory included synchronization of all gaitdata, full custom clinical software and rapid processingof data. Scott Tashman, MS (now PhD), validated andcontinued the United Technologies software package.Follow-up gait studies were regularly performed onpatients who had undergone preoperative gait studies,thus opening the way for a great many clinical papers[54–59] and a book entitled ‘Gait Analysis in CerebralPalsy’ [60].

Following Dr Gage’s move to Gillette Children’sHospital in 1990, the Newington laboratory continuedunder Peter DeLuca, MD, as Medical Director, withRoy B. Davis, PhD, as Director until 1998. SylviaO� unpuu, M Sc., a prior student of David Winter, is the

current Director. The Newington Gait Laboratorymoved to Hartford Connecticut, with the opening ofChildren’s Hospital of Connecticut as an integral partof the University of Connecticut Medical Center. Manypapers have been published both by Gillette Children’sHospital and Newington/Children’s Hospital of Con-necticut covering a variety of subjects, including: run-ning patterns of normal children [61] [62], outcome ofmultilevel surgery in cerebral palsy [63], stiff-knee gait[64,65], the utility of basing treatment decisions incerebral palsy on preoperative gait analysis, [66], and agait analysis data collection and reduction technique,which includes Davis’ much referenced joint centerdetermination method [67]. Gage has pushed the envel-ope in advocating gait analysis routinely in patientswith cerebral palsy [60]. Some surgeons, none of whomhave gait laboratories, have criticized this. An anecdoteillustrating this point follows:

At a course jointly sponsored by the AmericanAcademy of Orthopaedic Surgeons (AAOS) and thePediatric Orthopaedic Society of North America(POSNA) in San Francisco, May 6, 1990, entitled Con-troversies in the Treatment of Cerebral Palsy, thecourse chairpersons were Dr Michael Sussman and DrWalter Greene, myself, Dr Simon, and Dr Gage, in thatorder. We had just finished giving views on the impor-tance of clinical gait analysis when, from near the backrow, Hugh Watts, MD, a pediatric orthopaedist andfriend, but never one to avoid controversy, rose tochallenge the clinical usefulness of gait analysis. Heclaimed that the gait laboratory setting is not a suitableenvironment for arriving at the true walking patterns ofchildren with cerebral palsy. He implied that observa-tional analysis of children in the playground, or inother familiar surroundings, is better. The entire backsection of the auditorium, filled mostly with ortho-paedists, burst into spontaneous applause. The heateddiscussion that followed resulted in back-to-back edito-rials by Gage and Watts in the Journal of PediatricOrthopaedics [59,68]. Be that as it may, acceptance ofclinical gait analysis has steadily increased and newlaboratories are being established throughout theworld. The new generation of orthopaedic surgeons,introduced to gait analysis in their training, increasinglydemands functional analysis, before and after treat-ment, in order to better understand the magnitude ofthe disability and to ascertain the impact of their inter-vention. As my ‘parting shot over the bow’ on thissubject, I would like to quote Max Planck, the famousGerman physicist, who pioneered modern physics byproposing the quantum theory and won the 1918 NobelPrize. He said, ‘‘An important scientific innovationrarely makes its way by gradually winning over andconverting its opponents. What does happen is that theopponents gradually die out’’ [69].


Murali Kadaba, PhD, became interested in gait anal-ysis when he joined Helen Hayes Hospital as a researchscientist in 1979. Dr George Van Cochran was influen-tial in his decision to work in the area of clinical gaitanalysis. Dr Kadaba states that his interest was inten-sified after a visit with Dr Perry at Rancho Los AmigosPathokinesiology Laboratory in Downey, California,and with me at the Motion Analysis Laboratory at SanDiego Children’s Hospital. He received a NIH grant in1984 to study the reproducibility and reliability of gaitdata [70]. Following completion of this study, he be-came interested in the numerical representation of kine-matic and kinetic data for pattern recognition in spasticdiplegia [71]. The Helen Hayes clinical software wascompleted in 1985. ‘‘This was a cooperative effort; theother team members were H.K. Ramakrishnan, MaryWootten, Janet Burn (Gainey) and Dr Van B.Cochran’’ [72]. This clinical software served a criticalneed for software to be used in a clinical setting. It wasimplemented at the following centers: Richmond Chil-dren’s; Shriners Hospital, Houston Unit; MethodistHospital, Houston, Texas; Children’s Memorial Hospi-tal, Chicago; Children’s Milwaukee; and Shriners Hos-pital, Portland Unit. The Helen Hayes Team, under theleadership of Dr Kadaba, deserve great credit for devel-oping and supporting clinical software (no small task),in the precarious early years. At that time, the writersof commercial software were attuned to the diverseneeds of researchers, but they lacked confidence in theability of clinical laboratories to agree on nomenclatureand formats for data presentation. As a consequence,new laboratories were forced to adapt commercial soft-ware to their own tastes. The Helen Hayes Softwarehelped fill this temporary void. Happily, the commonneeds of most of us are now met with commerciallyavailable software.

