Comparison of action observation therapy and virtual reality combined with conventional therapy for upper extremity rehabilitation in children with unilateral cerebral palsy

Highlights

• •AOT and Xbox added to physiotherapy improve upper limb in cerebral palsy.
• •AOT and XBOX boost motor skills and daily independence when added to physiotherapy.
• •AOT is more effective for distal joints, selectivity; XBOX proximal and gross skills.
• •These therapies in rehabilitation improve kids' quality and social participation.

Abstract

Background

This study aims to evaluate and compare the effects of two neurorehabilitation methods—action observation therapy (AOT) and XBOX 360 Kinect™ (XBOX) on upper extremity body structure functions, activity, and functional participation levels in children with Unilateral Cerebral Palsy (UCP).

Methods

Forty-five children with UCP (aged 5–15 years) were randomly divided into three groups: AOT, XBOX, and Conventional (CNV). All groups received conventional physiotherapy (45 minutes, twice weekly for 3 weeks). The AOT group had additional 30-minute daily telerehabilitation sessions, while the XBOX group used XBOX 360 Kinect for the same duration. Evaluations included the Modified Ashworth Scale (MAS), Jamar Hand Dynamometer, Baseline Mechanical Grip Meter, Selective Control of the Upper Extremity Scale (SCUES), Jebsen Taylor Hand Function Test (JTHFT), Abilhand-Kids, and Pediatric Quality of Life Inventory (PedsQL). Statistical significance was set at p<0.017.

Results

Both AOT and XBOX groups showed greater improvements than CNV(p<0.017). AOT achieved clinically meaningful gains in all grasp types and selectivity, while XBOX mainly improved pinch, lateral grip, and forearm selectivity. Both groups improved in ABILHAND-Kids and PedsQL, with broader gains in AOT.

Conclusion

AOT and XBOX are effective supplementary therapies for upper extremity rehabilitation in children with UCP, complementing conventional physiotherapy.

Clinical trial number

NCT05327179.

Introduction

Cerebral Palsy is a non-progressive movement disorder caused by non-progressive brain damage during prenatal, perinatal, or postnatal periods. Approximately 30% of children with CP have unilateral CP (UCP), characterized by significant motor and sensory deficits on one side, particularly affecting the upper extremity.¹ In UCP, decreased joint mobility, muscle strength, increased spasticity, the presence of mirror movements and decreasing selectivity significantly impair upper limb functioning.²

Since the 2000s, approaches based on motor learning principles have increasingly been used in rehabilitation.³ Motor learning is a neuroplasticity-based process that enables lasting behavioral changes through practice and experience.⁴ These approaches are based on intensive and repetitive practice, increasing task difficulty, active participation, and feedback. Furthermore, programs tailored to the individual's needs and focused on goal-setting and functioning are important components that support motor learning.⁵

Action Observation Therapy (AOT) and virtual reality-based interventions (XBOX 360 Kinect™), are examples of approaches grounded in motor learning principles.³ AOT involves the observation and subsequent performance of simple actions commonly used in daily life activities. Studies have reported similar activation patterns in the sensorimotor cortex when an activity is observed or performed.⁶ AOT aims to facilitate the reinforcement of preserved neural circuits and the reactivation or reorganization of impaired networks through observational learning mechanisms.⁷ The XBOX-based interventions, on the other hand, represent a virtual reality approach where children perform tasks against the computer through games. This training engages movement practice with repetition, motivation, and immediate feedback, supporting motor learning with a 'learning-by-doing' approach,⁸ while avatar-based feedback reinforces ''learning-by-imitation' by increasing input to the primary motor cortex.⁹ Considering all of this, the main goal of both treatment approaches is to promote neuroplastic changes, facilitate motor learning, address functional impairments, and ultimately enhance activity and participation.

Improving upper extremity functioning is critical for independence in activities of daily living such as dressing, eating, and playing.¹⁰ Motor learning-based approaches support this goal. The literature has reported that AOT and XBOX-based interventions increase upper extremity functioning and participation in children with UCP.^11,12 In the literature, AOT and XBOX-based interventions are generally examined separately, and randomized controlled trials directly comparing these two methods, particularly in children with UCP, are lacking. This makes it difficult to determine which approach is more effective and creates a significant research gap in the literature. The fact that they share similar foundations but are effective with different therapeutic approaches necessitates determining which approach is superior in terms of functional gain and independence. Such comparisons should focus not only on motor performance but also on the individual's activities of daily living and level of participation.

At this point, the International Classification of Functioning, Disability and Health (ICF) framework provides a useful guide for holistically addressing the multidimensional nature of functioning. The ICF guides assessments encompassing not only motor performance but also different dimensions of functioning, such as activity and participation.¹³ This allows for a clearer assessment of the extent to which intervention methods provide meaningful and contextual experiences for children and their impact on learning and rehabilitation.¹⁴

This study aims to compare the effects of two innovative approaches based on motor learning principles, AOT and XBOX-based interventions, when integrated with conventional physiotherapy on upper extremity functioning, activity, and participation, guided by the ICF. This study aims to address the uncertainties surrounding the clinical effectiveness of contemporary neurorehabilitation approaches in children with UCP, fill a significant research gap in the literature, and contribute to the evidence base for innovative neurorehabilitation practices in pediatric populations.

