Augmented effects of EMG biofeedback interfaced with virtual reality on neuromuscular control and movement coordination during reaching in children with cerebral palsy

Abstract

BACKGROUND: The purpose of the present study was to compare therapeutic effects of an electromyography (EMG) biofeedback augmented by virtual reality (VR) and EMG biofeedback alone on the triceps and biceps (T:B) muscle activity imbalance and elbow joint movement coordination during a reaching motor task

OBJECTIVE: To compare therapeutic effects of an electromyography (EMG) biofeedback augmented by virtual reality (VR) and EMG biofeedback alone on the triceps and biceps muscle activity imbalance and elbow joint movement coordination during a reaching motor task in normal children and children with spastic cerebral palsy (CP).

METHODS: 18 children with spastic CP (2 females; mean±standard deviation = 9.5 ± 1.96 years) and 8 normal children (3 females; mean ± standard deviation = 9.75 ± 2.55 years) were recruited from a local community center. All children with CP first underwent one intensive session of EMG feedback (30 minutes), followed by one session of the EMG-VR feedback (30 minutes) after a 1-week washout period. Clinical tests included elbow extension range of motion (ROM), biceps muscle strength, and box and block test. EMG triceps and biceps (T:B) muscle activity imbalance and reaching movement acceleration coordination were concurrently determined by EMG and 3-axis accelerometer measurements respectively. Independent t-test and one-way repeated analysis of variance (ANOVA) were performed at p < 0.05.

RESULTS: The one-way repeated ANOVA was revealed to be significantly effective in elbow extension ROM (p = 0.01), biceps muscle strength (p = 0.01), and box and block test (p = 0.03). The one-way repeated ANOVA also revealed to be significantly effective in the peak triceps muscle activity (p = 0.01). However, one-way repeated ANOVA produced no statistical significance in the composite 3-dimensional movement acceleration coordination data (p = 0.12).

CONCLUSIONS: The present study is a first clinical trial that demonstrated the superior benefits of the EMG biofeedback when augmented by virtual reality exercise games in children with spastic CP. The augmented EMG and VR feedback produced better neuromuscular balance control in the elbow joint than the EMG biofeedback alone.

Keywords

Arm swing movement central pattern generator cerebral palsy gait training muscle activity

1 Introduction

Muscle imbalance in the elbow joint is an important bio-maker for neuromuscular impairments and associated reaching movement incoordination in children with spastic cerebral palsy (CP) (Johnson, 2002). As many as 3 million children with CP suffer from chronic physical disabilities including age-appropriate reaching and grasping movements due to neuromuscular impairments (Feltham et al., 2010). Neurophysiologically, muscle imbalance is caused by cortical disinhibition in central nervous system disorders, which modulates the reciprocal inhibition in the spinal cord (Isobe, 1989). The inability to mediate the reciprocal inhibition results in overactive or hypertonic activation of the antagonist biceps muscle and inhibitive activation of the agonist triceps muscle, which generates rather inappropriate coactivation during functional reaching (elbow extension) movements (Janda, 1987; Rassafiani & Sahaf, 2011). Subsequently, this impaired motor control or imbalance muscle activation between the agonist and antagonist muscles affect elbow joint arthro/osteokinematic movement coordination during reaching movements, which is often manifested as “incoordinated movement acceleration” (Janda, 1996; Lee et al., 2009, 2013). Biomechanically, recent clinical studies showed altered elbow movement kinematics (Lee et al., 2013) and motor recruitment patterns (Yoo et al., 2014) in children with CP during reaching movements. Kinematic variables including movement time (MT), mean velocity (MV), normalized jerk score (NJS), mean angular velocity (MAV) and NJS of the shoulder, elbow and wrist joint were significantly impaired, but improved after a 10-week functional movement strengthening exercise (Lee et al., 2013). Similarly, a recent electromyography (EMG) study revealed a remarkable increased muscle activation imbalance between the triceps and biceps during elbow extension movements when compared with normal controls (Yoo et al., 2014).