The original Helen Hayes software could now benamed the Helen Hayes marker set as software isavailable that can handle both the Helen Hayes markerset and the Cleveland Clinic marker set. The competingmarker set is the Cleveland Clinic, credited to KevinCampbell of the Cleveland Clinic Foundation. Thedifferences in the two marker sets are briefly outlined asfollows:

Both marker sets are used to define joint centers andsegmental coordinate systems (SCS) needed to calculateangular kinematics. The main difference between bothsets is in the way the joint centers and coordinatesystems are defined. Helen Hayes (HH) is a ‘wand-based’ marker set, in which joint centers and segmentalcoordinate systems are defined using a wand marker oneach segment (i.e. thigh and shank). As joint markersare shared between segments and, therefore, each seg-

ment has at least three markers for its definition, theHelen Hayes is considered a simplified marker set,which along with a static trial of markers on the medialand lateral sides of each joint (ankle and knee), giveseverything that is needed to calculate joint angles (i.e.location and orientation of each joint axis) [73].

The Cleveland Clinic Foundation (CCF) marker setis a ‘cluster-based’ marker set, in which clusters orarrays of three (or sometimes four) markers are used todefine joint centers and segmental coordinate systems.With this marker set, the clusters are placed on eachsegment along with the medial and lateral markers,which define the flexion-extension axis of each joint,during the static trial. As the clusters define a coordi-nate system to reference the positions of the medial andlateral markers, all medial and lateral joint markers canbe removed after the static trial, and a dynamic trialcan be collected, while still maintaining the locationand orientation of each joint axis [74].

The advantage of the Helen Hayes marker set is thatit is relatively simple to use and more applicable to gaitanalysis of children. The arrays used in the ClevelandClinic marker set have been known to hit each other insmaller children. In a recent comparison study, con-ducted in our Motion Analysis Laboratory by ArnelAguinaldo, MA, ATC, and the laboratory team, weobserved less variability in the transverse plane kine-matics with the Cleveland Clinic marker set. This wasprobably because there was less marker movement, dueto the fact that there are at least three markers fixed toa rigid frame, although skin motion over the segmentdefined by the array is still a factor.

2.6. Methods of joint angle calculation

Although, there are two methods of joint angle calcu-lation most frequently used: Euler/Cardan and Helicalscrew axis, there are at least three additional methods,with definitions and pros and cons noted in Table 1.

Further use and description of joint movement androtational three-dimensional motion by Kenton Kauf-man, PhD, includes the transformation matrix methodfor complete representation [78]. The Euler/Cardanmethod lends itself well to clinical interpretation.Therefore, we utilize it for most of our clinical studiesthat do not involve translational movement, such asthat found in patients with anterior cruciate insuffi-ciency. For persons with knee instability, which mayadd more than normal translation to angular rotation,the use of the helical screw axis method of joint anglecalculation is more appropriate.

John Greaves, PhD, a graduate in electrical engineer-ing from the University of California, Santa Barbara,while he was a student with Glen Culler, worked on aproject they dubbed ‘the bug watcher’. The researchersin the biological sciences department were interested in


monitoring the motion of the microscopic marineplankton known as dynoflagellates. Greaves tackled theproblem for them by designing and building a systemusing a video camera to look into a microscope coupledwith a computer to process the video signals andprovide position and velocity information about thedynoflagellates. The system worked. After some inter-vening post-doctoral activities, John Greaves foundedMotion Analysis Corporation; a company that provides3-D motion capture systems for many clinical gaitlaboratories, sports analysis research facilities, andequipment for use in the movie industry. This companyutilizes ORTHOTRAK software for processing of move-ment, developed jointly by the Cleveland Clinic Foun-dation and Motion Analysis Corporation, inconjunction with Chet Tylkowski, MD, then at theHuman Motion Laboratory, Department of Ortho-paedics, University of Florida, Gainesville. The currentversion is a joint development effort with JamesRichards, PhD, Freeman Miller, MD, and Patrick Cas-tagno, MS, from the University of Delaware and theAlfred I. duPont Hospital for Children.

Chester Tylkowski, MD, became interested in gaitanalysis while working as a fellow with Dr SheldonSimon at Boston Children’s Hospital, 1978–79. He setup a clinical gait laboratory at the University of Flor-ida, Gainesville, in 1980. Subsequently, in 1989, DrTylkowski and the laboratory moved to the MiamiChildren’s Hospital. His third move was to the ShrinersHospital, Lexington, Kentucky, in 1995, where he holdsthe position of Chief of Staff. He holds the unofficialrecord for the physician responsible for forming thelargest number of clinical gait laboratories. This is anindication of his commitment to the importance ofclinical gait analysis throughout his orthopaedic career[79–81]. A kinesiologist, Susan Sienko Thomas, MA, aformer student of David Winter, holds a similar recordfor involvement in three gait laboratories. In 1985,Susan was hired to oversee the operation of the South-ern Illinois University Motion Analysis Laboratory inSpringfield, Illinois, after the tragic death of MaxineCovert. The laboratory had been established with aVICON system in 1983. In 1989, she assisted in the setup and operation of the Motion Analysis Laboratory atChildren’s Memorial Hospital, Chicago, where the He-

len Hayes clinical software package was used. Shecurrently holds the position of Clinical Research Coor-dinator at the Shriners Hospital in Portland, Oregon[82,83].

Steven J. Stanhope, PhD, from the BiomechanicsLaboratory, National Institutes of Health, has stronglyadvocated inter-laboratory reliability and the need fordemonstrating proven means of testing kinematicmovement measurements. His efforts have lifted thestandards for all gait laboratories in the U.S. Similarefforts are occurring in Europe under initial funding bythe European Union. The testing of laboratories shouldbe a requirement for certification. Although many ef-forts have been exerted to move this important taskforward, it still remains in the discussion stage. Thereshould be no objection to the concept that patients,referring physicians, and payers have a right to knowthat a laboratory is capable of providing accurate andreliable data. There is little doubt that testing andcertification will be instituted, hopefully in the nearfuture.

