Potential of Low-Cost Decoupled, Bilateral Steering System to Assess Arm Bias after Stroke

Ruta P. Paranjape, BS1, 2 and Michelle J. Johnson, PhD.1, 2,3
1Rehabilitation Robotics Research and Design Lab, Zablocki VA Medical Center, Milwaukee, WI.
2Department of Biomedical Engineering, Marquette University, Milwaukee, WI.
3Department of Physical Medicine and Rehabilitation, Medical College of Wisconsin, Milwaukee, WI


Stroke is the leading cause of disability for adults in the United States and often results in hemiparesis or weakness of the arm opposite to the side of the lesion. This weakness often leads to a psycho-physiological condition called learned nonuse. Currently there are five million people living in US with post-stroke effects, and this number continues to rise in spite of new life saving medical advances. Because the death rate after stroke has reduced remarkably, there is an extra need to develop training paradigms to improve the functional capacity of impaired arm after stroke. Quantification of the extent of disability is the first step in the development of a new paradigm of therapy. These quantification methods have to be effective, the environment used to test has to be reliable, cost effective, repeatable and precise. This paper reports on the potential of a novel bilateral steering environment with two wheels to quantify learned bias after stroke.


Robot Therapy, Bimanual Co-ordination, Stroke Rehabilitation, Upper arm


Stroke is the leading cause of disability for adults in the United States [(1)]. A common impairment after stroke, hemiparesis, is the leading cause of motor impairment and functional disability in the upper and lower limbs. The disabilities caused by hemiparesis often disrupt the individual’s ability to use both the arms effectively and spontaneously. The side opposite to brain injury gets severely affected resulting in compromised use of that particular side. This phenomenon is termed as learned nonuse.. There is a need to implement the strategies, which will initiate coordinated and spontaneous use of both the arms in stroke survivors.

Force and position-based metric derived have been used to quantify impairment after stroke [(2), (3), (4), (5), (6)]. In addition to force and position matrices quantification of impairment in terms of strength, movement time, reaction time, and speed of movement is calculated in coupled wheel environment [(7)]. Their focus was on improving the coordinated use of both the arms in a coupled environment [(7), (8), (9)].Johnson and colleagues used Driver’s SEAT, a custom steering environment with a split-steering wheel to measure torques exerted by the impaired and non impaired arm of subjects during bimanual and unimanual steering [(8)]. Differences in these torques indicated the presence of learned nonuse. Subjects were encouraged by the distortion to apply effort beyond their initial capacity. This quantification method uses coupled driving environments in bimanual training. Coupled environment project the use of both the hands together but fail to list the role of an individual arm in the whole process. Decoupled wheel environment (two separate wheels used for each arm moment and combined movement projected on the screen as cursor movement) will help quantify the individual role of each arm during target acquisition tasks. Studies show in the decoupled training environment reliability of impairment arm increases and non impaired arm decreases at the end of the training session [(9), (10)]. Thus decoupled environments quantify role of each arm in the given task and are proved useful in further therapy selection.

The long term goal of this research study is to quantify learned nonuse in stroke survivors and handedness in able bodied subjects using different platforms and develop an effective training paradigm, which is home based, easily accessible, simple and low cost and effective in reversing this learned bias in stroke survivors and handedness in able bodied subjects. This paper reports on evaluation of the sensitivity of our decoupled bimanual steering robotic system to detect changes in arm preference and accuracy in able-bodied persons with strong right handedness which will mimic hemiparetic subjects after stroke and discusses implications for stroke assessment.


BiTheraDrive Hardware and Software

PHOTO 1(BiTheraDrive System) shows the BiTheraDrive system setup: a-Commercially available Logitech Wheels b-Wooden arm rest resting on side metal frame c-Height adjustable front metal bar d-Graphical display of the unidirectional tracking task.) Fig 1 shows the BiTheraDrive system setup a-Commercially available Logitech Wheels b-Wooden arm rest resting on side metal frame c-Height adjustable front metal bar d-Graphical display of the unidirectional tracking task. (Click image for larger view)

The BiTheraDrive system as shown in Fig.1 is used as the experimental apparatus for the study protocol. The existing TheraDrive system, which consisted of front height adjustable bar for wheel mount and side bar frame [(11)], was adapted to accommodate a decoupled steering strategy. Theradrive front bar was replaced with a height adjustable bar structure, which is able to acccept two commercially available force-feedback Logitech wheels. Wheels are mounted symmetrically from the midline and the positions of the wheels can be adjusted horizontally on the bar. The adjustable side metal frames of the Theradrive structure is used to mount a wooden support. Subjects use this support to rest the elbow during and between trials. The wheels are connected to a custom software program, which records the angular movement of the wheels as the subjects complete the tracking tasks displayed on the screen. The change in the wheel angle is transferred to the computer. Data transfer protocol is built in feature of the used wheels. The signal from the wheel is given to the program in the MATLAB (Simulink) and is displayed in the form of graphical information on the Visual Display as seen in Figure 2.Subjects were asked to move the cursor to the target using one wheel or both the wheels as per the mode of tracking. The graphical representation of the movement is shown in Fig 2. A custom MATLAB program was written to extract the position data from the file and analyze different parameters such as RMS error and contribution.