Materials and methods

Study design

This is a parallel group, randomized controlled, single-blind study, in which the outcome assessor was blinded to group allocation, and it was approved by the Gazi University Ethics Committee (Date: December 28, 2020, #E-91610558.302.08.01-). Written informed consent was obtained from all parents of the participants. The study is registered with ClinicalTrials.gov (NCT05327179). Participants were randomly assigned to three groups: Conventional (CNV) (n=15), AOT (n=15), and XBOX (n=15) using Research Randomizer (https://www.randomizer.org/).

Participants

The study included children aged 5 to 15 years with a diagnosis of UCP confirmed by a pediatric neurologist, who were able to cooperate sufficiently to follow verbal instructions, comply with test procedures, and maintain effective communication throughout the intervention. Additionally, participants were required to be able to actively grasp and stabilize an object with the affected hand (MACS (Manual Ability Classification System) ≤ IV and HHFCS (House Hand Functional Classification System) ≥ 4).^15,16 Although individuals with MACS level IV were eligible, only children who could actively grasp and stabilize objects with the affected hand (HHFCS ≥ 4) were included in the study. Those who used the affected hand only for passive assistance were excluded. Participants using antiepileptic medication, having received Botulinum Toxin Type A (BoNT-A) injections or undergone orthopedic surgery within the past 6 months, with vision or hearing impairments affecting cooperation, or with bilateral lower extremity involvement were excluded from the study. Out of screened 56 children with UCP, 45 met the inclusion criteria. Fig. 1 presents the CONSORT flow chart.

Fig. 1.

Flowchart (UCP: Unilateral Cerebral Palsy, n: number of participants, AOT: Action Observation Therapy, XBOX: XBOX 360 Kinect™, CNV: Conventional).

Interventions

All participants received conventional physiotherapy. The intervention groups additionally received either AOT or XBOX-based training, whereas the conventional group underwent only conventional physiotherapy without any supplementary interventions. All sessions were conducted under the supervision of a physical therapist. Data were collected in different settings according to the type of intervention: AOT sessions were delivered remotely via telerehabilitation, while XBOX-based applications were conducted at the Developmental Physical therapy Pediatric Rehabilitation Unit of XXX University. Conventional physical therapy sessions were delivered in special education centers where the children routinely attended and where collaborating physical therapists worked.

CNV Protocol. Involved exclusively conventional physical therapy targeting the upper extremities. This therapy was administered twice a week, with each session lasting 45 minutes, over a period of three weeks. As part of upper extremity rehabilitation, various play-based and task-oriented activities were utilized to promote functional use and active engagement. Exercises that would facilitate motor learning were selected. These included weight-bearing, pushing and pulling tasks, dribbling while maintaining balance, object manipulation, bilateral hand use, dough kneading, goal-directed reaching, ball throwing into a target, adhesive placement at varying heights, water pouring, card turning, puzzle assembly, grasp-and-release exercises, clothespin opening and closing, ball throwing and catching, bead stringing, as well as paper folding and cutting.

AOT Protocol. AOT was administered via telerehabilitation in conjunction with conventional physical therapy over a period of 3 weeks. The protocol entailed the division of functional tasks into smaller, sequential steps to support motor learning. Each activity was structured into 3 or 4 clearly defined subcomponents of the activity, with videos created to focus on one subcomponent at a time for a duration of 3 minutes. While videos shot from various perspectives are used in the literature, egocentric videos were preferred in this study.¹⁷ This approach aimed to provide more visual cues, prevent mental rotation and effective motor imagery. In the practical implementation, two pre-existing versions of varying difficulty were used for children categorized under HHFCS levels 4–5 and 6–8. In the practical implementation, HHFCS levels 4–5 represent children with significant impairment in upper extremity function; HHFCS levels 6–8 represent children with better hand use and fewer limitations. The activities were selected from daily life activities, and the three subcomponents of a sample activity are as follows: “Putting sugar into the glass”, “Filling the glass with water”, “Bringing the glass to the mouth”. Although functional goals within these two protocols remained identical, the subcomponents of the activity within HHFCS 4–5 were facilitated by the materials.¹⁸ (Supplementary Material 1) Children with UCP were asked to first engage in a 3-minute passive observation phase for each subcomponent of a task. They were instructed to focus entirely on the visual and motor aspects of the movement in order to support motor imagery. This was followed by a 3-minute practice period in which they performed the observed movement using the same materials shown in the video. The process continued sequentially for each subcomponent of the activity. All activities were designed to include movements across the entire upper extremity (Supplementary Material 2). During both phases, the child sat with his or her feet flat on the floor and arms on the table. During the observation phase, videos were shown on a screen at least 22 inches wide, 1 meter away (Fig. 2).

Fig. 2.

Some images from Action Observation Therapy and Virtual Reality (a) Observation phase of Action Observation Therapy. (b) Action phase of Action Observation Therapy. (c) XBOX Kinect 360 training.

XBOX Protocol included XBOX 360 Kinect application, alongside conventional physical therapy, supervised by a physical therapist for 3 weeks (Fig. 2). Games, aligned with literature on upper extremity rehabilitation for UCP children,^19,20 focused on shoulder flexion, elbow extension, forearm supination, wrist extension and flexion, and grip release movements. These games were 'Kinect Adventure, Kinect Sport, Kinect Disneyland Adventures, and Kinect Forza Motorsport'. Each day, six different games were played, with each game lasting 5 minutes. The games offered three difficulty levels, adjusted based on the child's functional ability (Supplementary Material 1).