The majority of clinical patients seen at the Biome-chanics Laboratory, National Institutes of Health, areadults referred by physiatrists or internists. Thus, theclinical applications have been primarily for subjectswith rheumatoid arthritis, osteoarthritis, limb defi-ciency, diabetes, stroke, and neuromuscular disorders[84–86]. The establishment of a three-dimensional mus-culoskeletal database is a research effort that will con-tribute greatly to clinical applications [87]. Software forclinical applications, as well as for research, were devel-oped in this laboratory, and provided to laboratoriesthroughout the world. The program is called Move 3D.This software is very robust and, according to DrStanhope, is the first software to provide six degrees offreedom gait measurements.

3. Shriners network of gait laboratories

There are 19 orthopedic Shriners Hospitals in theU.S., Canada and Mexico. Of these, 12 currently havegait laboratories, with a 13th laboratory being estab-lished in Tampa, all of them carrying out clinicalanalysis before and after treatment. In addition, they

Table 1Methods of joint angle calculation (A. Aguinaldo, MA, ATC, Motion Analysis Laboratory, Children’s Hospital, San Diego)

Method Pros ConsDefinition

Angles projected onto plane Simple ParallaxPlane projectionsNot clinically relevantComplete representationDirection cosines Transformation matrixGimbal locksEuler/Cardan Sequence of rotations Clinical interpretation

Euler with floating axis Not suitable for kineticsSequence independentGrood and SuntayScrew axisHelical Rotation and translation Sensitive to noise

References: [75,76,73,77].


are involved in collaborative research. This confirma-tion of the usefulness of clinical gait laboratories hasnot gone unnoticed by physicians in the children’shospitals of North America. This powerful statementby a major block of children’s hospitals may wellhave helped spur the establishment of centers foranalysis in the other Children’s Hospitals, both pri-vate and University affiliated.

3.1. Acti�e marker systems

The Selcom Company of Sweden developed theSelspot System, which used active markers taped tothe limb segments, prior to the advent of VICON andMotion Analysis Corporation. The first Selspot move-ment capture systems in the U.S. were installed in thelaboratory of Thomas Andriacchi, PhD, at RushPresbyterian Medical Center in Chicago, and in thelaboratory of Robert Mann, PhD, at MassachusettsInstitute of Technology, Cambridge, MA. HermanWoltring, PhD, did some of the early work with theSelspot System as did Andrew Dainas. Robert Mann,PhD, and Eric Antonsson, PhD, established a clinicallaboratory at Massachusetts General Hospital in 1984,along with several orthopedists, William Harris, MD,Henry Mankin, MD, Donald Madeiras, MD, andMichael Erlich, MD, PhD. Their system was designedfor cerebral palsy gait analysis. David Krebs, PhD,PT, continues this work, collecting data on humansubjects with a variety of problems, including cerebralpalsy and disorders of posture and balance. This lab-oratory utilizes Selspot active markers, arranged inclusters on a fixed base, applied to each of the bodysegments being studied. A large database of childrenand adults with neuromuscular disorders, includingmany with problems of balance, has been established.One area of investigation is sitting-to-stand move-ments.

Although the Selspot system eliminates the need formarker identification and tracking, it contains otherless positive features including its propensity to pickup reflections, the necessity for the subjects to carrycumbersome apparatus, and the trade-off betweensampling rate and the number of markers. These in-herent drawbacks in this active marker system servedto energize the proponents of passive marker systemsand to keep them working to solve the difficultieswith marker identification and tracking. In spite ofthe current preponderance of passive marker systems,some active marker systems are emerging to competewith the passive models. Examples of active markersystems are those by CODA and Skill Technologies,(see Appendix A). The brief discussion contained inthis manuscript will be useful to clinicians, but engi-

neers and physicists will be well served by readingFurnee’s very complete, and highly technical, descrip-tions of passive and active motion capture systems[35,36].

3.2. Three-dimensional marker identification andtracking

Andrew Dainas deserves much credit, along withDoug McGuire, for more efficient processing of rawTV data to 3-D coordinates. In a personal communi-cation Dainas writes,

‘‘By 1988 we had finished the first version of theAMASS software. The software development wascarried independently of NIH, and was not sup-ported by NIH. At that time, we installed it on theNIH VICON system and offered it as a replace-ment to Oxford Metrics for their aged system. As itturned out, in 1988, Oxford Metrics had completedtheir new VAX–VX hardware system but lackedappropriate software, and they agreed to marketAMASS with their systems. Between 1988 and 1993,Oxford Metrics sold some 70 VICON systems bun-dled with the AMASS hardware. In 1993, OxfordMetrics announced the VICON 370 system, withtheir own software (which replicates many functionsof AMASS). We (at ADTECH) in turn portedAMASS to the PC computer, and adapted it to workwith raw data from both Motion Analysis Corpora-tion systems and ELITE systems. Currently, we of-fer AMASS as alternative software for these systems.AMASS can claim to be the first software used inclinical and gait applications to provide:1. Intelligent marker reduction to 2-D centers in the

camera image data by fitting circles to the pixelsoutlining the markers.