To improve muscle imbalance and associated elbow movement incoordination in children with CP, neurodevelopmental treatment (NDT), constraint-induced movement therapy (CIMT), EMG biofeedback, virtual reality (VR) game exercise, and strengthening or resistance training have been used, but outcome results were variable (Butler & Darrah, 2001; Dursun, Dursun, & Alican, 2004; You et al., 2005). NDT has been conventionally used to normalize abnormal movement patterns by means of utilizing various sensorimotor neurofacilitation, but lacks clinical evidence to support its superiority (Butler & Darrah, 2001). CIMT uses the repetitive practice of the ‘normalized shaping targeted functional movement’, but poses inherent issues with compliance and generalizability (Van der Lee, 2003). Moreover, NDT and CIMT are specialized to normalize functional movement patterns rather than to focus on neuromuscular imbalance (Butler & Darrah, 2001; Van der Lee, 2003). Strengthening training and VR or game exercise, which are often administered in conjunction, effectively mitigate muscle imbalance with enriched motivation and fun in children with CP (Boyd, Morris, & Graham, 2001; Miller & Reid, 2003; Rizzo & Kim, 2005; Shurtleff, Standeven, & Engsberg, 2009; You et al., 2005). However, strengthening training combined with VR or game exercise does not provide accurate or quantified biofeedback about muscle activation imbalance as in the EMG feedback training (Giggins, Persson, & Caulfield, 2013; Yoo et al., 2014). Furthermore, it is extremely difficult to control undesirable compensatory movement patterns during strengthening training or VR or game exercises because of a lack of real-time biofeedback (Chen et al., 2006; Giggins, Persson, & Caulfield, 2013).

A hybrid model of EMG-VR system (QEMG-4XL, Laxtha, Seoul, Republic of South Korea) has recently been developed to provide accurate biofeedback and motivation to restore muscle imbalance between the triceps and biceps during elbow reaching movements in children with CP. The EMG-VR is designed to provide real-time visual feedback about triceps and biceps muscle activation patterns (e.g., onset time and amplitude, force exerted) during functional VR games. The VR games comprised of enjoyable, self-motivating, and functional strengthening exercises for reaching movements. Hence, the present study undertook to determine the effects of EMG feedback and EMG-VR on the triceps and biceps (T:B) muscle imbalance and elbow joint movement coordination during a reaching motor task in normal children and children with spastic CP. The basic premise was that EMG-VR would show greater therapeutic effects on T:B muscle imbalance and movement coordination when compared with EMG feedback alone.

2 Methods

Eighteen children (mean age ± standard deviation = 9.60 ± 2.17 years), including 10 children with spastic CP (4 diplegia, 1 hemiplegia, and 5 quadriplegia; 2 females and 8 males; mean age ± standard deviation = 9.5 ± 1.96 years) and 8 children (3 females and 5 males; mean age ± standard deviation = 9.75 ± 2.55 years) were recruited from a local pediatric rehabilitation center in Gimpo city, Republic of South Korea. The study protocol was approved by the Korea Ministry of Health and Welfare of Human Studies Committee. Written informed consent was obtained from their parents prior to this study. The inclusion criteria were as follows: (1) aged between 7∼15 years old, (2) a diagnosis of spastic CP, (3) manual ability classification system (MACS) level I - III, (4) grade 1 at the Modified Ashworth Scale (MAS), and (5) ability to follow verbal commands. The exclusion criteria were (1) visual disorder, (2) cognitive disorder, (3) a history of orthopedic surgery, and (4) botulinum toxin injections performed in the past 12 months.