Outcome measures

Assessment of body functions and structures

In our study, muscle tone was evaluated using the Modified Ashworth Scale, gross grip strength was measured using the Jamar Hand Dynamometer, fine grip strength was assessed using the Baseline Mechanical Pinchmeter, and selectivity was evaluated utilizing the Selective Control of the Upper Extremity Scale (SCUES).²¹

Assessment of activity and participation

Hand functions were evaluated using the Jebsen Taylor Hand Function Test (JTHFT),²² daily life activities were assessed with the ABILHAND-Kids scale ²³ and Participation was measured using the Pediatric Quality of Life Inventory (PedsQL).²⁴ Although the PedsQL primarily measures quality of life, it was used in this study to obtain indirect information related to children’s participation through its home, school and social functioning subscales. The primary outcome of this study was determined to be the ABILHAND-Kids scale, and the power analysis was conducted considering these results.

Statistical analysis

Power analysis conducted before the study assumed a hypothetical η² level of 0.6. With a minimum of 15 participants in each group, the study's power was calculated to be 85% with a margin of error of 0.017. The Post Hoc analysis, based on 15 participants in each group, revealed a large effect size (η²=0.59) with a margin of error of 0.017.

Repeated measurements within the group were analyzed using the Wilcoxon test and Paired t-test, and significance was determined as p<0.05. More than two independent groups were compared using One-Way Analysis of Variance (ANOVA) and the Kruskal-Wallis tests. Bonferroni correction, Games Howell tests, and Pairwise Comparisons Post-Hoc analysis were used in pairwise comparisons based on multiple comparison results. Results were considered statistically significant at p<0.017.

The formula r =Z$\sqrt{n}$ was used for the effect size of the non-normally distributed data (r = 0.1, small effect; r = 0.3, medium effect; r = 0.5, large effect). The formula d=$\overline{X1}-\overline{X2}/\text{SDpooled}$ (Cohen’s d) was used for the effect size of normally distributed data. ANOVA was used to compare the three independent groups in terms of numerical data.

Intraclass correlation coefficient (ICC) were obtained from the literature and used to calculate the standard error of measurement (SEM) and the Smallest Detectable Difference (SDD) in order to interpret the clinical significance of change following the intervention. Calculations were based on standard deviation (SD) from our sample and applied using the following formulas: SEM = SD × $\sqrt{(1-ICC}$) and SDD = SEM × 1.96 × $\sqrt{2}$, corresponding to a 95% confidence level. This approach allowed us to distinguish real change attributable to the intervention from potential measurement error.²⁵

Results

No adverse events were reported by the children or their parents throughout the treatment program applied to 45 children with UCP. The demographic characteristics of children are presented in Table 1. All groups were comparable at baseline across all outcome measures (p>0.017).

Table 1.

Demographic characteristics of participants.

		AOT n=15 (%)	XBOX n:15 (%)	CNV n=15 (%)	TOTAL n:45 (%)	p
Gender	Female	7 (%46.7)	8 (%53.3)	9 (%60)	24 (%53.3)	0.929
	Male	8 (%53.3)	7 (%46.7)	6 (%40)	21 (%46.7)
Affected side	Right	8 (%53.3)	9 (%60)	7 (%46.7)	24 (%53.3)	0.765
	Left	7 (%46.7)	6 (%40)	8 (%53.3)	21 (%46.7)
MACS	I	0 (%0)	1 (%6.7)	0 (%0)	1 (%2.2)	0.613
	II	7 (%46.7)	7 (%46.7)	9 (%60)	23 (%51.1)
	III	7 (%46.7)	5 (% 33.3)	5 (%33.3)	17 (%37.8)
	IV	1 (%6.7)	2 (%13.3)	1 (%6.7)	4 (%8.9)
HHFCS	IV	1 (%6.7)	2 (13.3%)	1 (%6.7)	4 (%8.9)	0.803
	V	7 (%46.7)	4 (%26.7)	5 (%33.3)	16 (%35.6)
	VI	7(%46.7)	7 (%46.7)	7 (%46.7)	21 (%46.7)
	VII	0 (%0)	2 (%13.3)	2 (%13.3)	4 (%8.9)
	VIII	0 (%0)	0 (%0)	0 (%0)	0 (%0)
Age (years). mean (SD)		10.27 (3.04)	9.33(3.13)	10.20(3.65)	9.33(3.24)	0.688
Height (m). mean(SD)		1.34 (0.20)	1.31(0.20)	1.33(0.22)	1.32(0.20)	0.891
Weight (kg). mean(SD)		36.6(14.03)	31.66(10.70)	32.10(13.40)	33.45(12.69)	0.509
Body Mass Index (SD)		18.00 (4.71)	18.38 (2.87)	17.50 (3.59)	17.92 (4.08)	0.291

MACS: Manual Ability Classification System, HHFCS: House Hand Functional Classification System, AOT: AOT Group, XBOX: XBOX Group, CNV: Conventional Group, n: number of participants, m: Meter, kg: Kilogram, p: Significance Level (<0.017).

Compared to the CNV group, both AOT and XBOX groups showed greater improvements in muscle tone, grip strength, selectivity, and functional outcomes. AOT exceeded the SDD value in all grip types and selectivity (except shoulder) and demonstrated clinical significance. XBOX showed clinically significant gains, particularly in pinch, lateral grip and forearm selectivity. In functional tests, clinically significant improvement was observed in the AOT group only in the “Simulated feeding” subheading of the JTHFT. ABILHAND-Kids and PedsQL scores showed statistically significant increases in both groups; however, the AOT group exceeded the SDD in more subdomains, providing clinical superiority (Tables 2, 3).

Table 2.

Mean within-group and between-group differences in Body Functions and Structures.