2. Provide the user with the ability to linearize eachcamera for distortions, etc.

3. Do automatic identification of reference markersin the 3-D camera calibration process.

4. Do completely hands-off 3-D reconstruction andtracking of unidentified camera image data.

In 1994, I wrote and incorporated into AMASS thefirst program to carry out the 3-D camera systemcalibrations using a large number of markers whoselocations in 3-D space need not be measured before-hand. This technique does away with the need forrigid calibration objects, or hanging strings or rods,and is capable of eliminating a chief source of inaccu-racies in most currently used 3-D measurement sys-tems. The method has since been also implementedby Oxford Metrics.’’


3.3. Present reality

Commercial hardware and software now availablehave nearly eliminated problems with marker identifica-tion and tracking, thus removing the chief objection topassive marker systems. As a consequence, the develop-ment and utilization of active marker systems was onhold for a time. In today’s state-of-the-art laboratory, asubject can be fitted with appropriate reflective markersand walk down a calibrated walkway, while the mark-ers are automatically tracked and thousands of compu-tations are performed by a dedicated high-end PCcomputer or a computer work station. The resultantjoint angles can be viewed within minutes from the endof collection of the data. An increase in the number ofcameras, plus 3-D identification and tracking of mark-ers, now enable laboratory personnel to examine thedata for reliability and potential errors while the subjectis still present in the laboratory. This represents anenormous evolution in automated movement measure-ments in the 34 plus years since the technology was firstdeveloped. There are still some problems to be workedout. For example, accurate timing of toe-off is prob-lematic with kinematic methods. The incorporation offorce platform input establishes the events of foot-strikeand toe-off accurately for those patients able to contacttwo or more force platforms. However, it is the slowwalkers, using crutches or a walker, who often exhibitvariable or even inaccurate foot-contact times, as calcu-lated from the trajectory velocities of markers on thefoot. Yet another problem is that of marker movementdue to skin movements over the underlying skeleton. Anumber of research studies address this problem, butnone have discovered a way to totally eliminate inaccu-racies due to skin movement [88–95].

If these reasons are not enough to convince thereader of the need for additional research, or eveninvestigation of other methods of measuring movement,there is yet another problem, that of placing markersaccurately and reliably. Mistakes can alter the calcula-tions of joint centers. The models for establishing hipcenter, used in all of the commercial software systems,have come from cadaver studies and are not patientspecific. This inherent flaw in patients with pathologyof the hip makes moment and power calculation of thehip suspect. Discussion of this topic will be included inThe Evolution of Clinical Gait Analysis Part III, Kinet-ics and Energy.

3.4. Future

It would be a mistake to assume from the rapiddevelopment of 3-D passive marker systems that tech-nological advances in active marker systems are notoccurring. There are currently several companies em-ploying active marker systems. Why, with good passive

systems dominating the field, is this occurring? Thepassive marker systems are expensive, considerabletraining is still required for optimal use of the hardwareand software, and flexibility in programming for specialstudies requires the talents of engineers. Our laboratorycurrently uses an 8-camera, passive marker system, andeven larger camera arrays are in use in somelaboratories.

Possible developments in the next decade are:

(1) The elimination of the need for either passive oractive targets and a reduction in the number of camerasnow in use. With increasing computer memory and disccapacity, markerless measurements of 3-D motion loomon the horizon as a possibility.

(2) The development of an active marker system withradio-frequency active emitters is a promising approachfor the economics of gait analysis hardware. This wouldbring about economies in the number of cameras re-quired. The technology for 3-D identification of radio-frequency signals is already well established in militaryapplications. The active emitters are lightweight andrelatively inexpensive and there is little to prevent theuse of a large number of markers. If such a system is tobe implemented, there must be initial research anddevelopment investment, following which the costs forpurchase of software and hardware would be wellbelow the present costs for passive marker systems.(Tera Research has a patent pending for this technol-ogy. For further information, contact Dr Walter Heineat [email protected].)

(3) Better methods of defining joint centers, especiallywith regard to hip joint center, will be required, possi-bly with the aid of CT and MRI scanning. Momentmeasurements are sensitive to the accuracy of jointcenter calculations; consequently, errors in joint centersdegrade the accuracy of moment calculations.

(4) The use of neural network statistical analysis isstill in its infancy in clinical gait analysis [96] andcomputer assisted diagnostic and problem identification[97] will surely expand in the next decade.

(5) In addition to whole body gait analysis, footmodels are emerging. Analyzing joints distal to theankle remains a major challenge for the future [98,99].

The treatment of individuals with pathological gaitwill steadily change as data are gathered and publishedfrom multiple sources. A quantum change has alreadyoccurred in the treatment of cerebral palsy. A newgeneration of multidisciplinary motion analysis teams isforging new standards of quality and pushing the limitsof application to a wide variety of disabilities. Rapidchanges are occurring in the treatment of myelodys-plasia, and improvements in the recognition and treat-ment of a large variety of neurological disorders are on


the horizon. We now have the tools to perform func-tional analysis and to replace guesswork with a scien-tific framework for evaluation and treatment. Oral drugtreatment, injections of Botulinum Toxin Type A, in-trathecal baclofen pump, physical therapy, orthoticmanagement, orthopedic and neurosurgery must all beevaluated on both a short and long term basis, and gaitanalysis must play a pivotal role. The changes in treat-ment will be incremental, wide-ranging, and will comefrom all parts of the globe. There is much work to doand the beneficiaries will be patients with disorders ofmovement. It is a source of great satisfaction to pa-tients, their parents, and their physicians to know thatlocomotion and movement disorders are at last receiv-ing the attention they deserve.