Procedure and Data Analysis

PHOTO 2(shows the direction of wheel movement during the given tracking task. Subjects move the wheel from A to B). Fig 2 shows the direction of wheel movement during the given tracking task. Subjects move the wheel from A to B. (Click image for larger view)

The BiTheraDrive study consists of a two day protocol. Day1 is a training and initial assessment day and Day-2 is an intervention and assessment day. Able-bodied subjects were seated at a comfortable distance from the wheel. Subjects were asked to hold on to a Vertical-Gripper through the study protocol (unimanual as well as bimanual steering task). Subjects were asked to perform unidirectional tracking task Fig 1where they have to move the wheels symmetrically and rhythmically from point A to B as shown in Fig 2 for 30 times for each of three modes (unimanual right(UR), unimanual left(UL), bimanual (BR, BL)). Both the days started with a pre assessment sessions and ended with post assessment sessions, where the bimanual as well as unimanual target acquisition tasks were performed. During day 1 subjects were tested with parameters such as dynamically varying force, target distance and weightings. Subjects perform 100 trials of each. Day 1 was 90 min session. In this paper, we are presenting day 1 pre and post assessment results for tracking accuracy and arm contribution. A contribution dependent metric for bimanual as well unimanual task was developed (Table 1). Tracking accuracy was derived from root mean square error (Table 1). This metric will give insight into a bias for dominant and non impaired arm in normal and stroke survivors. The data was averaged across the thirty trials. The position data is collected for pre and post assessment sessions. The Root Mean Square (RMS) error was calculated for the bimanual as well as unimanual target acquisition tasks. Percentage Contribution of each arm in bimanual and unimanual mode was calculated by dividing individual contribution by half the target distance in bimanual mode and by target distance in unimanual mode.

Nine neurologically normal subjects participated in the study with written informed consent. The study was approved by the Institutional Review Board at Marquette University. The average age of all subjects was 24 years (Standard Deviation of 2.96). All the subjects were strongly right handed as assessed by the Handedness Questionnaire [(12)].

Table1: Pre and Post Assessment parameters such as Contribution in % and Error in degrees
  Pre-Assessment Post-Assessment
  Contribution(%) Error(Degrees) Contribution(%) Error(Degrees)
  Contribution Standard Deviation Max Min RMS error Contribution Standard Deviation Max Min RMS error


Bimanual tracking tasks:

GRAPH1(RMS Error Plot Comparing Pre and Post Assessment Sessions for Bimanual Right (BR), Bimanual Left (BL), Unimanual Right(UR) and Unimanual Left(UL) mode RMS error ) Graph 1 RMS error plot comparing pre and post assessment sessions for bimanual right (BR), bimanual left (BL), unimanual right (UR) and unimanual left(UL) mode RMS error (Click image for larger view)

It can be seen from Graph 1 that the data collected for bimanual tasks, errors for both the arms together as well as individual arm was decreased remarkably in post assessment sessions. It can be seen from the Graph 2 that the contribution of dominant arm was decreased from pre to post session and contribution of non dominant arm was increased from pre to post assessment session. Unimanual Tracking Tasks: It can be seen from Graph 1 the rms error for dominant arm was increased and non dominant error was decreased from pre to post assessment session. It can be seen from Graph 2 that the contribution of dominant arm was increased slightly and contribution of non dominant arm was decreased.

GRAPH2(Contribution Plot in % for BR, BL, UR and UL) Graph 2 Contribution plot in % for BR, BL, UR and UL (Click image for larger view)


RMS error, a position based metric, was used to quantify accuracy of the arm use in able bodied subjects. In bimanual mode minimization of rms errors indicate the increased accuracy of both the hands in that mode. Minimization of errors in bimanual task in both the hands proves that the system is sensitive to change and subject became familiar with the system at the end of the post session. This also proves the ability of the bimanual tracking task to improve accuracy in both the arms together. Non dominant arm errors are maximum for all the tasks which indicate the presence of strong bias in the given population. The given system is accurate in quantifying the existing handedness in the given subject population. In unimanual mode error was increased from pre to post session in dominant tracking and was decreased in non dominant tracking. This must have happened due to the interaction with the system during the whole day. This proves that the system is simple to get used to. In bimanual mode, contribution of dominant arm as well as non-dominant arm was decreased but non dominant arm was overshooting thought the day. Subjects didn’t undergo any training protocol hence this result is due to the familiarization of the system. In unimanual mode, contribution of dominant arm as well as non dominant arm remained almost the same from pre to post assessment sessions.


