There is spasticity in 80% of cerebral palsy (CP) cases, ranging from muscle stiffness and pain to an uncontrollable degree of voluntary control [1]. In children with spastic CP, motor dysfunction is more severe in the upper than in the lower extremities. Upper extremity (UE) dysfunction, paralysis of the affected arm, muscle weakness, convulsions, dystonia, and damage to proprioceptive sensation appear [2].

These symptoms result in difficulty reaching or handling objects, dressing, eating, washing, self-care activities, and lower social participation and quality of life [3]. Children with spastic CP gradually ignore the paralyzed hand as they develop and use only the non-paraplegic hand. Neuropsychologically, the movement opportunities of the paralyzed hand are limited using the non-paralyzed hand predominantly as it is easier to move. This is known as non-use syndrome [4].

These neurological mechanisms cause detrimental reorganization of the cerebral cortex and subcortical sensorimotor areas. Additionally, continuous use of only the non-paralyzed hand is accompanied by impaired attention, language, memory, visual perception, and cognitive and motor control [5]. Visual, tactile, and proprioceptive sensory feedback affects most eye-hand coordination tasks [6]. The proper judgment of visual perception and recognition of information about the external environment in daily life is complex in patients with spastic CP. Virtual reality (VR) facilitates learning of limb movements that are necessary for activities of daily living by exposing an individual to environments with sensory and visual perception challenges. [7]. Feedback on VR can be classified as intrinsic or augmented. Augmented feedback provides essential information verbally or non-verbally, and errors can be corrected when defective or distorted information is received after movements [8].

UE function and visual perception disorders in children with CP can be treated using constraint-induced movement therapy, hand-arm intensive bimanual training, neurodevelopment treatment, upper and lower coordination training, and sensation integration therapy [9]. However, these interventions show qualitative differences based on therapists’ handling and proficiency. Moreover, children with CP with severely impaired motor skills show limited training participation and have difficulty receiving sensory and motor inputs for experience-dependent neural plasticity [10].

Electroencephalography (EEG) is beneficial for understanding which brain regions are activated during the movements performed in virtual reality. Particularly, alpha waves dominate during arousal, whereas beta waves are closely related to attention and concentration [11]. Therefore, it is necessary to check the status of the awakening, attention, and concentration of brain activation in immersive VR programs [12].

VR is a system in which learners perform and solve tasks in an environment like real-life using computer software while providing intrinsic and augmented feedback [13]. Real-time task processing in VR stimulates intrinsic feedback in children with CP. A play environment, such as a game, provides augmented feedback by encouraging interest, motivating, and expanding active participation. This improves motor function and movement of paralyzed extremities using motor control and learning [14]. The VR-based immersive UE rehabilitation program facilitates real-time exteroceptor stimulation (visual, auditory, and tactile) for purposeful movements. VR-based immersive UE training is more effective for eye-hand coordination, visual perception, UE function, proprioception, and posture control than traditional VRUE training programs [15]. Studies on visual perception, attention, cognitive function, motor control, and motor learning by applying VR-based intervention to children with CP are lacking [16].

Therefore, this study investigates the changes in UE function, visual perception, and brain activity in children with spastic CP by Leap Motion ® UE training.

1. Patient History

The study participant selected for this case report was a 14-year-old girl with spastic CP who received rehabilitation treatment at secondary hospital H, Gwangju. The patient was 153 cm tall and weighed 47 kg, with typical vital signs. Neurological examination confirmed the existence of CP, but it was otherwise unremarkable. The patient was born at full-term, weighing 3.4 kg, and had a history of periventricular leukomalacia. The patient was admitted to the hospital at the age of 3 years, the neurologist diagnosed spastic CP for dystonia/dyskinesia. The study purpose was explained in detail to the subject and her guardians under the ethical principles of the Declaration of Helsinki, and written consent was obtained before participation.

Generally, the participant had no difficulty changing her posture in daily life and could walk independently. She was able to walk on sloping roads without physical assistance and climb and descend stairs while holding onto the railing. There was no difficulty in fast walking or running activities involving the global muscle groups. Yet, there were limitations in the delicate and sophisticated one-side arm and hand movements involving local muscles.

The participant had a modified Ashworth scale score of 1 and a gross motor function classification system of stage 1. She was at the first stage of the classification systems for manual ability and communication function, without problems noted in communication.