The surface EMG (QEMG-4XL, Laxtha, Seoul, Republic of South Korea) was used to determine neuromuscular imbalance of triceps and biceps muscles by analyzing muscle activity amplitudes of triceps and biceps muscles during reaching movements before and after the interventions. The child was comfortably seated at a chair with both arms supported on the arm rests and asked to reach forward to the target (Figs. 1 and 2). The target’s skin location was first prepared by shaving and rubbing the skin using a disposable alcohol-soaked pad. A pair of hydrogel electrodes with Ag/AgCl electrodes was affixed to the triceps and biceps on the more affected side with an inter-electrode distance of 2 cm. The electrodes were applied at the landmarks established in Cram’s surface EMG guideline.

As the child performed three consecutive reaching tasks, raw EMG signals were collected at a sampling rate of 1,000 Hz, and band-pass filtered between 20–450 Hz, and notch-filtered at 60 Hz. The collected EMG signals were filtered and processed by using the “root mean square” value (RMS) using the MyoResearch Master Edition 1.08 XP software (Noraxon, Scottsdale, AZ, USA). The mean of peaks during the reaching movement served as the reference voluntary contraction (RVC) for each muscle activity. The root mean square (RMS) was integrated with a window of 100 ms epoch (Merletti & Parker, 2004) and then normalized with a RVC. The EMG signals were represented as RMS processed percentage of RVC (% RVC). EMG data were recorded three times for 5 seconds during the reaching movement.

A 3-axis accelerometer (3-axis accelerometer, Vernier Software & Technology, Oregon, USA) was used to determine movement coordination during the reaching movement by computing movement acceleration variability. Movement acceleration has been utilized to estimate movement acceleration coordination (Michaelsen et al., 2013) which was concurrently collected with EMG muscle activity. The 3-axis accelerometer sensor was attached on the third metacarpal bone of the dominant or more involved hand where the X-axis represents anterioposterior motion; Y-axis, mediolateral; and Z-axis, vertical motion, respectively. LabPro interface (Vernier Software & Technology, Oregon, USA) was used to connect with the sensor and software (Logger Pro3, Vernier Software & Technology, Oregon, USA) was used to calculate the data and graphically exhibit on the computer monitor. The collected data was exported and stored as a CSV file for further statistical analysis. The Logger Pro software was used to compute X, Y, and Z-acceleration, which was expressed: where represents the changing speed, represents the changing direction, is the unit (outward) normal vector, and R is its instantaneous radius. Among the total acceleration movement data acquired for 24 seconds (1200 data points), the most consistent and representative acceleration data points obtained for 2 seconds of movement (100 data points) was used. To evaluate elbow movement coordination, standard deviations (SDs) and coefficient of variations (CVs: the extent of relative variability in relation to mean) of the X-axis, Y-axis, and Z-axis acceleration movements) were computed and used for further statistical analysis. SD and CV analyses have been used to estimate movement coordination variability and smaller SD and CV indicate better movement acceleration coordination or smoothness (Chen et al., 2012). Additionally, the test-retest reliability for movement acceleration measurement was performed using the 3-axis accelerometer as described in the present protocol and revealed an excellent reliability (r = 0.92).

Strength testing was used to examine triceps and biceps elbow muscle strength (Hislop & Montgomery, 2007) using a PowerTrack IITM commander (J-tech Medical, Salt Lake City, Utah). The biceps muscle strength test was performed in a seated position with back support and the knees and hips flexed at approximately 90 degrees with the forearm supinated and flexed slightly more than 90 degrees. The resistance was given over the flexor surface of the proximal wrist site, and the examiner applied a counterforce over the anterior superior aspect of shoulder to avoid any potential compensation. The muscle test was performed three times and averaged for further analysis. Intra-session (Intraclass correlation coefficient (Intraclass correlation coefficient (ICC1,1 = 0.70) and inter-session (ICC1,1 = 0.79) reliabilities were established (Crompton, Galea, & Phillips, 2007).

