PattonEtAl(2008)TopicsInStrokeRehabilitation15(2)131-139

8/13/2019 PattonEtAl(2008)TopicsInStrokeRehabilitation15(2)131-139

1/9

131

KineAssist: Design and Development

of a Robotic Overground Gaitand Balance Therapy DeviceJames Patton, David A. Brown, Michael Peshkin, Julio J. Santos-Munn,

Alex Makhlin, Ela Lewis, J. Edward Colgate, and Doug Schwandt

Background and Purpose:Balance and mobility training consists of activities that carry a high risk for falling. The purpose

of this article is to describe a novel robotic system for allowing challenging, yet safe, balance and mobility training inpersons at high risk for falls. Method:With no initial preconceptions of what device we would build, a user-needs analysisled us to focus on increasing the level of challenge to a patients ability to maintain balance during gait training and alsoon maintaining direct involvement of a physical therapist (rather than attempting robotic replacement). The KineAssist

is a robotic device for gait and balance training that has emerged from a unique design process of a start-up product of asmall company and a team of therapists, engineers, mechanical design experts, and rehabilitation scientists. Results:TheKineAssistprovides partial body weight support and postural control on the torso; allows many axes of motion of thetrunk and pelvis; leaves the patients legs accessible to a physical therapists manipulation during walking; follows a patientswalking motions overground in forward, rotation, and sidestepping directions; and catches an individual who loses balanceand begins to fall. Discussion andConclusion:Design and development of the KineAssistproceeded more rapidly in the

context of a small company than would have been possible in most institutional research contexts. A prototype KineAssisthas been constructed and has received US Food and Drug Administration (FDA) classification and institutional review boardclearance for initial human studies. The acceptance of KineAssistwill ultimately depend on improved patient outcomes,the use of this new tool by therapists, the ease of use of the system, and the recognition of the unique value it brings totherapeutic recovery. Key words:device development, gait and balance training, robotics

James Patton,is Associate Professor, Bioengineering, University

of Illinois at Chicago, and Associate Director, Center for RehabRobotics, Rehabilitation Institute of Chicago (RIC), Chicago,

Illinois.

David A. Brown, PT, PhD,Feinberg School of Medicine,

Northwestern University, and Kinea Design LLC, Evanston,

Illinois.

Michael Peshkin,Northwestern University Mechanical

Engineering and Kinea Design LLC, Evanston, Illinois.

Julio J. Santos-Munn,is Director of Operations, Kinea Design

LLC, Evanston, Illinois.

Alex Makhlin,is Senior Controls Engineer, Kinea Design LLC,Evanston, Illinois.

Ela Lewis,is Clinical Project Manager, Kinea Design LLC,

Evanston, Illinois.J. Edward Colgate,Northwestern University MechanicalEngineering and Kinea Design LLC, Evanston, Illinois.

Doug Schwandt,is Mechanical Engineer Consultant.

The job of a physical therapist is physicallydemanding and challenging. Working withpatients who have balance and mobility im-

pairments puts therapists at risk for falling duringtraining sessions. Clinicians use their bodies to lift,move, and provide safety nets for patients whomay be up to three times larger than they are. Theintensity and duration of physical therapy sessionsare often limited due to exhaustion of the clini-

cian. Safety concerns sometimes limit the extent towhich the clinician is able to challenge the patientas much as possible to enhance learning, becausefalls and other injuries are not desirable.

Robots are tireless, precise devices that can dorepetitive motion. In these rather early days inthe development of humanmachine interactions,there are many unrealized functions that robotictechnology can do for rehabilitation. The device

Top Stroke Rehabil2008;15(2):131139 2008 Thomas Land Publishers, Inc.www.thomasland.com

doi: 10.1310/tsr1502-131


2/9

132 TOPICSINSTROKEREHABILITATION/MAR-APR2008

can facilitate, rather than replace, the efforts of atherapist. This collaborative approach in rehabili-tation robotic design was utilized by starting withthe end user (clinician) and implementing thefeedback received to create a device that assistswith functional mobility in stroke rehabilitation.