2. Examination

The participant’s grip strength was measured using a grip force meter. Participants measured the right side, the non-paraplegic side first, and the left and paralyzed side using a grip force meter (Jamar hydraulic hand dynamometer, PC 5030 JI, Sammons Preston, Bolingbrook, IL, USA). We employed the measurement method suggested by the American Society of Hand Therapists. Measurements were obtained while the subject was sitting in a chair without armrests with both shoulder joints in a neutral position, the elbow joint at 90° flexion, and the wrist joint in a neutral position. Measurements were repeated three times, and the average value was recorded [17]. The pinch force was measured using a pinch gauge (B&L Engineering pinch gauge, Sammons Preston). Measurements were obtained in the same posture as in the pinch strength test. We measured the maximum tip pinching, three-point pinch, and lateral pinch repeatedly while sitting, and then averaged the results. The grip and pinch internal correlation coefficients for static and dynamic contraction of this evaluation tool were .94–.96 and .91–.98, respectively [18, 19].

The Jebsen-Taylor test of hand function (JTTHF) was performed to examine UE function and agility. This inspection test includes seven items: writing, card turning, small common objects, simulated feeding, stacking checkers, moving large light objects and large heavy objects, and the time required to perform each task was measured. The dominant hand reliability of this evaluation tool was r = .67–.99, and the non-dominant hand reliability r = .60–.92 [20].

The Purdue pegboard test examined the subject's hand agility and coordination. This inspection tool measures the process of inserting and assembling pegs into holes using pins, collars, and washers. The measurement method proceeds in the order of the non-paraplegic side, paralyzed side, and use of both hands. The test-retest reliability of this test tool was reported at r = .60–.79 [21].

Visual perception ability was tested using the motor-free visual perception test-3 (MVPT-3). This instrument measures visual discrimination, spatial relationships, visual memory, figure-ground, and visual closure. It includes 65 items, ranging from 1–40 items up to the age of 10 years and 1–65 items for ages ≥11 years. The subject here was 14 years old and was scored based on all items 1–65, as specified above. Higher scores indicate better visual perception. The inter-inspector reliability of this test tool was reported to be r = .86–.90 [22].

3. EEG Data Acquisition

Brain activation changes were examined using EEG. BIOS-S24 (Bio Brain Inc., Daejeon, Korea) was used for brain waves. The subject's EEG data were stored in a computer at a 256 Hz sampling frequency, .5–50 Hz pass filter, and 12-bit (A/Dg-to-digital A/D) conversion. EEG activity was measured according to the frequency range using band power and power spectral density (PSD) classified by frequency band.

We collected data measuring spontaneous EEG 3 times each for 2 min and measuring induced brain waves for 30 s. The subjects were given a break of 1 min for each data collection. The first and last 5 s of the EEG were removed to exclude the possibility of the subject’s attention deteriorating or the influx of miscellaneous waves.

The electrical signal was transformed from the EEG time to the frequency domain composed of amplitude and frequency using a discrete fast Fourier transform. After identifying the converted signal's PSD, each frequency band's size on the EEG signal was summed and reclassified into the previously defined EEG type. EEG data were collected by attaching 19 electrodes to the subject’s head using the international 10–20 electrode placement method. Electrodes were attached to sensory input and motor execution functional areas Fp1 and Fp2, cognitive thinking processing areas F3 and F4, auditory information receptive areas T3 and T4, motor information receptive areas P3 and P4, and visual perception input areas O1 and O2 areas (Fig. 1).

Fig. 1. Electroencephalography (EEG) based on the International 10–20 system method. Electrodes were attached to Fp1, Fp2, F7, F3, FZ, F4, F8, T3, C3, CZ, C4, T4, T5, P3, PZ, P4, T6, O1, and O2.

The EEG electrodes (Bio Brain Inc.) comprised plate-shaped disks coated with gold. To minimize contact resistance with the skin when attaching the electrodes, foreign substances on the head surface were washed with a pre-prepared alcohol swab, and an electrode paste (paste Z401CE, Laxtha, Daejeon, Korea) was applied and attached. Gauze was placed to cover and fix the attached electrode so that the electrode remained in place. After attaching the electrodes, the subject sat in a comfortable chair in a quiet examination room with no noise and watched the VR screen. When the color change on the screen appeared along with a “beep” sound in VR, the subject extended and touched the object with her paralyzed arm. PSD of the alpha and beta bands were obtained for 30 s at each brain region while the subject reached an object in VR [23].