Acknowledgements

My special thanks to all of the individuals whoresponded to my letters and phone calls, supplyingdetails that give life to this account. For the administra-tive assistance provided by Sherill Marciano, Jill Jor-dano, and Kit Holm, who put up with my manychanges to the manuscript, and for Kit’s tenacity withresearch, which helped immensely with reviewer re-sponse and final publication requirements. To bioengi-neer, Arnel Aguinaldo, and physical therapists,Marilynn Wyatt and Janet Buttermore, for their assis-tance with the search for details and review of themanuscript. To John Hagy who filed and saved corre-spondence and other documents from the early days ofthe San Francisco Gait Lab. Finally, my thanks to DrHank Chambers, Medical Director of the Motion Anal-ysis Laboratory at Children’s Hospital, San Diego, forhis helpful comments.

Appendix A. A partial list of commercial kinematicsystems

Ariel dynamics6 Alicante StreetTrabuco Canyon, CA 92679USATel.: (949) 858 4216Fax: (949) 858 5022

BTSVia C. Columbo, 1A 20094 CorsicoMilano, ItalyTel.: +39 02458751Fax: +39 0245867074

CODACharnwood Dynamics Ltd.

17 South Street, Barrow on SoarLeicestershire, LE12 8LYTel.: +44 (0) 116 230 1060Fax: +44 (0) 116 230 1857

Motion analysis corporation3617 Westwind BlvdSanta Rosa, CA 95403USATel.: (707) 579-6500Fax: (707) 526-0629

Peak performance7388 S. Revere ParkwaySuite 603Englewood, CO 80112USATel.: (303) 799 8686Fax: (303) 799 8690

PrimasMotion Studies LaboratoryDelft University of TechnologyP.O. Box 52600 AA DelftThe NetherlandsTel.: +31 (0) 15 278 9111Fax: +31 (0) 15 278 6522

Qualisys Inc.148 Eastern Blvd, Suite 110Glastonbury, CT 06033USATel.: +1 860 627 5060Fax: +1 860 627 4041

Qualisys ABDrottninggatan 31Goteborg 41114SwedenTel.: +46 (u) 317743830Fax: +46 (u) 317014145

Selspot, ABSallarangsgatan 3S-431 37 Molndal, Sweden

Skill Technologies, Inc.1202 E. Maryland Ave., Suite IGPhoenix, AZ 85014USATel.: 602-277-7678Fax: 602-277-2326

VICON motion systemsOxford Metrics LimitedUnit 14, MINNS ESTATE7 West WayOxford OX20JB


UKTel.: +44 (1865) 26 1800Fax: +44 (1865) 24 05 27

Appendix B

B.1. Copy of letter to Raymon Linder, from DrDavid Sutherland, dated August 23, 1965.


B.2. Copy of letter to John Hagy, from Dr E. R.Schottstaedt, dated November 29, 1965.

D.H. Sutherland / Gait and Posture 16 (2002) 159–179 175A

ppen

dix

C

D.H. Sutherland / Gait and Posture 16 (2002) 159–179176D

escr

ipti

onof

our

first

use

ofa

dedi

cate

dco

mpu

ter

for

‘Dia

gnos

tic

Gai

tA

naly

sis’

,by

Ele

ctro

nic

Pro

cess

ors,

Inc.

(tw

opa

ges)

.


References

[1] Braune W, Fisher O. Ueber den schwerpunkt des menschlichenkorpers, mit rucksicht auf die ausrustung des deutschen infanter-isten. ABH Math Phys Cl K Sachs Ges Wissensch 1889;15:559.

[2] Braune W, Fischer O. Die rotationsmomente der Beugemuskelnam Ellbogengelenk des Menschen, Abh d Konigl Sachs Ges dWissensch Math Phys Klasse 1890;15.

[3] Braune W, Fischer O. Der gang des menschen. Berlin: B.G.Teubner, 1895.

[4] Braune W, Fischer O. Untersuchungen uber die elenke desmenschlichen Armes, in: Abh Sachs Akad Wiss, editor. Praktis-che Anatomie Bd. I, Teil III. Berlin: Julius Springer, 1935. p. 94.

[5] Braune W, Fischer O. The human gait. Berlin: Springer-Verlag,1987.

[6] Eberhart HDaA, Fundamental studies of human locomotionand other information relating to design of artificial limbs.Berkeley: University of California, 1947.

[7] Levens AS, Inman VT, Blosser JA. Transverse rotation of thesegments of the lower extremity in locomotion. J Bone JointSurg 1948;30-A:859.

[8] Murray MP, Drought AB, Kory RC. Walking patterns ofnormal men. J Bone Joint Surg 1964;46A(2):335–60.

[9] Murray MP, Kory RC, Sepic SB. Walking patterns of normalwomen. Arch Phys Med Rehabil 1970;51:637–50.

[10] Murray MP, Mollinger LA, Sepic SB, Gardner GM, Linder MT.Gait patterns in above-knee amputee patients: hydraulic swingconstant-friction knee components. Arch Phys Med Rehabil1983;64(8):339–45.

[11] Murray MP, Gore DR, Sepic SB, Mollinger LA. Antalgic ma-neuvers during walking in men with unilateral knees. Clin Or-thop 1985;1:192–200.

[12] Murray MP, Jacobs PA, Gore DR, Gardner GM, Mollinger LA.Functional performance after tibial rotationplasty. J Bone JointSurg 1985;67(3):392–9.