Despite our small population we can confirm that the system is a good quantification tool for the measurement of handedness. Different training protocols designed on this system will play an important role in reversing the existing bias in able bodied as well stroke survivors. System can be used as home based low cost training tool. Statistical significance is still unknown for the given results. We will need more subjects to prove any statistical significance. This system with smaller modification can be used to quantify learned bias after stroke.


We would like to thank Department of Physical Medicine & Rehabilitation, Medical College of Wisconsin (WI), Falk Neurorehab, Marquette University (WI) and Whitaker foundation. We thank Dominic Nathan BS, Xin Feng MS and Shantanu Karnik BS for their help with the experimental setup and data collection for BiTheraDrive System. We also would like to thank Dr. Judith Kosasih and Jayne Johnston, RN, OTR for clinical assistance


Michelle J. Johnson, PhD
Rehabilitation Robotics Research and Design Lab
Zablocki VA Medical Center
5000 W. National Ave.
Milwaukee, WI, 53295.


  1. G. Gresham and P. W. Duncan (1995). Post-stroke Rehabilitation. Clinical Practice Guideline. Washington DC: US Department of Health Services, AHCPR Publication vol. 16, no. 95-0662.
  2. James H. Cauraugh Sang Bum Kim, Aaron Duley (2005).Coupled bilateral movements and active neuromuscular stimulation:Intralimb transfer evidence during bimanual: Neuroscience Letters 382 ,39-44.
  3. Barbara Quaney.Impaired Grip Force Modulation in the ipsilesional hand after Unilateral Middle Cerebral Artery after Stroke. American Society of Neurorehabilitation.
  4. Anne Martine Bertrand Catherine Mercier, Priscilla Lam Wai hun, Daniel Bourbonnais, and Johanne Desrosiers(2004). Effects of Weakness on Symmetrical Bilateral Grip Force Exertion in Subjects with Hemiparesis: J Neurophysiol 91: 1579–1585.
  5. Anne Martine Bertrand Catherine Mercier, Priscilla Lam Wai hun, Daniel Bourbonnais (2004). Differences in the magnitude and direction of forces during a submaximal matching task in hemiparetic subjects. Exp Brain Res, 157:32-42.
  6. S.H. Kim, P. S. Pohl, C. W. Luchies, A. P. Stylianou, Y. Won(2003). Ipsilateral deficits of targeted movement after stroke. Arch Phys Med Rehabil vol 84, pp.719-724, 2003.
  7. R. D. Jones,I. M. Donaldson, N.B. Sharman (Oct. 1996.). A Technique for Removal of the Visuoperceptual Its Application to Parkinson’s Disease Component from Tracking Performance and Its Application to Parkinson’s Disease” IEEE Transactions on Biomedical Engineering, vol.. 43, pp-1001-1010.
  8. M. J. Johnson, H. F. M. Van der Loos, C. G. Burgar, P. Shor and L. J. Leifer (2005). Experimental Results using Force-Feedback Cueing in Robot-Assisted Stroke Therapy. IEEE Transactions on Neural Systems and Rehabilitation Engineering, vol. 13, no. 3, pp. 335-48.
  9. P. Nair, C. Jadhav, V. Krovi(2003). Development and Testing of a LowCost Diagnostic Tool for Upper Limb Dysfunction. IEEE RSJ, pp. 2600-2605.
  10. S. Rick, J.R. Tresillian, M. Mon-Williams, V. L. Coppard, R. G. Carson (2003). Bimanual Aiming and Overt Attention: One Law for Two Hands. Exp Brain Res, 153:59-75.
  11. M. J. Johnson, M. Trickey, E. Brauer, X. Feng (2003). TheraDrive: A New Stroke Therapy Concept for Home-based, Computer-Assisted Motivating Rehabilitation. IEEE EMBC, pp.4844-4847.
  12. Oldfield RC (1971). The assessment and analysis of handedness: the Edinburgh inventory. Neuropsychologia 9:97-113.


  • Source Ordered
  • No Tables
  • Very Compatible


Disney produced a television show in the mid 1990s called Gargoyles. It's a great show and I'm a big fan. A few years ago Disney started to release the show on DVD. The last release was of season 2, volume 1. That was two years ago. Volume 2 has not been released. Why? Poor sales. So if you should find yourself wanting to support my work, instead I ask you pick up a copy of season 2, volume 1. It's a great show and you might find yourself enjoying it.