4. Intervention

The VR devices used here were Oculus Rift (Oculus, Menlo Park, CA, USA) and Leap Motion (LM-010, CA, USA). The VR-based Leap Motion ® for UE training used here is part of an immersive VRUE rehabilitation program using sensors developed by Leap Motion (Fig. 3). The Leap Motion sensor captures and tracks minute movements of the hands and fingers. This device can model the hand and finger kinematics. It is cheaper and more convenient than other instruments [24]. UE rehabilitation training was implemented by wearing Oculus Rift on the face and linking the VR system and Leap Motion. The subject sat in a chair and performed five randomized tasks, including convex stacking, word guessing, putting fruit in a box, jigsaw puzzles, and playing a musical instrument (Fig. 4). The next task was performed when the previous one was completed. The total intervention period was six weeks, and 24 sessions were conducted four times a week. The processing time of the five tasks was 20 min, and the VR UE rehabilitation program was performed twice a day for a total of 40 min, set as one session. The rest period was set as 10 min at the first session end, 20 min, and the second session was performed after the break. Subjects took the Oculus Rift off their face, rested comfortably during the break, and played a video or music to boredom (Fig. 2).

Fig. 2. Experimental procedure.
VR: virtual reality; GST: grip strength test; PST: pinch strength test; JTTHF: Jebsen Taylor test of hand function; PPT: Purdue pegboard test; MVPT-3: motor free visual perception test-3; EEG: electroencephalography

Fig. 3. Oculus Rift and Leap Motion sensor.

Fig. 4. The virtual reality-based Leap Motion training. (A) Main menu screen, (B) blocks of columns, (C) crossword puzzle, (D) puzzle making, (E). playing an instrument, (F) putting fruit in a box.

5. Data Analysis

Statistical tests were performed using SPSS version 21.0. The mean values for each variable before and after the intervention are presented, and the changes are compared and presented as descriptive statistics in tables.

1. Changes in Grip Strength

After intervention, the subject's grip strength increased on her non-paraplegic side from 17.78 kg to 18.72 kg and on the paraplegic side from 2.50 kg to 3.22 kg (Table 1).

Table 1

Changes in grip strength

Grip	Pre-test	Post-test	Difference	Difference (%)
Unaffected side (kg)	17.78	18.72	.94	5.29
Affected side (kg)	2.50	3.22	.72	28.8

2. Changes in the Pinch Function

After the intervention, the subject’s tip pinch increased from 4.67 kg to 5.78 kg on the non-paraplegic side and from 2.11 kg to 3.06 kg on the paraplegic side.

The subject’s three-point pinch increased from 6.06 kg to 6.81 kg on the non-paraplegic side and from 2.94 kg to 3.78 kg on the paraplegic side after the intervention.

After the intervention, the participant’s lateral pinch increased from 7.67 kg to 7.83 kg on the non-paraplegic side and from 4.11 kg to 5.44 kg on the paraplegic side after the intervention (Table 2).

Table 2

Changes in pinch strength

Pinch	Side	Pre-test	Post-test	Difference	Difference (%)
Two pinches	Unaffected side (kg)	4.67	5.78	1.11	23.77
	Affected side (kg)	2.11	3.06	.95	45.02
Three pinches	Unaffected side (kg)	6.06	6.81	.75	12.38
	Affected side (kg)	2.94	3.78	.84	28.57
Lateral pinch	Unaffected side (kg)	7.67	7.83	.16	2.09
	Affected side (kg)	4.11	5.44	1.33	32.36

3. Changes in JTTHF Scores

The subject’s execution time for the writing portion of the JTTHF decreased from 10.21 s to 9.27 s on the non-paraplegic side and from 31.78 s to 22.62 s on the paraplegic side after the intervention.

The subject’s execution time for simulated page-turning decreased from 3.70 s to 3.56 s and from 8.87 s to 5.21 s on the non-paraplegic and paraplegic side after the intervention, respectively.

The subject’s execution time for small common objects decreased from 7.13 s to 6.21 s on the non-paraplegic side and from 15.23 s to 11.24 s on the paraplegic side after the intervention.

The subject’s execution time for simulated feeding decreased from 12.49 s to 8.40 s on the non-paretic side and from 21.55 s to 14.16 s on the paraplegic side after the intervention.

The subject’s execution time for stacking checkers decreased from 4.66 s to 4.35 s on the non-paraplegic side and from 11.12 s to 7.40 s on the paraplegic side after the intervention.

The subject’s execution time for moving large light objects decreased from 3.95 s to 3.24 s and from 6.33 s to 4.71 s on the non-paraplegic side after the intervention.

The subject’s execution time for moving large heavy objects decreased from 3.91 s to 3.22 s and from 6.19 s to 3.99 s on the non-paraplegic side after the intervention (Table 3).