The elbow flexion and extension range of motion (ROM) were recorded as the child attempted to perform terminal elbow joint flexion and extension movement from the seated position. The detailed methods and reliability (ICC3,1 = 0.91) of the ROM test was well described elsewhere (Norkin & White, 1995; Rothstein, Miller, & Roettger, 1983). In short, the fulcrum of the goniometer was positioned over the lateral epicondyle of the humerus, with one part of the goniometer along the length of the humerus to the tip of the acromion process and the other part along the length of the radius.

Box and Block Test (BBT) is reliable and commonly was used to measure upper extremity motor function. The child comfortably sat at the chair in front of the table and was asked to move the cubes from one box to another designated box as quickly as possible in a given time (60 s). The total numbers of cubes moved were counted. The test was performed twice, and then the average was calculated. BBT was found to have good inter-rater reliability (ICC 2,1 = 0.99) and test-retest reliability (ICC 2,1 = 0.96) (Platz et al., 2005).

An experimental procedural checklist was used to implement the study consistently. Prior to data acquisition, all of the subjects underwent clinical tests (i.e., ROM, strength testing, BBT), muscle imbalance, and movement coordination experiments before and after the interventions. Muscle imbalance and movement coordination were concurrently determined by the EMG and 3-axis accelerometer measurements, respectively. This study used normal children as controls because no previous data was available to compare with the present data; hence, we collected EMG and movement coordination data from the age-matched, normal children. Interventions consisted of EMG feedback alone and EMG-VR feedback game exercises. The experimental paradigm and flowchart are presented in Fig. 3. A sample of 18 children (8 normal children; 10 children with spastic CP) underwent an identical baseline EMG testing where the normal children served as controls. All children with CP first received one session of EMG feedback (30 minutes), followed by one session of the EMG-VR feedback (30 minutes) after a 1-week washout period.

As in the EMG muscle imbalance testing, a child was comfortably seated at a chair with back support and with 90° hip-knee-ankle flexion during the interventions. Two surface EMG electrodes were attached on the triceps and biceps of the dominant (in normal children) or more involved arm (in children with CP). The hand dominance was determined by asking a preferred hand used during play and activities. The child was then asked to practice a repetitive, cyclic, and maximal range of elbow extension-flexion movement with a neutral wrist joint position using an arm ergometer exercise machine, amounting to 30–40 cycles. For the EMG feedback, real-time audiovisual sensory biofeedback information about muscle activation patterns and muscle imbalance between biceps and triceps were provided during the ergometer-induced elbow extension exercise, which enabled the child to acquire the motor task more explicitly or consciously at the cortical motor control level (Fig. 6). During the EMG-VR feedback, identical EMG feedback information, which was integrated with a “blowing balloon” VR game was provided for triceps muscle activation (Figs. 4 and 5). This VR game was designed to provide a fun and motivating experience for children to perform the elbow extension reaching movement more implicitly or subconsciously without undesirable compensatory movement patterns. The specific goal of the game was to pump up the balloon as the child increased triceps muscle contraction or activity while reciprocally inhibiting the bicep muscle activity.

One-sample Kolmogorov-Smirnov tests were performed to test the assumption of normal distribution. An independent t-test was used to compare the mean differences in EMG muscle activity imbalance data between the normal and children with CP. One-way repeated-measure analysis of variance (ANOVA) was used to assess the potential mean differences in clinical tests, EMG, and 3-dimensional MAC between the EMG and EMG-VR feedback. The Fisher’s least significant difference (LSD) post-hoc was used to examine if the interaction effects were observed. An independent t-test was used to determine mean differences in EMG muscle activity imbalance data and movement acceleration coordination data between normalchildren and children with CP. The collected data was analyzed by a statistical package for the social sciences (SPSS) version 18.0 software (SPSS Chicago, IL, USA). The statistical significance level was set at p < 0.05.

3 Results

The demographic and general clinical characteristics of the 18 children including; age, gender, diagnosis, more affected side, manual ability classification system are presented in Table 1.