[13] Karpovich PV, Karpovich GP. Electrogoniometer: a new devicefor study of joints in action. Fed Proc 1959;18:79.

[14] Bajd T, Stanic U, Kljajic M, Trnkoczy A. On-line electrogonio-metric gait analysis. Comp Biomed Res 1976;9(5):439–44October.

[15] McLeod PC, Kettelkamp DB, Srinivasan V, Henderson OL.Measurement of repetitive activities of the knee. J Biomech 1975;8(6):369–73.

[16] Tata JA, Quanbury AO, Steinke TG, Grahame RE. A variableaxis electrogoniometer for the measurement of single planemovement. J Biomech 1978;11:421–5.

[17] Johnston RC, Smidt GL. Measurement of hip joint motionduring walking. J Bone Joint Surg 1969;51A:1083–94.

[18] Lamoreux LA. Kinematic measurements in the study of humanwalking. Bull Prosthet Res 1971;3:10–5.

[19] Kinzel GL, Hillberry BM, Hall ASJ, Sickle V, Harvey WM.Measurement of the total motion between two body segments,II. Description of application. J Biomech 1972;5:283–93.

[20] Townsend MA, Izak M, Jackson RN. Total motion knee go-niometry. J Biomech 1977;10:183–93.

[21] Foort J, Cousins S, Hannah R. Electronic recording of jointfunction, in: Workshop on human locomotion and clinical anal-ysis of gait. University of British Columbia, Vancouver, B.C.1976.

[22] Chao YS. Justification of triaxial goniometer for the measure-ment of joint rotation. J Biomech 1980;13:989–1006.

[23] Linder R. Data aid, San Jose, girl friday letter shop, 1965.[24] Sutherland DH, Schottstaedt ER, Vreeland RW, Oyama FR,

Hagy JL. Measurement of movements and timing of musclecontraction from movie film. J Bone Joint Surg 1967;49A:1248–9.

[25] Sutherland DH, Olshen R, Cooper BA, Woo SLY. The develop-ment of mature gait. J Joint Bone Surg 1980;62-A(3):336–53.

[26] Sutherland DH, Cooper L. The pathomechanics of progressivecrouch gait in spastic diplegia. Orthop Clin North Am1978;9:143–54.

[27] Sutherland DH. Gait analysis in cerebral palsy. Dev Med ChildNeurol 1978;20(6):807–13.

[28] Sutherland DH, Cooper L, Daniel D. The role of the ankleplantar flexors in normal walking. J Bone Joint Surg1980;62(3):354–63.

[29] Sutherland DH, Olshen RA, Cooper L, Wyatt M, Leech J,Mubarak S, Schultz P. The pathomechanics of gait in Duchennemuscular dystrophy. Dev Med Child Neurol 1981;23:3–22.

[30] Furnee E. TV/computer motion analysis systems the first twodecades, in: Signal processing department. Delft, The Nether-lands: Delft University of Technology, 1989. p. 247.

[31] Furnee EH. Hybrid instrumentation in prosthetics research. in:Proc. 7th. Intl. Conf. on Med. and Biol. Eng., Stockholm, 1967.

[32] Furnee E, Halbertsma J, Klunder G, Miller S, Nieukerke K, vander Burg J, van der Meche F. Proceedings: automatic analysis ofstepping movements in cats by means of television system and adigital computer. J Physiol 1974;240(2):3P–4P July.

[33] Steilberg PC. Het ontwerp van een video-naar-digitaal omzettervoor de analyse van armbewegingen, in: Faculty of AppliedPhysics. Delft: Delft University of Technology, 1967.

[34] Steilberg PC. Televisiereregistratiesysteem voor de analyse vanarmbewegingen, in: Faculty of Applied Physics. Delft: DelftUniversity of Technology, 1968.

[35] Furnee EH. TV/computer motion analysis systems: the first twodecades. Delft University of Technology, 1989.

[36] Furnee E. Opto-electronic movement measurement systems: as-pects of data acquisition, signal processing and performance, in:Otto Bock Foundation International Symposium. Berlin, Ger-many: Mecke Druck and Verlag, Duderstadt, 1990.

[37] Andrews BJ, Jarrett MO, Paul JP. On-line kinematic data acqui-sition, in: B.E.S. Conference. Edinburgh, Scotland, 1975.

[38] Jarrett MO, Andrews BJ, Paul JP. A television–computer systemfor the analysis of human locomotion, in: I.E.R.E. GoldenJubilee Conference. An exhibition on Application of Electronicsin Medicine. Southampton University, Southampton, England,1976.

[39] Paul JP, Jarrett MO, Andrews BJ. Quantitative analysis oflocomotion using television, in: World Congress I.S.O. Mon-treux, 1974.

[40] Whittle MW, Herron RE, Cuzzi JR, Hugg JE. Effects of ex-tended space flight on body form of Skylab astronauts usingbiostereometric methods. pp. 588–589, in: Herron RE, editor.Biostereometrics 1974;74(708):30.

[41] Whittle MW. Gait analysis: an introduction, 2nd ed. Oxford,Boston: Butterworth-Heinemann, 1996.

[42] Ferrigno G, Pedotti A. ELITE: a digital dedicated hardwaresystem for movement analysis via real-time TV signal processing.IEEE Trans Biomed Eng 1985;32(11):943–50.

[43] Cappozzo A, Leo T, Pedotti A. A general computing method forthe analysis of human locomotion. J Biomech 1975;8(5):307–20.