	Groups	Baseline Median (IQR) / Mean ± SD Min - Max	3 weeks	*Effect size / p value	Change between baseline and 3 weeks Median (IQR) Mean±SD	SEM / SDD	**Effect size / p value	POSTHOC AOT-CNV	POSTHOC XBOX-CNV	POSTHOC AOT-XBOX
Modified Ashworth Scale (min-max: 0-5)
Elbow Flexion	AOT	1.00 (1.00) 0.00 - 3.00	1.00 (1.00)	0.71/ 0.006*	-1.000 (2.00) 0.86±0.83	0.38/1.05	0.33/ 0.001⁎⁎	0.001	0.016	1.000
	XBOX	1.00 (1.00) 0.00 - 3.00	1.00 (1.00)	0.73/ 0.005*	-1.000 (1.00) 0.53±0.51	0.31/0.85
	CNV	1.00 (1.00) 0.00 – 2.00	2.00 (1.00)	0.35/ 0.180	0.00 (1.00) 0.20±0.56	0.23/0.63
Pronators	AOT	2.00 (1.00) 0.00 – 3.00	1.00 (1.00)	0.81/ 0.002*	-1.000 (1.00) 1.13±0.74	0.44/1.21	0.48/ 0.000⁎⁎	<0.001	0.068	0.068
	XBOX	1.00 (1.00) 0.00 – 4.00	1.00 (1.00)	0.63/ 0.014*	0.000 (1.00) 0.40±0.50	0.53/1.46
	CNV	1.00 (1.00) 0.00 – 2.00	2.00 (1.00)	0.42/ 0.102	0.000 (1.00) 0.26±0.59	0.28/0.77
Wrist Flexors	AOT	1.00 (1.00) 0.00 – 3.00	0.00 (1.00)	0.85/ 0.001*	-1.000 (0.00) 0.93±0.59	0.44/1.22	0.27/ 0.010⁎⁎	<0.001	0.092	0.223
	XBOX	1.00 (2.00) 0.00 – 4.00	1.00 (1.00)	0.65/ 0.011*	0.000 (1.00) 0.60±0.82	0.55/1.52
	CNV	1.00 (1.00) 0.00 – 3.00	1.00 (0.00)	0.25/ 0.317	0.000 (0.00) 0.06±0.25	0.34/0.94
Gross and Fine Grip Strength (N/kg)
Gross Grip (N)	AOT	59.10 ± 52.30 23.53 – 119.93	88.30 (51.54)	0.88/ 0.001*	28.53 ± 35.59	6.26/17.35	0.46/ 0.000⁎⁎	<0.001	<0.001	0.330
	XBOX	57.20 ± 61.93 10.26 – 200.66	85.66 (66.93)	0.88/ 0.001*	13.27 ± 16.0	9.58/26.55
	CNV	57.20 ± 91.80 13.13 – 206.00	97.90 (97.80)	0.17/ 0.495	0.80 ± 5.40	10.92/30.26
Pinch Grip (kg)	AOT	3.66 ± 1.66 1.33 – 5.66	5.66 (1.84)	0.88/ 0.001*	1.67 ± 0.84	0.10/0.27	0.58/ 0.000⁎⁎	<0.001	0.002	0.057
	XBOX	3.33 ± 1.66 2.00 – 7.66	4.66 (1.66)	0.85/ 0.001*	0.67 ± 0.99	0.11/0.30
	CNV	3.00 ± 3.25 1.00 – 6.00	3.00 (3.00)	0.08/ 0.755	0.00 ± 0.50	0.14/0.38
Lateral Grip (kg)	AOT	5.86 ± 1.64 3.33 – 8.66	8.02 ( 1.63	0.88/ <0.001*	2.00 ± 1.67	0.07/0.19	0.57/ 0.000⁎⁎	0.000	0.012	0.049
	XBOX	6.32 ± 2.26 1.33 – 11.33	7.34 ( 2.52	0.78/ <0.001*	1.00 ± 1.33	0.10/0.27
	CNV	4.98 ± 2.75 1.00 – 11.5	5.03 ( 2.69	0.19/ 0.458	0.00 ± 0.25	0.12/0.33
Tripod Grip (Affected Side) (kg)	AOT	4.00 ± 2.33 1.00 – 6.66	6.66 (2.34)	0.88/ 0.001*	2.34 ± 2.66	0.28/0.77	0.55/ 0.000⁎⁎	<0.001	0.068	0.009
	XBOX	3.66 ± 3.00 1.00 – 9.33	4.66 (1.66)	0.75/ 0.004*	1.00 ± 1.34	0.38/1.05
	CNV	4.25 ± 3.00 1.00 – 11.25	4.00 (3.00)	0.25/ 0.319	0.00 ± 0.25	0.50/1.38
Tripod Grip (Unaffected Side) (kg)	AOT	9.33 ± 1.66 5.33 – 12.00	9.33 (3.34)	0.19/ 0.454	0.00 ± 1.01	0.37/1.02	0.10/ 0.007⁎⁎	0.008	0.070	0.071
	XBOX	7.66 ± 3.67 3.66 – 11.33	8.00 (4.33)	0.12/ 0.148	0.67 ± 0.66	0.44/1.21
	CNV	6.25 ± 4.00 5.33 – 12.00	6.33 (5.00)	0.12/ 0.636	0.00 ± 0.59	0.76/3.82
Selective Control of the Upper Extremity Scale
Shoulder	AOT	3.00 (1.00) 2.00 – 3.00	3.00 (0.00)	0.45/ 0.830	0.20 ± 0.41	0.18/0.49	0.18/ 0.021	-	-	-
	XBOX	3.00 (1.00) 2.00 – 3.00	3.00 (0.00)	0.63/ 0.014*	0.40 ± 0.51	0.19/0.52
	CNV	3.00 (1.00) 1.00 – 3.00	3.00 (1.00)	0.25/ 1.000	-0.13 ± 0.52	0.24/0.66
Elbow	AOT	2.00 (0.00) 1.00 – 3.00	3.00 (2.00)	0.86 /<0.001*	0.73 ± 0.46	0.23/0.63	0.49/ 0.001⁎⁎	0.002	0.262	0.016
	XBOX	2.00 (1.00) 1.00 – 3.00	3.00 (1.00)	0.45/ 0.014*	0.20 ± 0.41	0.21/0.58
	CNV	2.00 (1.00) 1.00 – 3.00	2.00 (1.00)	0.00/ 1.000	0.00 ± 0.53	0.20/0.55
Forearm	AOT	1.00 (1.00) 1.00 - 2.00	2.00 (1.00)	0.90/ <0.001*	0.93 ± 0.46	0.15/0.41	0.64/ 0.001⁎⁎	0.001	0.165	0.033
	XBOX	2.00 (0.00) 1.00 – 3.00	3.00 (1.00)	0.63/ 0.014*	0.40 ± 0.51	0.13/0.36
	CNV	2.00 (2.00) 1.00 – 3.00	2.00 (2.00)	0.25/ 1.000	0.13 ± 0.52	0.24/0.66
Wrist	AOT	2.00 (1.00) 1.00 – 2.00	3.00 (1.00)	0.79/ 0.002*	0.73 ± 0.59	0.10/0.27	0.51/ 0.009⁎⁎	0.010	0.366	0.051
	XBOX	2.00 (0.00) 0.00 – 3.00	2.00 (1.00)	0.25/ 0.317	0.33 ± 0.48	0.15/0.41
	CNV	2.00 (0.00) 1.00 – 2.00	2.00 (0.00)	0.44/ 1.000	0.20 ± 0.41	0.07/0.19
Fingers	AOT	2.00 (1.00) 1.00 – 3.00	3.00 (1.00)	0.83/ 0.001*	0.80 ± 0.56	0.16/0.44	0.45/ 0.002⁎⁎	0.001	0.135	0.084
	XBOX	2.00 (0.00) 1.00 - 3.00	2.00 (1.00)	0.42/ 0.102	0.27 ± 0.59	0.16/0.43
	CNV	2.00 (1.00) 1.00 – 3.00	2.00 (1.00)	0.11/ 1.000	-0.07 ± 0.59	0.18/0.49
TOTAL	AOT	10.00 (3.00) 6.00 – 12.00	13.00 (3.00)	0.87/ 0.001*	3.40 ± 1.40	0.32/0.88	0.61/ 0.000⁎⁎	<0.001	0.025	0.002
	XBOX	11.00 (2.00) 7.00 – 14.00	13.00 (2.00)	0.55/ 0.031*	1.60 ± 1.24	0.31/0.85
	CNV	10.00 (3.00) 7.00 – 13.00	10.00 (3.00)	0.14/ 1.000	0.13 ± 1.36	0.31/0.85