The clinical motor function tests collected from 10 children with CP at the pretest and posttest are presented in Table 2. One-way repeated ANOVA was revealed to be significantly effective in elbow extension ROM (p = 0.01), biceps strength (p = 0.01), and BBT (p = 0.03). However, no statistical difference was observed in the elbow flexion ROM and triceps strength (Table 2). Post-hoc analysis yielded significant increases (post-EMG feedback < post-EMG-VR feedback) in the elbow extension ROM (p = 0.03), biceps strength (p = 0.01), and BBT (p = 0.01) (Figs. 7 –9). Post-hoc analysis yielded significant increases (pre-EMG feedback <post-EMG feedback) in the elbow extension ROM (p = 0.02) (Fig. 7). Finally, post-hoc analysis yielded significant increases (pre-EMG-VR feedback <post-EMG-VR feedback) in the elbow extension ROM (p = 0.01) (Fig. 7). However, muscle strength was not affected regardless of the intervention.

One-way repeated ANOVA was revealed to be significantly effective in the peak triceps muscle activity (p = 0.01), but no statistical difference was observed in the peak biceps EMG amplitude (Table 3). Post-hoc analysis yielded significant increases (pre-EMG-VR feedback <post-EMG-VR feedback) in the peak triceps EMG amplitude (p = 0.03) (Fig. 10). Furthermore, post-hoc analysis yielded no significant differences in triceps and biceps muscle amplitudes between pre-EMG feedback and pre-EMG-VR feedback, indicating that the washout following the initial EMG biofeedback was successful. An independent t-test showed no significant difference in T:B ratio between the post-EMG-VR feedback and the normal control data (Table 4).

One-way repeated ANOVA produced no statistical significance in the composite 3-dimensional MAC data (p > 0.05) (Table 5). An independent t-test showed statistical differences in the 3-dimensional X-axis MAC data between the normal pretest and the CP pretest (p = 0.01) as well as between the normal pretest and the post-EMG feedback (p = 0.01). However, the independent t-test showed no statistical difference in the 3-dimensional X-axis MAC data between the normal pretest and post-EMG-VR feedback (p = 0.08) (Table 6).

4 Discussion

The present study compared the immediate effects of the EMG biofeedback and EMG-VR feedback on elbow extension ROM, spasticity, muscle strength, muscle imbalance, and movement acceleration coordination in normal children and children with CP. As anticipated, the EMG-VR feedback showed superior improvements on elbow extension ROM, BBT, muscle imbalance in children with CP. Most importantly, the EMG-VR feedback improved elbow movement, muscle balance and coordination, which were consistently reflected in the hand function test. To the best of our knowledge, the present study represents the first clinical trial to highlight and compare the immediate effects of the EMG feedback and EMG-VR feedback on improving muscle imbalance and abnormal coordination in children with spastic CP. Because there is a dearth of information about the therapeutic effects EMG biofeedback and VR on upper extremity motor function, muscle imbalance, and movement coordination in children with CP, it was difficult to compare our findings with those of previous studies. Hence, we compared the effectiveness of the EMG feedback and EMG-VR with the age-matched control data.

The triceps muscle activation in children with CP was more significantly improved after the EMG-VR feedback than the EMG feedback. Such muscle activation improvements were comparable to normal controls. This results suggest that the EMG-VR feedback was more effective in restoring triceps muscle activation than EMG alone because the EMG biofeedback augmented by VR may have facilitated underactive agonist triceps muscle activation while reciprocally inhibiting hypertonic biceps muscle activation (Yoo et al., 2014). The present findings support previous studies (Bolek, 2003; Chen et al., 2007; Dursun, Dursun, & Alican, 2004; You et al., 2005). Chen et al. (2007) examined the effects of VR on upper extremity kinematics in children with CP. This study showed some improvement in the quality of upper extremity behavior during the VR feedback and these training effects were maintained for about 4 weeks. You et al. (2005) showed promising effects of VR on elbow reaching movements and neuroplasticity in a child with spastic CP. Another study investigated the effects of EMG feedback on gait function in children with spastic CP and found increased triceps surae muscle activity and toe clearance during the swing phase of gait (Bolek, 2003). Similarly, Dursun, Dursun, and Alican (2004) demonstrated that children with CP who received EMG biofeedback decreased more muscle tone and increased ankle joint motion than children who received the exercise only.