The resulting KineAssistdevice works with atherapist during a multitude of functional tasksthat typically take place during a physical therapysession. The therapist is the person guiding thetraining session, based upon their wide reper-tory and knowledge of appropriate exercises. Therobot remains dormant until either the patientor the therapist decides to act. The robot canalso provide assistance using several program-

ming modes, depending on need. Meanwhile,the therapist can administer training challengesand guidance rather than worrying about fallsand preventing a person from losing his or herbalance. In fact, a therapist may choose to chal-lenge a patient so that loss of balance occurs, iftherapeutically appropriate.

The purpose of this article is to describe the de-sign process that led to this exciting prototype sys-tem and to discuss some of the problems, issues,and successes of this interdisciplinary project.

Assessment of Needs

To identify opportunities for robotics, our ap-proach was first to assemble an inclusive team.Several engineers, therapists, and designers metand planned for an immersive understandingof therapists and patients needs and what newtechnology might offer. We partnered with severalindustrial design specialists at IDEO, a Palo Alto,California, consulting company with a branch

office in Evanston, Illinois. An interesting philoso-phy emphasized by IDEO was that often the mar-ket does not know what the market really needs,and so observing, asking questions, and learningby example lead to great new insights. Sometimesa question is asked for which there really isnt asolid answer, pointing to an unmet need or a newmanner of thinking about solutions.

Based upon funding from the Advanced Tech-nology Program at the US National Institute ofStandards and Technology (NIST) of the US De-partment of Commerce, and additional financial

support from the Rehabilitation Institute of Chi-cago, the original proposal had no stipulation ofany particular device at the outset. Instead, the en-tire team, through direct observation mixed withperiods of discussion, attempted to discover a listof user needs through observations and interviewsduring physical therapy settings. The team con-ducted observations in a broad range of physicaltherapy settings in the Chicago area. These siteswere chosen to encompass the range of patients,tasks, and environments where physical therapyis delivered. We observed inpatient, outpatient,and home settings. We visited both hospital-basedand private clinics. We observed therapeutic tasksby patients with both orthopedic and neurological

diagnoses. Diagnostic categories included jointreplacement, lumbar diagnosis, traumatic braininjury, spinal cord injury, stroke, burns, pediatrics,and Parkinsons disease.

Interlaced between these sessions, the team wasinvolved in discussions to determine key tasks forwhich needs are the greatest and attempted to as-sess the kinematics required to support these taskswith a robotic device. This often involved drawingcartoons and pictures, a rather unconventionaland enjoyable approach to device design. Our

mix of backgroundsfrom physical therapy tocognitive psychology to engineering to industrialdesignwas useful in generating a rich range ofperspectives around what we saw and allowed usto have a shared understanding of the physicaltherapy world. It was well into the third monthof the project before there was any focus on anyspecific device.

We discovered that therapists tend to be verypractical. They often want to use very simple ma-chines or no machine at all, and they often prefer

to use their own hands. Devices that allow themto actually do the work that they need to do areimportant. Clinicians are skilled at guiding themotion of the legs manually, and they prefer tomanage the treatment that way.

We also learned that there was nothing out thereto help clinicians train a variety of balance activi-ties with substantial challenges that might initiate afall. As a result, we identified that the area of great-est need that we could address with techniquesfrom humaninteractive robotics was the practice

of overground walking in combination with bal-


3/9

Robotic Overground Gait & Balance Therapy Device 133

ance, while maintaining close interaction betweenpatient and clinician. Practicing gait in functionaloverground contexts, as opposed to treadmills, iswidely desirable. Perhaps most important, clini-cians wish to challenge patients at or beyond theirlevel of comfort, without risk of falls that couldinjure the clinician or patients.