Table 3

Changes in the Jebsen-Taylor test for hand function (JTTHF) scores

Item	Side	Pre-test	Post-test	Difference	Difference (%)
Writing	Unaffected side (s)	10.21	9.27	.94	9.21
	Affected side (s)	31.78	22.62	9.16	28.82
Card turning	Unaffected side (s)	3.70	3.56	.14	3.78
	Affected side (s)	8.87	5.21	3.66	41.26
Small common objects	Unaffected side (s)	7.13	6.21	.92	12.9
	Affected side (s)	15.23	11.24	3.99	26.2
Simulated feeding	Unaffected side (s)	12.49	8.40	4.09	32.75
	Affected side (s)	21.55	14.16	7.39	34.29
Stacking checkers	Unaffected side (s)	4.66	4.35	.92	19.74
	Affected side (s)	11.12	7.40	3.99	35.88
Moving large light objects	Unaffected side (s)	3.95	3.24	.71	17.97
	Affected side (s)	6.33	4.71	1.62	25.59
Moving large heavy objects	Unaffected side (s)	3.91	3.22	.69	17.65
	Affected side (s)	6.19	3.99	2.20	35.54

4. Changes in Purdue Pegboard Test Performance

Before the intervention, the subject spent 14.89 pcs on the non-paraplegic side and 5.33 pcs on the paraplegic side on the Purdue pegboard, and 15.67 pcs on the non-paraplegic side and 5.83 pcs on the paraplegic side after the intervention. There was a 9.21% increase in two-handed performance (Table 4).

Table 4

Changes in the Purdue pegboard test

Side	Pre-test	Post-test	Difference	Difference (%)
Unaffected side (pieces)	14.89	15.67	.78	5.24
Affected side (pieces)	5.33	5.83	.50	9.38
Both hand (pieces)	4.78	5.22	.44	9.21

5. Changes in the MVPT-3

The pre-test score was 106.33 points, and the post-test score was 138.83 points, an increase of 32.05 points. There was a 30.14% increase in MVPT-3 (Table 5).

Table 5

Change in motor-free visual perception

Item	Pre-test	Post-test	Difference	Difference (%)
Standard score	106.33	138.83	32.05	30.14

6. Changes in Alpha And Beta Wave Activity After Intervention

Analysis of the EEG in the PSD of alpha and beta waves revealed distinct changes in brain activity following the intervention. Specifically, the PSD of alpha waves showed a decrease in each brain region (Table 6), while the PSD of beta waves indicated an increase across all regions (Table 7). Mapping these changes before and after the intervention demonstrated that alpha wave activity was suppressed in the forehead area F3 and temporal area T4, whereas beta wave activity increased in the frontal region Fp2, forehead region F4, temporal region T4, parietal region P4, and occipital region O2 (Fig. 5 and 6).

Table 6

Changes in alpha wave

Area	Pre-test	Post-test
Fp1	.13 ± .04	.06 ± .02
Fp2	.11 ± .04	.05 ± .02
F3	.25 ± .12	.09 ± .06
F4	.16 ± .05	.11 ± .05
T3	.14 ± .08	.11 ± .06
T4	.19 ± .03	.11 ± .05
P3	.17 ± .08	.12 ± .06
P4	.02 ± .00	.14 ± .03
O1	.18 ± .05	.12 ± .07
O2	.18 ± .05	.13 ± .05

¹Values are shown as mean ± standard deviation

Table 7

Changes in beta wave

Area	Pre-test	Post-test
Fp1	.11 ± .081	.17 ± .05
Fp2	.13 ± .03	.21 ± .03
F3	.15 ± .07	.21 ± .04
F4	.12 ± .06	.24 ± .07
T3	.14 ± .10	.22 ± .06
T4	.12 ± .06	.25 ± .03
P3	.13 ± .08	.18 ± .11
P4	.11 ± .06	.21 ± .10
O1	.14 ± .07	.21 ± .02
O2	.12 ± .05	.22 ± .05

¹Values are shown as mean ± standard deviation

Fig. 5. Changes in alpha power mapping. (A) Pre-test, (B) Post-test.

Fig. 6. Changes in beta power mapping. (A) Pre-test, (B) Post-test.