The composite movement acceleration coordination analysis reached no statistical difference in X-axis (anterioposterior) MAC between the normal control and post-EMG-VR biofeedback. However, no statistical difference was observed between the interventions. These findings corroborate a previous study that investigated effects of VR training on children with CP and showed increased movement assessment battery for children-2 (mABC-2) score (p = 0.04). Especially, the manual dexterity score increased (p < 0.05) (Sandlund, Waterworth, & Hager, 2011).

The clinical hand motor test demonstrated a significant improvement in ROM and BBT, supporting that improved muscle imbalance and movement coordination may have contributed to overall elbow extension range of motion and functional hand motor performance. This finding was in agreement with Yoo et al.’s preliminary study (2014), which showed more increased muscle activity balance between the triceps and biceps after the EMG-VR intervention than the EMG biofeedback intervention in children with CP. This study’s results showed that triceps activation improved but biceps activation was inhibited during EMG-VR feedback conditions. However, triceps activation was inhibited while biceps activation increased during EMG feedback alone.

Taken together with the previous results, the present findings suggest that EMG-VR feedback helped restore muscle imbalance thereby improving movement coordination while reaching. Moreover, augmented VR appears to increase implicit learning as the children acquired the skill while playing the fun, functional, and motivating game (Chen et al., 2007; Dursun, Dursun, & Alican, 2004; Mumford & Wilson, 2009; You et al., 2005). The jerky and unstable reaching movement trajectories were likely represented at the baseline in different coordinate systems, reflecting the positive movement coordination characteristics of CP. It thus appears that concurrent EMG biofeedback integrated in the VR game may have helped to organize multi-sensory integration for motor planning and execution which requires a sensorimotor control process akin to successful coordinate transformations as evidenced in decreased movement acceleration deviation (Chen et al., 2007; Lee et al., 2010; Mumford & Wilson, 2009; You et al., 2005). However, the strength test showed no significant change from either intervention because the present study involved an immediate training effect.

A few shortcomings in this present study should be considered in future research. First, the present study examined the immediate effect of the EMG-VR in CP and hence the long-term effects remain unknown and warrant further investigation. Second, it may be of interest to investigate other wrist joint muscles and movement coordination because the reaching motor task involves multi-joint motor control and movement coordination. Lastly, the present study demonstrated a relatively superior effect of the EMG biofeedback when it was augmented by virtual reality, but the underlying neural substrates and control mechanism remain unknown. Perhaps, functional neuroimaging studies could help ascertain this mechanism.

5 Conclusion

The present study is the first clinical trial that demonstrated the superior benefits of EMG biofeedback when augmented by virtual reality exercise games in children with spastic CP. The augmented EMG and VR feedback produced better neuromuscular balance control in the elbow joint than the EMG biofeedback alone. These novel findings showed promising clinical merits for children with spastic CP who often experience abnormal neuromotor control and imbalance with associated movement incoordination during reaching movements. EMG biofeedback interfaced with VR feedback may be used to magnify therapeutic effects as an innovative, alternative neuro rehabilitation technique.

Conflict of interest

The authors have no conflict of interest to report.

Footnotes

Acknowledgments

The authors would like to acknowledge financial and administrative supports provided from “Brain Korea 21 PLUS Project (Grant NO. 2016-51-0009)” sponsored by the Korean Research Foundation for Department of Physical Therapy in Graduate School, Yonsei University.

References

Bolek

J. E.

(2003). A preliminary study of modification of gait in real-time using surface electromyography. Appl Psychophysiol Biofeedback, 28(2), 129–138.