[44] Cappozzo A. Gait analysis methodology. Hum Mov Sci1984;3:27–50.

[45] Dinn DF, Winter DA, Trenholm BG. CINTEL—computerinterface for television. IEEE Trans Comput 1970;C-19:1091–5.

[46] Letts RM, Winter DA, Quanbury M. Locomotion studies as anaid in clinical assessment of childhood gait. CMA J1975;112(May 3):1091–4.

[47] Winter DA, Greenlaw RK, Hobson DA. Television–computeranalysis of kinematics of human gait. Comp Biomed Res1972;5:498–504.

[48] Winter DA, Quanbury AO, Hobson DA, Sidwall HG, ReimerG, Trenholm BG, Steinke T. Kinematics of normal locomo-


tion—a statistical study based on TV data. J Biomech1974;7(6):479–86 November.

[49] Winter DA, Yack HJ. EMG profiles during normal humanwalking: stride-to-stride and inter-subject variability. Electroen-cephalogr Clin Neurophysiol 1987;67(5):402–11.

[50] Winter DA. Biomechanics and motor control of human move-ment, 2nd ed. New York: Wiley, 1990.

[51] Simon SR, Knirk J, Mansour JM, Koskinen MF. A comprehen-sive clinical system for four-dimensional motion analysis. BullHosp Joint Dis 1977;38(1 April):41–4.

[52] Simon SR, Deutsch SD, Nuzzo RM, Mansur MJ, Jackson JL,Koskinen M, Rosenthal RK. Genu-recurvatum in spastic cere-bral palsy. J Bone Joint Surg 1978;60A:882–94.

[53] Simon SR, editor. Orthopaedic basic science. AmericanAcademy of Orthopaedic Surgeons, 1994.

[54] Gage JR, Fabian D, Hicks R, Tashman S. Pre- and postopera-tive gait analysis in patients with spastic diplegia: a preliminaryreport. J Pediatr Orthop 1984;4:715–25.

[55] Gage JR, Perry J, Hicks RR, Koop S, Werntz JR. Rectusfemoris transfer to improve knee function in children withcerebral palsy. Dev Med Child Neurol 1987;29:159–66.

[56] Gage JR. Surgical treatment of knee dysfunction in cerebralpalsy. Clin Orthop 1990;253:45–54.

[57] Gage JR. An overview of normal walking. Instr Course Lect1990;39:291–303.

[58] Gage JR. Gait analysis. An essential tool in the treatment ofcerebral palsy. Clin Orthop 1993;288:126–34.

[59] Gage JR. The role of gait analysis in the treatment of cerebralpalsy. J Pediatr Orthop 1994;14(6):701–2.

[60] Gage JR. Gait analysis in cerebral palsy. New York: MacKeithPress, 1991.

[61] Ounpuu S. The biomechanics of running: a kinematic andkinetic analysis, Inst. Course Lect. 1990:305–318.

[62] Ounpuu S, Gage JR, Davis RB. Three-dimensional lower ex-tremity joint kinetics in normal pediatric gait. J Pediatr Orthop1991;11:341–9.

[63] DeLuca PA, Ounpuu S, Davis RB, Walsh JH. Effect of ham-string and psoas lengthening on pelvic tilt in patients with spasticdiplegic cerebral palsy. J Pediatr Orthop 1998;18(6):712–8.

[64] Ounpuu S, Muik E, Davis RB, Gage JR, DeLuca PA. Rectusfemoris surgery in children with cerebral palsy. Part II: a com-parison between the effect of transfer and release of the distalrectus femoris on knee motion. J Pediatr Orthop1993;13(3):331–5.

[65] Ounpuu S, Muik E, Davis RB, Gage JR, DeLuca PA. Rectusfemoris surgery in children with cerebral palsy. Part I: the effectof rectus femoris transfer location on knee motion. J PediatrOrthop 1993;13(3):325–30.

[66] DeLuca PA, Davis RB, Ounpuu S, Rose S, Sirkin R. Alterationsin surgical decision making in patients with cerebral palsy basedon three-dimensional gait analysis. J Pediatr Orthop1997;17(5):608–14.

[67] Davis RB, O� unpuu S, Tyburski DJ, Gage JR. A gait analysisdata collection and reduction technique. Hum Mov Sci1991;10:575–87.

[68] Watts HG. Gait laboratory analysis for preoperative decisionmaking in spastic cerebral palsy: is it all it’s cracked up to be? JPediatr Orthop 1994;14:703–4.

[69] Bartlett J. In: Kaplan J, editor. Bartlett’s familiar quotations,16th ed. Little, Brown and Co, London, UK 1992.

[70] Kadaba MP, Ramakrishnan HK, Wootten ME, Gainey J, Gor-ton G, Cochran GV. Repeatability of kinematic, kinetic, andelectromyographic data in normal adult gait. J Orthop Res1989;7(6):849–60.

[71] Kadaba MP, Ramakrishnan HK, Jacobs D, Chambers C, Scar-borogh N, Goode B. Pattern recognition in spastic diplegia.Biomecanica 1995;III(4):49–58.

[72] Kadaba MP, Wootten ME, Gainey J, Cochran GV. Repeatabil-ity of phasic muscle activity: performance of surface and intra-muscular wire electrodes in gait analysis. J Orthop Res1985;3(3):350–9.

[73] Kadaba MP, Ramakrishnan HK, Wootten ME. Measurement oflower extremity kinematics during level walking. J Orthop Res1990;8(3):383–92.