AOT: AOT Group, XBOX: XBOX Group, CNV: Conventional Group, N: Newton, kg: kilogram, IQR: Inter Quantile Range, SEM: Standard Error of Measurement SDD: Smallest Detectable Difference, SD: Standard Deviation

*: Shows the significance value of pre-post difference p˂0.05, **: Show the significance value of the delta score p˂0.017

Table 3.

Mean within-group and between-group differences in Activity and Participation.

	Groups	Baseline Median (IQR) / Mean ± SD Min - Max	3 weeks	*Effect size / p value	Change between baseline and 3 weeks Median (IQR) Mean± SD	SEM / SDD	**Effect size / p value	POSTHOC AOT-CNV	POSTHOC XBOX-CNV	POSTHOC AOT-XBOX
Jebsen Taylor Hand Function Test (sec)
Card turning	AOT	9.12 ± 5.30 6.18 – 23.45	6.41 ± 4.40	0.88/ 0.001*	-2.76 ± 1.74	1.53/4.24	0.60/ 0.000⁎⁎	<0.001	0.480	0.001
	XBOX	7.02 ± 6.55 5.08 – 17.17	7.15 ± 7.33	0.62/ 0.016*	-0.38 ± 1.91	1.15/3.18
	CNV	9.98 ± 10.31 6.77 – 20.57	9.09 ± 11.20	0.04/ 0.865	0.03 ± 0.82	1.40/3.88
Picking up small common objects	AOT	13.21 ± 12.28 9.15 – 26.15	10.98±13.69	0.62/ 0.017*	-1.39 ± 1.79	1.54/4.26	0.10/ 0.103	-	-	-
	XBOX	17.43 ± 19.01 7.17 – 42.75	16.92±20.57	0.56/ 0.031*	-1.13 ± 1.73	2.67/7.40
	CNV	17.55 ± 9.91 7.37 – 31.13	17.61±10.68	0.23/ 0.363	-0.41 ±0.95	1.90/5.26
Simulated feeding	AOT	42.01 ± 13.01 17.72 – 52.37	30.21±24.87	0.88/ 0.001*	-8.53 ± 9.73	2.93/8.12	0.31/ 0.000⁎⁎	<0.001	0.389	0.009
	XBOX	42.18 ± 23.31 18.45 – 58.26	29.13±23.48	0.59/ 0.023*	-2.15 ±8.83	3.54/9.81
	CNV	28.32 ± 29.05 11.48 – 52.52	27.84±34.33	0.28/ 0.281	-1.23 ± 2.14	3.81/10.56
Stacking checkers	AOT	8.65 ± 6.88 3.37 – 29.22	7.20±8.22	0.65/ 0.012*	-1.66 ±1.35	1.98/5.48	0.17/ 0.019⁎⁎	0.140	0.140	0.787
	XBOX	5.62 ± 7.04 2.70 – 14.77	4.76±7.49	0.53/ 0.038*	-0.95 ± 1.47	1.01/2.79
	CNV	7.42 ± 7.16 3.16 – 32.19	7.61±6.58	0.22/ 0.394	0.37 ± 1.51	1.92/5.32
Moving light objects	AOT	6.35 ± 2.60 4.47 – 31.82	4.75±3.61	0.79/ 0.002*	-1.74 ± 2.11	1.73/4.79	0.11/ 0.007⁎⁎	0.010	0.038	1.000
	XBOX	7.12 ± 6.34 3.72 – 29.29	5.23±5.62	0.60/ 0.020*	-1.56 ± 1.68	1.66/4.60
	CNV	7.82 ± 5.87 4.47 – 27.13	7.12±5.66	0.16/ 0.551	0.03 ± 1.40	1.90/5.26
Moving heavy objects	AOT	7.35 ± 3.74 4.59 – 16.56	6.23±2.23	0.54/ 0.036*	-0.97± 1.51	0.97/2.68	0.07/ 0.212	-	-	-
	XBOX	8.73 ± 7.53 4.12 – 30.12	6.54±6.05	0.64/ 0.013*	-0.92 ± 1.21	1.92/5.32
	CNV	9.66 ± 10.97 3.13 – 53.61	9.13±10.90	0.29/ 0.256	-0.26 ± 0.80	3.75/10.39
TOTAL	AOT	93.80 ± 31.79 50.68 – 170.42	75.56±33.34	0.56/ <0.001*	18.24 ± 7.29	8.26/22.89	0.48/ 0.000⁎⁎	<0.001	0.003	0.011