Focus on Balance Activities

The design approach led us to understand therole of therapy that focuses on dynamic standingas a pathway to recovery of function. In one of themost prevalent causes of motor disability, stroke,the majority of individuals must recover from

significant gait and balance limitations. These in-clude the inability to effectively control standingand walking.1Only 37% of stroke survivors areable to walk after the first week2,3and even amongthose who achieve independent ambulation, sig-nificant residual deficits persist in balance andgait speed 6 months after stroke.4,5 Commondeficits that prevent ambulation include theinability to bear loads, to generate propulsiveforces, to move limbs swiftly through s trajectory,and to control lateral stability.69 Most patients

do not walk with a normal locomotor gait pat-tern. Rather, they use compensatory strategiesthat include the assistance of another person ordevices such as walkers or orthotics. Internallybased compensatory strategies include reducedgait velocity, increased stance and double sup-port time, knee hyperextension in stance, and hipcircumduction during swing phase. Although themajority of stroke survivors will achieve some levelof ambulation, there continues to be a strong needfor therapeutic interventions that can reduce their

need for long-term physical assistance. Such bio-mechanically inefficient and unstable locomotorgait patterns can lead to secondary damage overtime. One readily identifiable deficit that occursas a result of hemiparesis is decreased speed oflocomotion.911Walking speed is an effective in-dicator of the degree of abnormality in gait qual-ity, overall functional status, and clinical progressin people with paraparesis.12,13Furthermore, gaitspeed has been found to correlate with ability tobalance on either one or both lower extremities,degree of lower extremity force recovery, Barthel

Index score, degree of ambulatory independence,cadence of gait, and rating of overall gait appear-ance.1416

Recent findings have suggested that repetitivetraining for locomotion is indeed effective in fos-tering recovery. Spinal transected cats recoveredthe ability to step on a moving treadmill belt afterthey were trained on a treadmill with some bodyweight support and stimulation and assistance inpaw placement.1719 Spinal locomotor pools mayinclude a central pattern generator for alternatingflexor and extensor leg muscle activity. Barbeauand collegues attempted such a technique on hu-man patients, suspending them over a treadmillusing an overhead lift for body weight support

and clinician-provided assistance to the legs.20Since then, many studies have reported the ef-fectiveness of treadmill training with individualspoststroke.2127

Most important, clinicians safety concerns oftenlimit the amount of challenge they can introducein a therapeutic setting, even though the patientsmust face the challenges of these balanced activitiesin therapy if they are to recover. For example, if apatient starts to lose balance and fails to recover, heor she has learned more than if he or she never had

the chance to experience the balance loss. How-ever, safety concerns over an injury from an actualfall, and the patients distrust that can result whena therapist fails to prevent a loss of balance, tendto limit challenge opportunities like this. The clini-cian is prevented from practicing with the patientalone in functional environments. Sessions areconfined to parallel bars or unrealistic harnessingsystems that are mounted over treadmills. In cur-rent practice, functional practice is not a realisticoption due to the need for multiple people to assist

in case safety is compromised. Realistic challengehas not been well integrated into clinical practicebecause of the limitations imposed by safety. In anideal setting, clinicians would try to get patientsinto more realistic, functional environments asquickly as possible. Patients who do not move outof the parallel bars at the appropriate stage in theirrehabilitation may use the bars as a psychologicalcrutch, preventing the development of skills andconfidence for a real-world setting. It is clear thattechnology may help move the therapists closer to

their goals.


