Design of NJIT-Robot-Assisted Virtual Rehabilitation System to Train the Hemiplegic Upper Extremity of Children with Cerebral Palsy

Qinyin Qiu, MS, Diego A. Ramirez, MS, Soha Saleh, MS, Heta D. Parikh, OT,
Donna Kelly, OT, Sergei Adamovich, PhD.

New Jersey Institute of Technology, Department of Biomedical Engineering
323 MLK Jr. Blvd. Newark, NJ 07102

ABSTRACT

This paper describes the design of NJIT-Robot-Assisted Virtual Rehabilitation System (NJIT-RAVR), which combines adaptive robotics with complex VR simulations for the rehabilitation of upper extremity function in children with CP. Five game-like simulations were developed to facilitate improvement in forearm pronation/supination and arm interjoint coordination during three-dimensional reaching and placement tasks. Arm kinematics and forces were recorded in real time to drive the interaction with virtual environments. In a pilot study, four subjects have been trained for three weeks to examine the proof of concept. Clinical testing was performed before and after the training to assess motor control and real-world upper extremity function. Subjects showed increase in movement speed and in smoothness of the hand trajectory. This ongoing study establishes the feasibility of the system for use by young children with mild to moderate hemiplegia.

Keywords:

Cerebral Palsy, Upper Extremity, Robotics, Neuroplasticity, Virtual Reality

BACKGROUND

Cerebral palsy (CP) is a non progressive neurodevelopmental disorder of motor control due to lesions or other dysfunctions of the Central Nervous System (CNS) [(4)]. The involved upper extremity significantly impacts playing and self-care activities such as eating and dressing [(5)]. Constraint induced therapy (CIT) is currently being used in children to accomplish the goals of intensive massed practice and shaping. It has demonstrated the ability to produce major and sustained motor function improvement in children with spastic hemiplegia secondary to CP [(1)].

Virtual environment (VE) is another technology used to accomplish intensive massed practice in children. VE therapy has the capability to create an interactive, motivating environment in which the therapist can manipulate the practice intensity and feedback to create individualized treatments [(9)]. Use of VE is thought to enhance children's motivation, enable age appropriate play/participation and sense of self-efficacy [(6)], which may result in a desire to practice more [(7)].

 This paper describes the NJIT-RAVR system, which combines adaptive robotics with complex VR simulations for the rehabilitation of upper extremity impairments and function in children with CP.

METHODS

Hardware

The Haptic Master (Figure 1a) with ring gimbal is a 6 degree of freedom admittance controlled (forced controlled) robot [(3)]. A three-dimensional force sensor measures the external force exerted by the user on the robot. Wrist motions: roll, pitch and yaw, are simultaneously recorded by ring gimbal. The Haptic Master Application Programming Interface (API) allows us to program the robot to produce haptic effects, such as springs, damping and constant global forces. The Haptic Master records position, orientation, force and velocity in three dimensions at a rate of up to 1000 Hz, allowing the movement arm to act as an interface between the participants and the virtual environments. 

Author did not supply
Figure 1: (a) The Haptic Master with ring gimbal; (b) forearm and hand based volar splints (Click for larger view)

Three different sized forearm and hand based volar splints (Figure1b) were fabricated to secure the hand to the ring gimbal for children between 5 to 16 years old. The hand based splints allowed for free movement of the digits and wrists and the forearm based splints allowed free movement of the digits while keeping the wrist immobilized. These splints increase stability and control while playing the games and prevent atypical movement patterns.

Participants were positioned in a pediatric Leckey Contour Seating system (Leckey®, UK) with modular foot supports. Their trunks were secured using an H-harness and a 45 degree seat belt. The subjects needed support for feet, seat belt for hips and chest vest to prevent trunk lateral movements and compensation of abnormal patterns of movement. The height of the Leckey Chair was oriented to the HapticMaster in order to obtain a starting position of approximately 90 degree of elbow flexion with humerus adducted to the trunk and the forearm rotated to a position of comfort according to the participants’ available active forearm range of motion.

Simulations

Bubble Explosion:

This simulation focuses on improving the speed and accuracy of shoulder and elbow flexion/extension during aiming and reaching movements towards a target. The participant moves a virtual cursor in a three-dimensional space in order to touch a series of ten haptically rendered bubbles to make them explode.

Cup Reach:

This simulation focuses on improving general upper extremity endurance to improve reaching accuracy. The participant uses their virtual hand (hemiparetic side) to lift virtual cups and place them onto one of three spots on three shelves with varying heights. Amount of resistance of the haptic arm can be adjusted simulating increased weight of the cups.

HammerHM:

This simulation focuses on improving forearm pronation and supination during shoulder flexion and elbow extension. In the simulation, the subject has to repeatedly rotate the wrist to control the virtual hammer on the screen to hammer the wood down into the ground.

CarRaceHM:

This simulation accepts inputs from the robot to command the cars in the game. Pronation and supination movements of the ring gimbal relative to a starting position turn the car a desired angle. A slight force either forwards (shoulder flexion with elbow extension) or backwards (shoulder extension with elbow flexion) is used to accelerate or stop the car respectively.

Participants

Four children with spastic hemiplegia secondary to Cerebral Palsy (CP) were recruited from the outpatient department of a comprehensive pediatric rehabilitation facility. Pre-participation data is summarized in Table 1.

Table 1: Subject information
Subjects

Age

Sex

Cognition

Diagnosis

Dominant Hand

S1

8y

M

Mild mental retardation

Left hemi

RIGHT

S2

16y

M

Within normal range

Right hemi

LEFT

S3

10y1m

M

Within normal

Left hemi

RIGHT

S4

7y7m

F

Within normal

Right hemi

LEFT

Training Procedure

Participants used the NJIT-RAVR System for one hour, 3 days a week for three weeks. For performance testing purposes, subjects performed four sets of ten reaches utilizing the Bubble Explosion simulation to initiate each session. The subjects played a combination of three or four of the other simulations depending on their therapeutic goals, tolerances and preferences for the remainder of the sixty minute session.