	XBOX	93.31 ± 36.99 47.56 – 173.23	83.55±39.62	0.25/ 0.003*	-8.34 ± 10.50	9.64/26.72
	CNV	95.78 ± 42.98 41.04 – 169.69	95.20±44.92	0.01/ 0.558	-2.23 ± 1.09	11.20/31.04
Abilhand-KIDS
AOT		0.95 ± 1.07 -1.96 – 1.96	3.14±1.30	1.84/ <0.001*	2.19 ± 0.26	0.15/0.41	0.47/ <0.001⁎⁎	<0.001	0.001	0.010
XBOX		1.96 ± 2.20 -1.02 – 6.68	2.91±1.72	0.48/ <0.001*	0.96 ± 0.78	0.31/0.85
CNV		2.01 ± 0.75 1.03 – 6.68	2.44±1.39	0.39/ 0.350	0.42 ± 1.70	1.95/5.40
Pediatric Quality of Life Inventory (Parent)
Physical health score	AOT	511.66 ±84.44 350.00 – 625.00	675.0±94.01	1.83/ <0.001*	163.33±112.94	41.36/114.64	0.48/ <0.001⁎⁎	0.011	0.020	0.018
	XBOX	465.00 ± 158.62 275.00 – 800.00	540.0±157.7	0.47/ 0.002*	75.00 ±76.76	77.70/215.37
	CNV	466.66 ± 117.51 225.00 – 775.00	458.3±121.9	0.07/ 0.290	-8.33 ±29.38	57.56/159.54
Psychosocial health score	AOT	334.44 ± 60.49 200.00 – 400.00	419.4±39.67	1.66/ <0.001*	85.00 ±51.97	23.42/64.91	0.41/ <0.001⁎⁎	0.004	0.001	0.245
	XBOX	330.56 ± 97.57 178.33 - 500.00	391.11±63.89	0.73/ 0.002*	69.56 ±60.46	37.78/104.72
	CNV	358.89 ± 78.71 158.33 – 466.67	354.44±77.3	0.06/ 0.217	-4.44 ±13.31	30.48/84.48
Scale total score	AOT	423.06 ± 49.52 254.17 – 495.83	547.22±60.2	2.25/ <0.001*	124.17 ±64.96	17.15/47.53	0.57/ <0.001⁎⁎	0.017	0.001	0.015
	XBOX	397.78 ± 122.57 216.67 – 650.00	465.6±100.63	0.60/ <0.001*	67.78 ±54.88	42.45/117.6
	CNV	412.78 ± 86.41 191.67 – 620.83	406.39±89.24	0.07/ 0.138	-6.38 ±15.74	29.95/83.01
Pediatric Quality of Life Inventory (Child)
Physical health score	AOT	531.66 ± 130.75 300.00 – 775. 00	681.66±121.39	1.89/ <0.001*	150.00±75.00	67.93/188.29	0.58 / <0.001⁎⁎	<0.001	0.003	0.005
	XBOX	531.66 ± 151.03 275.00 – 800.00	601.66±155.40	0.46/ 0.001*	70.00 ±75.00	78.47/217.50
	CNV	460.00 ± 100.36 225.00 – 675.00	451.66±97.04	0.08/ 0.173	0.00 ±25.00	52.14/114.52
Psychosocial health score	AOT	336.67 ± 63.52 225.00 – 433.33	415.56±38.94	1.50/ <0.001*	78.89 ±39.07	27.68/76.72	0.71 / <0.001⁎⁎	<0.001	0.003	0. 002
	XBOX	371.67 ± 85.84 233.33 – 500.00	512.22±80.60	1.69/ <0.001*	140.56 ±54.45	37.41/103.69
	CNV	356.11 ± 75.29 158.33 – 433.33	351.67±70.86	0.06/ 0.192	-4.44 ±12.54	32.81/90.94
Scale total score	AOT	429.17 ± 112.50 304.17 – 604.17	566.67±104.17	1.26/ 0.001*	108.33 ±58.33	36.22/100.39	0.73 / <0.001⁎⁎	<0.001	<0.001	0.371
	XBOX	425.00 ± 112.50 254.17 – 650.00	562.50±150	0.18/ 0.001*	95.00 ±58.33	47.87/132.68
	CNV	408.33 ± 62.50 191.67 – 550.00	408.33±50.00	0.00/ 0.065	0.00 ±12.50	32.89/91.16