4/9


Design Development UsingState of the Art Tools

Rehabilitation technology based on treadmillsand robotics has been introduced recently. Co-lombo and colleagues have designed a roboticallydriven gait orthosis that operates by computerinterface with a standard treadmill28 and is nowcommercially available under the name Lokomat.The AutoAmbulator, funded by HealthSouth, is atreadmill-based device that features an overheadharness system to support body weight, mechani-cally powered braces to move the patients legs,and numerous sensors to track vital signs, move-ment, and contact speed. Leg exoskeletons have

been constructed to generate spring-like assistivejoint torques using Bowden cables and series elas-tic actuator technology in the LOPES system.29Schmidt and colleagues have developed a systemthat uses moving programmable footplates.30Theankles can also be moved using linear actuatorsmounted on the handrails of a treadmill.31Whilethese machines are merely a scattering of examplesand the remainder of developed or developingrobotic devices are too numerous to mention here,it is important to note that all have given minimal

attention to balance. Furthermore, most therapeu-tic devices have been designed to simulate the rolethat a clinician could play, eliminating the skills ofthe clinician. Consequently, the design focused onnovel methods of assisting the balance componentof standing activities.

In addition to many design and developmentdiscussions, one critical design philosophy we usedwas to develop experience prototypes as early aspossible with the quickest means possible. This ledto our experiencing mistakes as early as possible

and learning from them. Our first attempts includedinclined devices and treadmills, planar joints, bi-cycle pedals, and harnesses that would lower theamount of gravitational force on someone whilethey were walking and provide them with assis-tance. Many concepts did not move in the directionof our ultimate design, but spending a small amountof money on a cheap treadmill and some other ma-terials from the hardware store led to a great deal oflearning. For example, we explored the possibilityof a gait training device that would allow people towalk while leaning back on an inclined surface. Our

experience prototype revealed that the mechanics ofgait were unrealistically altered and quite difficulteven for a healthy individual. The therapists thattried it told us they would have little interest in sucha therapy device.

Design Description

Our resulting design, the KineAssist, shownin Figure 1, has two major components: a mobile

base system and a brace system. The mobile baseis omni-directional and uses Cobot technologyoriginally developed by Peshkin and Colgate atNorthwestern University for assistive devices inmaterials handling.32 Cobotics is software thatforms the basis of a new class of technology thatsenses human movement and allows devices tofollow and take direction from this movement.This adds precision and safety to lifting, guiding,and positioning. This admittance control method-ology renders a haptic display that compensatesfor the inertial effects of the robot, rendering the

Figure 1.Artists rendition of the design concep-tion of the KineAssist. A physical therapist workswith a patient on balance exercises, and value oftechnology is realized only when both individualswork together with the robot.


5/9


system virtually undetectable and allowing easyforward and turning motions while the machinemoves in response to the motion of the patient.Force sensed at the pelvic harness is used to drivethe motion of the system. The trunk and pelvismechanism allows the patients bending motionsboth left-right, forward-backward, rotations abouta persons transverse axis, and hip rotations aboutthe forward axis. A torso mechanism attaches atchest level and can prevent collapse of the trunk.

A software-driven safety zone limits the patientsupper body range of motion where the trunksupport implements an adjustable, compliantconstraint that catches the patients when they losebalance. In addition to simply acting as a fall-ar-resting device, this device can partially supportthe patients weight at the level of the pelvis, andthe system is also capable of comfortably applyingforces to the body. The therapist has the freedomto change parameters and assist or challenge thepatient to the level that is necessary to gain the bestclinical outcomes.

The legs of the mobile base are shown in Figure2 in their parallel configuration, which allowsthe KineAssist to pass through a standard dooropening. The legs may also be angled outward 30to allow a patient more room for side stepping.The motion of the mobile base is powered and ishighly responsive to the patients desires for mo-tion, so that the patient does not have to pull thebase. The patients intent for motion is detectedby a combination of passive sliders and integrated

force sensors incorporated into the pelvic part ofthe patient support structure. Control algorithmsmove the base in response to the patients forcesand motions, so that the patients walking andturning motions are unconstrained.

Pelvic and torso harnesses serve as the interfacebetween the machine and the patient and providethe means of comfortably applying desired forcesto the body as well as acting as a fall arrest device.The harnesses were custom designed; they aremodified circus harnesses. Harnesses may seem

commonplace, but they can be a key factor in

Figure 2.The initial KineAssistalpha prototype.