Measurements

Clinical testing was performed just prior to and immediately following the training period to assess motor control and real-world upper extremity function. Measurements included Melbourne assessment, upper extremity active range of motion and strength. Kinematics and kinetic measurements including hand movement speed, smoothness of endpoint trajectory, and movement duration were calculated using data collected by the robot during the Bubble Explosion activity on the first and the last day of training as well as at the first day of each training week.

RESULTS

Three participants completed 9 hours training in 3 weeks. One participant completed 8 hours training because he had cold-related fever in the second day of the training and missed one session. One participant (S4) showed improvement on the overall Melbourne Assessment test score (see Table 2) while subjects S2, S3 and S4 showed improvement on three timed Melbourne subtests. S3 showed 15 degrees increase on shoulder flexion (from 0-130 degrees to 0-145 degrees), and 50 degrees increase on forearm pronation/supination (from 60-90 degrees to 10-90 degrees). Both S3 and S4 had almost 100% increases on strength test. S3’s lateral pinch strength increased from 2 lbs to 4 lbs, and 3-jaw pinch strength increased from 1 lb to 2 lbs. S4’s grip strength increased from 6 lbs to 16 lbs, lateral pinch strength increased from 3 lbs to 7 lbs, and 3-jaw pinch strength creased from 1 lb to 2 lbs. All participants showed improvement on several kinematics measures of the movement recorded directly by the robot. The percentage of improvement between pre-test and post-test is shown in Table 3.

Table 2: Percentage score of the Melbourne Assessment clinical test, plus the results of three timed subtests

Subjects

Melbourne assessment

Reach Forward time (s)

Reach sideways Time (s)

Hand to Mouth Time (s)

 

Pre

Post

Pre

Post

Pre

Post

Pre

Post

S1

40.2%

42.6%

1.9

1.2

n/a

1.4

n/a

n/a

S2

74.6%

75.4%

3.6

2.9

3.7

2.3

15.1

10.7

S3

76.2%

77.1%

4.5

1.5

2.4

1.8

2.2

1.6

S4

59.8%

67.2%

2.9

1.5

2.2

0.8

5.4

4.6

Table 3: Percent improvement in the kinematic measures of the arm reaching in three-dimensional space recorded directly by the robot
Subjects

Duration (s)

Path Length (m)

Smoothness

S1

6.53%

13.46%

21.09%

S2

32.30%

4.39%

49.26%

S3

67.01%

64.25%

93.87%

S4

0.94%

18.02%

-0.99%

DISCUSSION

This study establishes the feasibility of this system for use by young children with mild to moderate hemiplegia secondary to CP. All subjects completed 9 hours of training without ill effects and all subjects demonstrated improvement on the timed subtests of the Melbourne test, and three subjects improved on the kinematics of the reaching test. The combination of adaptive robotics and game-like virtual environments offers promise in the ability of both approaches to expand the volume and intensity of practice a participant can perform. Adaptive robotics will allow lower functioning children to access these training activities, while engaging game-like environments will facilitate attention to task. As these children fatigue physically and their performance decays, assistance levels provided by the robot can increase adaptively, allowing the subjects to complete the number of repetitions that might be necessary to facilitate beneficial cortical adaptations and promote a positive impression of the training experience.

REFERENCES

  1. Charles J. R., Wolf S. L., Schneider J. A., Gordon A. M. (2006). Efficacy of a child-friendly form of constraint-induced movement therapy in hemiplegic cerebral palsy: a randomized control trial. Dev Med Child Neurol, 48, 635-642.
  2.  Cusick A., Vasquez M., Knowles L., Wallen M. (2005). Effect of rater training on reliability of Melbourne Assessment of Unilateral Upper Limb Function scores. Dev Med Child Neurol, 47, 39-45.
  3.  Moog FCS Corporation (2006), Haptic Master, http://www.fcs-cs.com
  4.  Morris C. (2007). Definition and classification of cerebral palsy: a historical perspective. Dev Med Child Neurol Suppl, 109, 3-7.
  5.  Reid D. (2004). The influence of virtual reality on playfulness in children with cerebral palsy: a pilot study. Occup Ther Int., 11, 131-144.
  6.  Reid D. T. (2002). Benefits of a virtual play rehabilitation environment for children with cerebral palsy on perceptions of self-efficacy: a pilot study. Pediatr Rehabil, 5, 141-148.
  7.  Rizzo A., Kim G. (2005). A SWOT analysis of the field of virtual reality rehabilitation and therapy. Presence, 14, 2005.
  8.  Rohrer B., Fasoli S., Krebs H. I., Volpe B., Frontera W. R., Stein J., Hogan N. (2004). Submovements grow larger, fewer, and more blended during stroke recovery. Motor Control, 8, 472-483.
  9.  Sveistrup H. (2004). Motor rehabilitation using virtual reality. J Neuroeng Rehabil, 1, 10.

Acknowledgment

This study is supported by the National Institute on Disability and Rehabilitation Research RERC on Technology for Children with Orthopedic Disabilities (Grant # H133E050011). Special thanks to Children’s Specialized Outpatient Center at Tom’s River, New Jersey for their support during the study.

Author Contact Information:

Qinyin Qiu, MS, Department of Biomedical Engineering, New Jersey Institute of Technology, University Heights, Newark, NJ 07102. Office Phone (973) 642-4098, EMAIL: qq4@njit.edu