Boyd

R. N

, Morris

M. E

, & Graham

H. K.

(2001). Management of upper limb dysfunction in children with cerebral palsy: A systematic review. Eur J Neurol, 8(5), 150–166.

Butler

, & Darrah

(2001). Effects of neurodevelopmental treatment (NDT) for cerebral palsy: An AACPDM evidence report. Dev Med Child Neurol, 43(11), 778–790.

Chen

, Huang

, Xu

, Wallis

R. I

, Rikakis

, & He

(2006). The design of a real-time, multimodal biofeedback system for stroke patient rehabilitation. ACM. Proceedings of the 14th annual ACM international conference on Multimedia 763–772.

Chen

Y. P

, Kang

L. J

, Chuang

T. Y

, Doong

J. L

, Lee

S. J

, Tsai

M. W

, Jeng

S. F

, & Sung

W. H.

(2007). Use of virtual reality to improve aupper-extremity control in children with cerebral palsy: A single-subject design. Phys Ther, 87(11), 1441–1457.

Chen

Y. T

, Pinto

, de

Miranda Marzullo, A. C

, Kennedy

D. M

, Fox

E. J

, & Christou

E. A.

(2012). Age-associated impairment in endpoint accuracy of goal-directed contractions performed with two fingers is due to altered activation of the synergistic muscles. Exp Gerontol, 47(7), 519–526. doi: 10.1016/j.exger.2012.04.007

Crompton

, Galea

M. P

, & Phillips

(2007). Hand-held dynamometry for muscle strength measurement in children with cerebral palsy. Dev Med Child Neurol, 49(2), 106–111 . doi: 10.1111/j.1469-8749.2007.00106.x

Dursun

, Dursun

, & Alican

(2004). Effects of biofeedback treatment on gait in children with cerebral palsy. Disabil Rehabil, 26(2), 116–120 .doi: 10.1080/09638280310001629679

Feltham

M. G

, Ledebt

, Deconinck

F. J

, & Savelsbergh

G. J.

(2010). Assessment o neuromuscular activation of the upper limbs in children with spastic spastic quadriplegic cerebral palsy during a dynamical task. J Electromyogr Kinesiol, 20(3), 448–456. doi: 10.1016/j.jelekin.2009.07.001

10.

Giggins

O. M

, Persson

U. M

, & Caulfield

(2013). Biofeedback in rehabilitation. J Neuroeng Rehabil, 10(1), 60–70 . doi: 10.1186/1743-0003-10-60

11.

Hislop

H. J.

Montgomery

(2007). Daniels and Worthingham’s Muscle Testing: Techniques of Manual Examination. New York: 8th edition: Elsevier Saunders.

12.

Isobe

R. D.

(1989). The effect of EMG biofeedback on cerebral palsy muscle activity. Thesis (Department of Psychology), Simon Fraser University.

13.

Janda

(1996). Evaluation of Muscular Imbalance. Rehabilitation of the Spine. Baltimore: Williams & Willkins.

14.

Janda

(1987). Muscles and Motor Control in Low Back Pain: Assessment and Management. New York: Churchill Livingstone.

15.

Johnson

(2002). Prevalence and characteristics of children with cerebral palsy in Europe. Dev Med Child Neurol, 44(9), 633–640. doi: 10.1017/S0012162201002675.

16.

Lee

D. R

, You

S. H

, Lee

N. G

, You

I. G

, Jung

M. Y

, & Han

B. S.

(2010). Strength training-induced changes in muscle size and motor improvement in bilateral schizencephaly: An experimenter-blind case report with 3-month follow-up. Phys Ther Korea, 17(4), 77–87.

17.

Lee

D. R

, You

S. H

, Lee

N. G

, Oh

J. H

, & Cha

Y. J.

(2009). Comprehensive Hand Repetitive Intensive Strengthening Training (CHRIST)-induced morphological changes in muscle size and associated motor improvement in a child with cerebral palsy: An experimenter-blind study. Neurorehabilitation, 24(2), 109–117. doi: 10.3233/NRE-2009-0460.