AOT: AOT Group, XBOX: XBOX Group, CNV: Conventional Group, sec: second, IQR: Inter Quantile Range, SEM: Standard Error of Measurement, SDD: Smallest Detectable Difference, SD: Standard Deviation

*: Shows the significance value of pre-post difference p˂0.05, **: Show the significance value of the delta score p˂0.017

Discussion

In this study, both AOT and XBOX-based interventions, combined with conventional physical therapy, supported upper extremity rehabilitation in children with UCP. Significant improvements were observed in muscle tone, grip strength, selectivity, and functional outcomes across both intervention groups. Both intervention groups were found to be more effective than conventional treatment alone. Additionally, both the AOT and XBOX groups demonstrated clinically meaningful improvements, exceeding the SDD threshold, on the study's primary outcome measure, the ABILHAND-Kids scale. No overall superiority has been demonstrated between AOT and XBOX, but both approaches have provided distinct benefits in different areas.

When examined in terms of clinical significance, AOT showed improvements above the SDD threshold in most motor and functional outcomes, particularly in distal joints, fine motor skills, and daily participation. XBOX exceeded the SDD threshold in fewer domains, mainly in proximal joint control and psychosocial outcomes. No statistically or clinically significant improvements were detected in the conventional physiotherapy group, despite some observable changes.

No differences were observed in age, gender, and functional level among the children with UCP in our study suggesting that the identified characteristics do not influence the treatment, and the observed changes after treatment are treatment-related.

Given that neurological development is age-dependent, the age range of 5–15 is reported to be the most suitable period for synaptic development, neural activation, and attention control.^26,27 Therefore, children in this age group were included in the study.

In our study, the CNV group received of 4.5 hours of treatment, whereas the AOT and XBOX groups received 12 hours of therapy in total. Jackman et al. reported that motor learning typically requires 14–25 hours of intervention, although this duration may vary depending on individual factors, the complexity of the goal, and the type of intervention.²⁸ In a study by Jackman et al., it was demonstrated that a minimum of 10 hours of functional intervention may be sufficient to achieve effective goal-oriented training.²⁹ The 12 hours intervention in our study aligns with the recommended dose for effective goal-directed training. The literature recommends that daily session duration for VR applications be 30–45 minutes, with a total duration of at least 360 minutes.^30,31 Furthermore, it has been reported that daily session duration between 30 and 60 minutes in AOT did not significantly differ in terms of improvement in grip strength and upper extremity function.¹¹ Our study is consistent with the recommended session durations of 30 minutes per day in AOT and 450 minutes in total in VR, along with the minimum goal-directed training time specified in the literature. Additionally Acar et al. reported that handheld controllers reduce the number of repetitions in VR applications.³² Therefore, in accordance with the principle of ‘learning by doing’, controller-free XBOX was used to increase functioning and repetitions.

In addition to the duration and intensity of intervention, the application environment also influences treatment effectiveness. In our study, AOT was implemented via telerehabilitation under the supervision of a physiotherapist, increasing accessibility, reducing costs, and providing individualized feedback. This hybrid model, which has gained importance in pediatric neurorehabilitation, strengthened family participation. Furthermore, the literature emphasizes that AOT can become a routine method through telerehabilitation, supporting the innovative approach of our study.³³

In our study, the initial disorder level assessment, guided by ICF, focused on muscle tone. Spasticity in UCP has been reported to be negatively correlated with motor functions, muscle strength and selectivity, and to reduce upper extremity participation.^21,34 Although statistical significance was observed at the end of treatment in the AOT and XBOX groups, clinical significance was not observed in any group. The AOT group showed more significant improvement in distal joints associated with fine motor skills requiring greater control due to the inclusion of such activities. In contrast, the XBOX group showed greater improvement in muscles in proximal joints such as the elbow. The repetitive, task-focused functional movements in XBOX are thought to regulate muscle tone, a finding consistent with other VR studies.^20,35 The improvement in both groups was linked to the strengthening of agonist muscles and a decrease in tone in antagonist muscles. Additionally, it is thought that the regulation of muscle tone in our study is related to increased selectivity. The lack of clinical significance may be due to the limited sensitivity of the MAS, the heterogeneity of the participants' age, motor level, and spasticity distribution, and the relatively short duration of the intervention. Moreover, clinical significance does not always reflect all dimensions of therapeutic effectiveness; even small changes can be important for an individual's quality of life and functioning.