A

B


6/9


the success or failure of a rehabilitation device.Because body weight support protocols can applylarge vertical forces, comfort is essential and hardto achieve. Many harness systems impinge on thegroin area and cause such pain that patients cannotwithstand the duration of training intended.

Harness design also plays a major role in thetime it takes to get a patient set up for a period oftherapy. Twenty minutes is a typical set-up timefor other body weight support systems. In a 1-hour session, a 20-minute set-up time is a majordeterrent to use. By a combination of good har-ness design and a system for harnessing while thepatients are seated (rather than reclined), patientsmay be set up in the KineAssist typically in 5

minutes. Our harnessing system is particularlycritical because the intent of the KineAssistis toallow the clinician to challenge the patient up to orbeyond the patients comfort zone. Falls are likely;the KineAssistmust be able to slow and stop thepatient without pain.

The KineAssistis able to produce unweightingof the patient (partial body weight support train-ing) up to 150 lbs of vertical force. The vertical col-umn is powered to provide this force continuouslyand at the same time to easily allow the vertical

motions of the pelvis and torso, which are a partof normal gait. The unweighting feature is ratedto 150 lbs, but the KineAssistis designed for pa-tients up to 350 lbs, and it can safely bring such apatient to a stop after only a few inches of fall. (Thethreshold distance for identifying and stopping afall is selected by the clinician.)

The trunk and pelvic mechanism allows bend-ing motions both left-right, forward-backward,rotations about a persons transverse axis, and hiprotations about the forward axis. The trunk and

pelvis mechanism is designed to allow the patientsnatural walking body dynamics to occur unim-peded while providing safety.

Postural control may be used by the clinician toset the trunk and pelvic components independent-ly of each other to maintain a person in a desiredposture. The prescribed posture is then actively(under computer and motor control) maintainedby the application of bias forces to the patientstorso. In addition, the clinician may perturb thepatient by, for example, pushing on the patients

shoulders or hips. However, even though the de-

vice does not actively perturb the patient, it will ac-tively monitor and allow the patient to make mo-tions necessary to recover from the perturbationwhile preventing him or her from falling outsideof the safety zone if the patient is unable to recoverfrom the perturbation. Finally a stabilization func-tion defines how much support the device givesthe patient at the trunk level. This can be adjusteddepending on how much stability the clinicianfeels the patient requiresfrom a somewhat rigidembrace to completely free.

We have discovered several significant chal-lenges in this design cycle. One challenge wasto design an appropriate pelvis interface that ac-commodates a wide variety of body shapes while

allowing the necessary degrees of freedom andcontrol from the robot. Another challenge was toinstantaneously move a heavy (approx 400 lbs)robotic mechanism, designed first for safety, inresponse to very small motions and forces ex-erted by the patient. Adaptive haptic algorithmsimproved this transparency aspect of the device.The controls also had to accommodate specificmodes of operation, because an impending fall isdefined differently for walking than for the sit-to-stand motion (Figure 3).

Future Goals

A currently funded technology transfer researchgrant from the National Institute of Child Healthand Development/National Institutes of Health isnow dedicated to showing that this device doeswhat it was designed to accomplish. More specifi-cally, does this device get in your way when youretrying to move freely? Ideally, the device shouldnot resist at all and should follow a person along,

however such idealizations are impossible and thelimits of the machine are currently being quanti-fied. We are identifying how some tasks are moredifficult for perfect control. For example, initiationof walking seems to produce a small amount of re-sistance in healthy people and in stroke patients.

A second goal addresses the important questionof how walking therapy can influence free walkingspeed and time to fatigue. Finally, the KineAssistallows us to address some more novel and radicalideas, such as the concept of repeated challenge asa therapy, where the therapist deliberately makes


7/9


walking more difficult in training using perturba-tions and forces that destabilize. When the patientleaves such rehabilitation settings, we hypothesizethat they will have learned more and perhapshave minimized a motor deficit. Such novel buteasily programmable strategies are now possiblewith such a robotic device, and a new and widelyimaginative set of possibilities are now realizable

with this new technology.