18.

Lee

J. A

, You

J. H

, Kim

D. A

, Lee

M. J

, Hwang

P. W

, Lee

N. G

, Park

J. J

, Lee

D. R

, & Kim

H. K.

(2013). Effect of functional movement strength training on strength, muscle size, kinematics and motor function in cerebral palsy: A 3-month follow-up. Neurorehabilitation, 32(2), 287–295. doi: 10.3233/NRE-130846.

19.

Merletti

, & Parker

P. A.

(2004). Electromyography: Physiology, Engineering, and Non-Invasive Applications. New Jersey: John Wiley & Sons.

20.

Michaelsen

S. M

, Gomes

R. P

, Marques

A. P

, Rodrigues

L. C

, Borges

N. G

Junior , Claudino

, & Dos Santos

M. J.

(2013). Using an accelerometer for analyzing a reach-to-grasp movement after stroke. Motriz: J Phys Educ, 19, 746–752. doi: 10.1590/S1980-65742013000400012

21.

Miller

, & Reid

(2003), Doing play: Competency, control,: Competency, control, and expression. Cyberpsychol Behav, 6(6), 623–632.

22.

Mumford

, & Wilson

P. H.

(2009). Virtual reality in acquired brain injury upper limb rehabilitation: Evidence-based evaluation of clinical research. Brain Inj, 3(3), 179–191. doi: 10.1080/02699050802695566

23.

Norkin

C. C

, & White

D. J.

(1995). Measurement of Joint Motion; A Guide to Goniometry. Philadelphia: FA Davis.

24.

Platz

, Pinkowski

, Van

W. F

, Kim

I. H

, di Bella

, & Johnson

(2005). Reliability and validity of arm function assessment with standardized guidelines for the Fugl-Meyer Test, Action Research Arm Test and Box and Block Test: A multicentre study. Clin Rehabil, 19(4), 404–411.

25.

Rassafiani

, & Sahaf

(2011). Hypertonicity in children with cerebral palsy: A new perspective. Iran Rehabil J, 9, 66–74. doi: http://irj.uswr.ac.ir/article-1-217-en.html

26.

Rizzo

, & Kim

G. J.

(2005). A SWOT analysis of the field of VR rehabilitation and therapy. Presence: Teleoperators Virtual Environments, 14, 119–146.

27.

Rothstein

J. M

, Miller

P. J

, & Roettger

R. F.

(1983). Goniometric reliability in a clinical setting: Elbow and knee measurements. Phys Ther, 63(10), 1611–1615.

28.

Sandlund

, Waterworth

E. L

, & Hager

(2011). Using motion interactive games to promote physical activity and enhance motor performance in children with cerebral palsy. Dev Neurorehabil, 14(1), 15–21. doi: 10.3109/17518423.2010.533329.

29.

Shurtleff

T. L

, Standeven

J. W

, & Engsberg

J. R.

(2009). Changes in dynamic trunk/head stability and functional reach after hippotherapy. Arch Phys Med Rehabil, 90(7), 1185–1195. doi: 10.1016/j.apmr.2009.01.026

30.

Van der Lee

J. H.

(2003). Constraint-induced movement therapy: Some thoughts about theories and evidence. J Rehabil Med, 41, 41–45.

31.

Yoo

J. W

, Lee

D. R

, Sim

Y. J

, You

J. H

, & Kim

C. J.

(2014). Effect of innovative virtual reality game and EMG biofeedback on neuromotor control in cerebral palsy. Biomed Mater Eng, 24(6), 3613–3618. doi: 10.3233/BME-141188

32.

You

S. H

, Jang

S. H

, Kim

Y. H

, Kwon

Y. H

, Barrow

, & Hallett

(2005). Cortical reorganization induced by virtual reality therapy in a child with spastic quadriplegic cerebral palsy. Dev Med Child Neurol, 47(9), 628–635.