[74] Campbell KR. Expert vision, data acquisition and analysis, in:ASME International Computers in Engineering Conference andExhibition, New York, NY, 1987.

[75] Craig JJ. Introduction to robotics. Addison-Wesley, Boston, MA1989.

[76] Grood ES, Suntay WJ. A joint coordinate system for the clinicaldescription of three-dimensional motions: application to theknee. J Biomech Eng 1983;105:136–44.

[77] http://guardian.curtin.edu.au/cga/faq/angles.html O, Methodsfor representation of angular rotations.

[78] Sutherland DH, Kaufman KR, Moitoza JR. Kinematics ofnormal human walking. In: Rose J, Gamble JG, editors. Humanwalking. Baltimore: Williams and Wilkins, 1994:23–43.

[79] Oeffinger DJ, Pectol RWJ, Tylkowski CM. Foot pressure andradiographic outcome measures of lateral column lengtheningfor pes planovalgus deformity. Gait and Posture 2000;12(3):189–95.

[80] Tylkowski CM, Simon SR, Mansour JM. The Frank StinchfieldAward Paper. Internal rotation gait in spastic cerebral palsy, inHip, 1982:89–125.

[81] Tylkowski CM, Rosenthal RK, Simon SR. Proximal femoralosteotomy in cerebral palsy. Clin Orthop1980;151(September):183–92.

[82] Mazur JM, Sienko-Thomas S, Wright N, Cummings RJ. Swing-through vs. reciprocating gait patterns in patients with thoracic-level spina bifida. Z Kinderchir 1990;45(Suppl. 1):23–5.

[83] Buckon CE, Sienko-Thomas S, Aiona MD, Piatt JH. Assess-ment of upper-extremity function in children with spastic diple-gia before and after selective dorsal rhizotomy. Dev Med ChildNeurol 1996;38(11):967–75.

[84] Cooper RA, Quatrano LA, Stanhope SJ, Cavanagh PR, MillerF, Kerrigan DC, Esquenazi A, Harris GF, Winters JM. Gaitanalysis in rehabilitation medicine: a brief report. Am J PhysMed Rehab 1999;78(3):278–80 May–June.

[85] O’Connell PG, Lohmann Siegel K, Kepple TM, Stanhope SJ,Gerber LH. Forefoot deformity, pain, and mobility in rheuma-toid and nonarthritic subjects. J Rheumatol 1998;25(9):1681–6September.

[86] Siegel KL, Kepple TM, O’Connell PG, Gerber LH, Stanhope SJ.A technique to evaluate foot function during the stance phase ofgait. Foot Ankle Int 1995;16(12):764–70 December.

[87] Kepple TM, Sommer H Jr., Lohmann Siegel K, Stanhope SJ. Athree-dimensional musculoskeletal database for the lower ex-tremities. J Biomech 1998;31(1):77–80 January.

[88] Holden J, Orsini J, Siegel K, Kepple TM, Gerber L, Stanhope S.Surface movement errors in shank kinematics and knee kineticsduring gait. Gait and Posture 1997;5:217–27.

[89] Reinschmidt C, van den Bogert AJ, Nigg BM, Lundberg A,Murphy N. Effect of skin movement on the analysis of skeletalknee joint motion during running. J Biomech 1997;30(7):729–32.

[90] Cappozzo A, Catani F, Leardini A. Skin movement artifacts inhuman movement photogrammetry, in: Abstracts of the XIVthCongress of the International Society of Biomechanics, Paris,1993.

[91] Angeloni C, Cappozzo A, Catani F, Leardini A. Quantificationof relative displacement between bones and skin and platemounted markers, in: Eighth Meeting of the European Society ofBiomechanics, Rome, 1992.

[92] Perry J, Enwemeka CS, Gronley JK. The stability of surfacemarkers during knee flexion, in: 34th Annual Meeting of theOrthopaedic Research Society, 1988.

http://guardian.curtin.edu.au/cga/faq/angles.html


[93] Lafortune MA, Cavanagh PR, Kalenak A, Skinner SM, SommerHJ. The use of intra-cortical pins to measure the kinematics ofthe knee joint, in: Proceedings of the Second Biennial Conferenceof the Canadian Society for Biomechanics, Kingston, 1982.

[94] Lucchetti L, Cappozzo A, Cappello A, Croce UD. Skin move-ment artefact assessment adn compensation in the estimation ofknee-joint kinematics. J Biomech 1998;31:977–84.

[95] Alexander EJ, Andriacchi TP. Correcting for deformation inskin-based marker systems. J Biomech 2001;34:355–61.

[96] Biafore S, Cottrell G, Focht L, Kaufman K, Wyatt M, Suther-land D. Neural network analysis of gait dynamics. Trans Orthop

Res Soc 1991;16(1):225.[97] Weintraub MA, Bylander T, Simon SR. QUAWDS: a composite

diagnostic system for gait analysis. Comput Methods ProgramsBiomed 1990;32(1):91–106.

[98] Kidder SM, Abuzzahab FS, Harris GF, Johnson JE. A systemfor the analysis of foot and ankle kinematics during gait. IEEETrans Rehab Eng 1996;4(1):25–32.

[99] Kaufman KR, Kitaoka HP, Hansen D, Shaughnessy WJ. Tech-nique for measurement of foot and ankle kinematics in children,in: IEEE/EMBS 19th International Conference, Chicago, IL,1997.

the evolution of clinical gait analysis part ii kinematics

Documents