Acknowledgments

We wish to thank Rehabilitation Institute ofChicago and the US National Institute of Standardsand Technology (NIST) of the US Department ofCommerce Advanced Technology Program for ini-tial support of this work. Amy Schwartz and BenRush of IDEO-Chicago skillfully guided our userneeds analysis and choice of technical target. Weappreciate the contributions of Tom Moyer and

Doug Schwandt to the mechanical design of theKineAssist. We thank the many clinicians, admin-istrators, and patients who shared their perspec-tives, experiences, and expertise with us.

This work was supported by the US Depart-ment of Commerce NIST Advanced Technol-ogy Program, award 70NANB3H3003, NIH R44HD051240-01, and by the Rehabilitation Insti-

tute of Chicago. Some of the robotic resources inour department were also supported by NIDRRH133E020724-03. The authors have financialinterests in and/or are employed by and/or are con-sultants to Kinea Design LLC, Evanston, Illinois.Kinea Design LLC was formerly namd Chicago PT.More information, including videos of the KineAs-sist at work, may be found at www.kineadesign.com (formerly www.chicagopt.com) and at www.smpp.northwestern.edu/robotLab. KineAssistis aregistered trademark. A US patent application hasbeen filed on the KineAssist.

Figure 3.The KineAssistin action as an individual plays catch with the therapist. (Both the patient andthe physical therapist are employees of Kinea Design LLC.)


8/9


REFERENCES

1. Perry J, Garrett M, Gronley JK, Mulroy SJ. Classifica-tion of walking handicap in the stroke population.Stroke.1995;26:982989.

2. Jorgensen HS, Nakayama H, Raaschou HO, et al.2. Jorgensen HS, Nakayama H, Raaschou HO, et al.Outcome and time course of recovery in stroke. PartI: Outcome. The Copenhagen Stroke Study. ArchPhys Med Rehabil.1995;76:399405.

3. Jorgensen HS, Nakayama H, Raaschou HO, et al.3. Jorgensen HS, Nakayama H, Raaschou HO, et al.Outcome and time course of recovery in stroke. PartII: Time course of recovery. The Copenhagen StrokeStudy. Arch Phys Med Rehabil. 1995;76:406412.

4. Forster A, Young J. Incidence and consequencesof falls due to stroke: a systematic inquiry. BMJ.1995;311:8386.

5. Keenan MA, Perry J, Jordan C. Factors affecting bal-ance and ambulation following stroke. Clin Orthop.1984;165171.

6. Lehmann JF, Esselman PC, Ko MJ, Smith JC, deLateurBJ, Dralle AJ. Plastic AFOs: evaluation of function.Arch Phys Med Rehabil.1983;64:402407.

7. Olney S, Richards C. Hemiparetic gait follow-ing stroke. Part I: Characteristics. Gait Posture.1996;4:136148.

8. Rogers MW, Hedman LD, Pai YC. Kinetic analysis ofdynamic transitions in stance support accompany-ing voluntary leg flexion movements in hemipareticadults. Arch Phys Med Rehabil.1993;74:1925.

9. Perry J. The mechanics of walking in hemiplegia.Clin Orthop.1969;63:2331.

10. Wagenaar RC, Beek WJ. Hemiplegic gait: a kinemat-ic analysis using walking speed as a basis.J Biomech.1992;25:10071015.

11. Brandstater ME, de Bruin H, Gowland C, Clark BM.Hemiplegic gait: analysis of temporal variables. ArchPhys Med Rehabil. 1983;64:583587.