The second assessment focused on gross-fine grip strengths. Statistical and clinical significance in grip strength across intervention groups generally overlapped. Improvement in the treatment groups is thought to be related to proximal stabilization, forced use, and increased motor selectivity, which is consistent with studies reporting the contribution of VR to object manipulation^11,18 and AOT to grip strength improvement.^19,36 The increase in tripod grasp on the less affected side at AOT may be explained by general motor development or the indirect effect of bimanual use, and this finding is consistent with the results of Perry et al.³⁷ Improvements in grip strength were also reflected in ABILHAND scores, indicating that changes in body functions improved activity performance, particularly during bimanual tasks and object grasping. However, the failure to exceed the SDD threshold in the gross and tripod grips in the XBOX group may be attributable to limited haptic feedback and lack of resistance in controller-free -based applications. This may be related to the inability to achieve sufficient finger-level strength due to the emphasis on proximal control.

The study focused on selectivity as the third domain of disorder. While there are numerous studies exploring the impact of selectivity on disorder, activity, and participation, research on treatment is limited. AOT demonstrated clinically significant improvement in all subscales except the shoulder. This improvement is related to the selective and targeted use of muscle groups in fine motor tasks. Feltham et al.³⁸ reported that MNS activation-based treatment reduced eccentric contraction of the biceps brachii muscle in distal joints, consistent with the limited improvement in the shoulder and the superiority of AOT in the elbow. Clinical significance was observed only in the forearm with XBOX, which may be explained by the selected games supporting movements in this region. Bryanton et al. noted that the virtual reality application improved selective motor control,³⁹ while Acar et al. reported its ineffectiveness on selectivity.³² The lack of improvement in other joints may be related to inadequate sensory and resistive input due to controller-free. The statistical significance in both groups may be attributable to the visual and multisensory feedback increasing attention and motivation, enabling more controlled and selective organization of movements.

In our study, the initial activity level assessment, guided by ICF, focused on JTHFT. After treatment, statistical significance was observed in both intervention groups in all subheadings. The inclusion of activities of daily living in the AOT and their assessment in the JTHFT may explain the observed improvement. Improvement in XBOX may be related to the support provided by games that encourage quick reactions, coordination, and visual tracking, which support hand function. Literature supports our study by reporting that AOT improves hand functions in children with UCP40, 41, 42 and that virtual reality applications increase the quality and speed of function in upper extremity rehabilitation.^12,32,43 However, only the AOT group demonstrated a clinically meaningful change exceeding the SDD value in the 'Simulated feeding' subscale. This situation can be attributed to several factors: feeding is a relatively simple and frequently practiced activity in daily life, the AOT protocol includes tasks directly related to feeding. Clinical significance was not found in the AOT and XBOX groups for other JTHFT subtests. This may be related to the interventions' lack of task specificity matching the JTHFT demands and the inadequacy of the interventions to produce clinical-level gains in JTHFT components such as speed, accuracy, and serial manipulation. Furthermore, potential ceiling effects may have prevented small gains from reaching clinical thresholds.

Both interventions produced both statistically and clinically significant improvements in Abilhand-KIDS scores, but AOT was statistically more effective. The superiority of AOT has been attributed to its targeting of daily living tasks and bimanual skills. Improvement in XBOX may be attributed to grasp-release, repetitive movements, proximal control, and increased upper extremity awareness during play. Our study is also supported by existing studies in the literature.^32,44,45 The clinical significance of the results indicates that improvements also increase perceived ability in daily activities.

Improvements in PedsQL scores, particularly in the school and social functioning subscales, indicate improved quality of life and may reflect better participation in daily routines. However, because the PedsQL is not a direct measure of participation, these findings should be interpreted as indirect indicators of participation. Our study used child and family forms. The improvement in the treatment groups is linked to decreased hand function impairments and neglect, fostering more active use of the affected extremity in ADL. Higher parental psychosocial scores on the AOT were associated with the perception that the app positively impacted children socially and emotionally. Higher child scores on the XBOX were attributed to its fun nature and its departure from routine. When delta scores were examined, clinically meaningful improvements (exceeding the SDD) were observed across all subdomains of the AOT group, whereas in the XBOX group, this was limited to the psychosocial subdomain of the child form. Given that the PedsQL reflects aspects of participation in daily life, these findings suggest that the AOT may have led to more comprehensive gains in overall participation, while the XBOX appears to have positively influenced the psychosocial experiences of children.

A limitation of this study is that individual differences such as severity of motor impairment and attention problems were not fully incorporated into the analysis; which may have influenced engagement and outcomes. Another limitation is that AOT was administered via telerehabilitation, whereas other interventions were administered face-to-face; this difference may constitute a potential source of bias. Additionally, our study could not assess the long-term effects of AOT and XBOX. Follow-up periods should be included in future research to examine the lasting impact of these applications. The inclusion of Functional Magnetic Resonance Imaging is believed to enhance the study's quality through objective evaluation. Comparative studies examining different durations and intensities for both XBOX and AOT would contribute valuable insights to the literature.

Conclusion

In conclusion, the addition of AOT and XBOX to conventional physical therapy appears effective in upper extremity rehabilitation in children with UCP. AOT is more effective in distal joints, fine motor skills, selectivity, and functional participation, while XBOX is more effective in proximal control and psychosocial well-being. The findings suggest that integrating both interventions into conventional treatment can improve motor and functional outcomes and that different therapeutic goals require different approaches.

Funding statement

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Declaration of competing interest

The authors report there are no competing interests to declare.