This study investigated the effect of VR-based Leap Motion ® UE training on UE function visual perception and brain activity in the participant with spastic CP. UE rehabilitation training using VR can easily adjust the virtual environment and level, stimulating interest and improving UE function in children with CP. VR can provide various sensory feedback and train arm movements necessary for daily life [25]. This study demonstrated improved grip strength, pinch ability, hand agility, hand coordination, and visual perception with VR training. EEG power of the alpha band of brain activity decreased, and the power of the beta band increased, showing a change in brain activity with the VR UE training.

There was no difference in grip strength on the paralyzed side after applying VR-based Leap Motion UE training, but the pinching ability improved. Tarakci et al. [26] investigated the effect of Leap Motion-based control training on UE rehabilitation in children with CP, juvenile idiopathic arthritis, and physical disabilities of the brachial plexus. The authors found a significant difference in grip force between the experimental and control groups, which is inconsistent with the results of a previous study [26]. This study differs from Tarakci's because it focuses solely on children with cerebral palsy and uses an immersive VR program. In contrast, Tarakci's study included various disabilities and used only Leap Motion training. Here, Oculus Rift was worn on the face, and both hands were placed on the Leap Motion sensor to manipulate the task. We speculate that the grip and pinch strength improved since both hands were used to complete the task.

The intervention improved the participant's JTTHF score and execution times of card turning, small common objects, simulated feeding, stacking checkers, moving large light objects, and moving large heavy objects. The results concur with the study of Acar et al. [27], which investigated the effect of a Nintendo intervention combined with neuromuscular development therapy in children with CP and demonstrated the JTTHF scores of all seven participants in the experimental group improved compared to those in the control group. The VR intervention in this study consisted of building blocks, guessing crosswords, putting fruit in boxes, jigsaw puzzles, and playing musical instruments like the tasks tested in the JTTHF. Therefore, improvement in the hand function score may have been caused by the learning effect.

Here, the subject’s Perdue pegboard performance improved with the intervention indicating that both hand agility and coordination improved with VR training on the paralyzed side. This result is consistent with a previous study [28] investigating the effects of occupational therapy interventions and game-based VR video programs on hand agility in patients with multiple sclerosis and showed that the Perdue pegboard test performance improved in the experimental compared to the control group. In this study, the paralyzed arm improved in agility and coordination due to VR intervention that required delicate control of both hands.

The subject’s visual perception also improved with VR training, consistent with a study by Alwhaibi et al. [29]. They investigated the effects of augmented feedback training through physical therapy on visual-motor integration, visual perception, and motor coordination in children with spastic CP. They found that visual perception ability improved in the experimental group compared to the control group. Here, the VR onscreen task required concentration on attention and problem-solving skills. This may activate the subject’s visual perception and visual-motor integration network.

The brain mapping results showed that the alpha waves were remarkably suppressed in forehead area F3 and temporal area T4. Beta waves showed changes in Fp2, frontal region F4, temporal region T4, parietal region P4, and occipital region O2. The data suggest that beta waves are preferentially active during concentration tasks or conscious activities [30]. The encouraging improvement in visual perception observed with the VR-based intervention can be attributed to enhanced attention and problem-solving abilities during task performance, which likely strengthened the visual perception network. This is in line with the findings of Alwhaibi et al. [29] However, the differences in EEG outcomes may be due to the distinct neural activation patterns induced by the immersive VR program compared to the use of the Leap Motion sensor. In contrast, the differences in EEG outcomes may result from the varying neural activation patterns induced by the immersive VR program compared to using only the Leap Motion sensor.

A limitation of this study is that it is a case study with a single participant, which restricts the generalizability of the findings. Additionally, the need for a control group makes it challenging to confirm the effectiveness of the VR intervention based solely on pre- and post-intervention comparisons. However, a key strength of this study is an immersive VR program that provided an engaging and motivating environment for the child with cerebral palsy, leading to positive changes in upper extremity function and visual perception and, notably, using EEG data to measure real-time brain activity allowed for an objective assessment of the intervention's impact on the participant's attention and concentration, a significant strength of this study.

This report investigated the case of a subject with spastic CP and the effects of VR-based Leap Motion UE training on UE function, visual perception, and brain activity. The results showed improvement in the subject’s UE function and visual perception. Alpha wave band power decreased, and Beta wave band power increased after the paralyzed arm's VR Leap Motion UE training.

This study demonstrated that VR-based UE training for children with CP could elicit positive changes in UE function and visual perception. It also caused changes in brain activity. These novel findings suggest that VR-based UE training can be used to improve UE function and visual perception for patients with CP. However, as this was a case study, there are limitations to generalizing the results. Future research requires an objective and scientific design, and investigations with large sample sizes remain warranted.