12. Dettmann MA, Linder MT, Sepic SB. Relationshipsamong walking performance, postural stability, and

functional assessments of the hemiplegic patient.Am J Phys Med Rehabil.1987;66:7790.

13. Knutsson E, Martensson A, Gransberg L. Influencesof muscle stretch reflexes on voluntary, velocity-controlled movements in spastic paraparesis. Brain.1997;120(pt 9):16211633.

14. Bohannon RW, Walsh S. Nature, reliability, andpredictive value of muscle performance measuresin patients with hemiparesis following stroke. ArchPhys Med Rehabil. 1992;73:721725.

15. Bohannon RW, Andrews AW. Correlation of kneeextensor muscle torque and spasticity with gaitspeed in patients with stroke. Arch Phys Med Rehabil.1990;71:330333.

16. Roth E, Merbitz C, Mroczek K, et al. Hemiplegicgait: relationships between walking speed andother temporal parameters. Am J Phys Med Rehabil.1997;76:128133.

17. Barbeau H, Rossignol S. Recovery of locomotionafter chronic spinalization in the adult cat. Brain Res.1987;412:8495.

18. de Leon RD, Hodgson JA, Roy RR, Edgerton VR.Full weight-bearing hindlimb standing followingstand training in the adult spinal cat.J Neurophysiol.

1998;80:8391.

19. Lovely RG, Gregor RJ, Roy RR, Edgerton VR. Effectsof training on the recovery of full-weight-bear-ing stepping in the adult spinal cat. Exp Neurol.1986;92:421435.

20. Barbeau H, Wainberg M, Finch L. Description andapplication of a system for locomotor rehabilitation.Med Biol Eng Comput.1987;25:341344.

21. Hesse S, Bertelt C, Jahnke MT, et al. Treadmill train-ing with partial body weight support comparedwith physiotherapy in nonambulatory hemipareticpatients. Stroke.1995;26:976981.

22. Hesse S, Malezic M, Schaffrin A, Mauritz KH. Resto-ration of gait by combined treadmill training andmultichannel electrical stimulation in non-ambula-tory hemiparetic patients. Scand J Rehabil Med.1995;27:199204.

23. Sullivan KJ, Knowlton BJ., Dobkin BH. Step training

with body weight-support: Effect of treadmill speedand practice paradigms on post-stroke locomotorrecovery. Arch Phys Med Rehabil.2002 83:683691

24. Kosak MC, Reding MJ. Comparison of partial bodyweight-supported treadmill gait training versus ag-gressive bracing assisted walking post stroke. Neu-rorehabil Neural Repair.2000;14:1319.

25. Nilsson L. Carlsson J, Danielsson A, et al. Walk-ing training of patients with hemiparesis at anearly stage after stroke: a comparison of walkingtraining on a treadmill with body weight supportand walking training on the ground. Clin Rehabil.2001;15:515527.

26. Teixeira Da Cunha Filho I, Lim PAC, Qureshy H,Henson H, Monga T, Protas EJ. A comparison of

regular rehabilitation and regular rehabilitation withsupported treadmill ambulation training for acutestroke patients. J Rehabil Res Dev. 2001;38:245255.

27. Visintin M, Barbeau H, Korner-Bitensky N, Mayo NE.A new approach to retrain gait in stroke patientsthrough body weight support and treadmill stimula-tion. Stroke.1998;29: 11221128.

28. Colombo G, Joerg M, Schreier R, Dietz V. Treadmilltraining of paraplegic patients using a robotic ortho-sis.J Rehabil Res Dev.2000;37:693700.

29. Veneman JF, Ekkelenkamp R, Kruidhof R, van derHelm FCT, van der Kooij H. Design of a series elastic-and Bowdencable-based actuation system for useas torque-actuator in exoskeletontype training. In:Proceedings of the IEEE 9th International Conferenceon Rehabilitation Robotics (ICORR); 2005; Chicago,IL:496499.

30. Schmidt H, Werner C, Bernhardt R, Hesse S, KrgerJ. Gait rehabilitation machines based on program-mable footplates. J Neuroengineering Rehabil.2007;4:2.

31. Reinkensmeyer DJ, Wynne JH, Harkema SJ. (2002) Arobotic tool for studying locomotor adaptation andrehabilitation. In: Proceedings of the IEEE Engineeringin Medicine and Biology Conference (EMBC); 2002;Houston, TX.

32. Peshkin MA, Colgate JE, Wannasuphoprasit W, et al.Cobot architecture. IEEE Trans Robotics Automation.

2001;17:377390.


9/9


SUGGESTED READINGS

Behrman AL, Harkema SJ. Locomotor training after hu-man spinal cord injury: a series of case studies. PhysTher. 2000;80:688700.

Bobath B. Adult Hemiplegia: Evaluation and Treatment.2nd ed. London: William Heinemann Medical BooksLtd; 1978.

Burgar CG, Lum PS, Shor PC, Machiel VHF. Develop-ment of robots for rehabilitation therapy: the Palo

Alto VA/Stanford experience. J Rehabil Res Dev.2000;37:663673.

Dobkin BH. Overview of treadmill locomotor trainingwith partial body weight support: A neurophysi-ologically sound approach whose time has come

for randomized clinical trials. Neurorehabil Repair.1999;13:157165.

Duncan PW, Goldstein LB, Matchar D, Divine GW, Feuss-ner J. Measurement of motor recovery after stroke.

Outcome assessment and sample size requirements.Stroke. 1992; 23(8):10841089.

Edgerton VR, de Leon RD, Tillakaratne N, RecktenwaldMR, Hodgson JA, Roy RR. Use-dependent plastic-ity in spinal stepping and standing. Adv Neurol.1997;72: 233247.

Edgerton VR, Roy RR, de Leon RD, Tillakaratne N, Hodg-son JA. Does motor learning occur in the spinalcord? Neuroscientist. 1997;3:287294.

Finch L, Barbeau H, Arsenault B. Influence of bodyweight support on normal human gait: devel-opment of a gait retraining strategy. Phys Ther.1991;71:842855.

http://www.healthsouth.com/hsus/HSUS/EN_US/cor-porate/abouths/pressroom/autoambulator.jsp

Krebs HI, Hogan N, Volpe BT, Aisen ML, Edelstein L, DielsC. Overview of clinical trials with MIT-MANUS: arobot-aided neuro-rehabilitation facility. Tech HealthCare. 1999;7:419423.

Laufer Y, Dickstein R, Chefez Y, Marcovitz E. The effect oftreadmill training on the ambulation of stroke survi-vors in the early stages of rehabilitation: a random-ized study.J Rehabil Res Dev. 2001;38:6978.

Pohl MM. Speed-dependent treadmill training in am-bulatory hemiparetic stroke patients: a randomizedcontrolled trial. Stroke.2002;33:553558.

Reinkensmeyer DJ, Takahashi CD, Timoszyk WK, Re-inkensmeyer AN, Kahn LE. Design of robot assis-tance for arm movement therapy following stroke.Adv Robotics. 2000;14: 625638.

Richards CL, Malouin F, Wood-Dauphinee S, et al. Task-specific physical therapy for optimization of gaitrecovery in acute stroke patients. Arch Phys MedRehabil.1993;74: 612620.

Visintin M, Barbeau H. The effects of body weight sup-port on the locomotor pattern of spastic pareticpatients. Can J Neurol Sci. 1989;16:315325.

Visintin M, Barbeau H. The effects of parallel bars,body weight support and speed on the modula-tion of the locomotor pattern of spastic pareticgait. A preliminary communication. Paraplegia.1994;32:540553.

PattonEtAl(2008)TopicsInStrokeRehabilitation15(2)131-139

Documents

Transcript of PattonEtAl(2008)TopicsInStrokeRehabilitation15(2)131-139