Virtual reality-based postural balance training in autistic children: a pilot randomized controlled trial

This dataset was collected to investigate the effects of an immersive virtual reality-based training administered with Gait Real-time Analysis Interactive Lab (GRAIL) on postural balance of autistic children. A total of 20 autistic participants (6 -13 y/o) were enrolled in a 5-week RCT and received either balance training with the GRAIL system or no training.

Specifically, the dataset includes:

-Demographic data collected at baseline (Sheet "Demographic data @ baseline");

-Primary outcomes of the study 

1. Static balance assessment (Sheet "Static COP")

Center of pressure (CoP) displacement parameters collected with the GRAIL system in four different conditions (standing with eyes open – wide feet (EOWF); standing with eyes (actively) closed – wide feet (ECWF); standing with eyes open – narrowed feet (EONF); standing with eyes (actively) closed – narrowed feet (ECNF)).

The following metrics are reported at T0 (at baseline) and T1 (follow-up): the sway path length, the area of the 95% confidence ellipse (area CE), the mean velocity of the CoP both in the ML and AP directions (Vel ML, Vel AP), and the jerk metrics.

2. Movement Assessment Battery for Children-2 (Sheet "MABC-2") 

Manual Dexterity, Aiming & Catching, Balance and Total scores are reported both at T0 and T1 for each subject.

-Secondary outcomes

1.Parent-report Developmental Coordination Disorder Questionnaire ("Sheet DCDQ")

Motor control, Fine motor, General coordination, and the Total score are reported both at T0 and T1 for each subject.

2. Balance assessment overground (Sheet "COM")

Center of mass sway path length is reported for the four standing conditions both at T0 and T1 for each subject.

3. Reactive balance assessment (Sheet "Reactive CoP")

Reactive postural balance to controlled perturbations was measured by means of CoP range displacement during the active perturbation and during the following recovery time, the peak value of CoP, the time to peak (TTP) and the time of recovery (TOR). Features are reported at T0 and T1 for each subject.

4. Gait parameters (Sheet "Gait parameters (Mean values)" and "Gait parameters (std values)")

Gait spatio-temporal, kineamatic (ROM of ankle, knee and hip during flexion) and kinetic (power of ankle, knee and hip) and are reported at T0 and T1 for each subject.

The mean and standard deviation of every gait parameter were computed by acquiring several steps during the trial. The standard deviation of the collected features was assessed as a measure of gait pattern variability.

If you use this dataset please cite the following manucript: Falivene, A.; Scaccabarozzi, G.; Busti Ceccarelli, S.; Molteni, M.; Klingels, K.; Verbecque, E.; Storm, F.A.; Biffi, E.; Crippa, A. Virtual Reality-Based Postural Balance Training in Autistic Children: A Pilot Randomized Controlled Trial. J. Clin. Med. 2025, 14, 5616. https://doi.org/10.3390/jcm14165616

本数据集旨在探究基于沉浸式虚拟现实、借助步态实时分析交互实验室（Gait Real-time Analysis Interactive Lab, GRAIL）开展的训练对自闭症儿童姿势平衡的影响。共计20名自闭症受试者（年龄6~13岁）参与了为期5周的随机对照试验（Randomized Controlled Trial, RCT），并被分配至GRAIL系统平衡训练组或无训练对照组。

具体而言，本数据集包含以下内容：
- 基线人口统计学数据（工作表"Demographic data @ baseline"）；
- 研究主要结局指标
1. 静态平衡评估（工作表"Static COP"）：
借助GRAIL系统在四种不同站姿条件下采集的压力中心（Center of Pressure, CoP）位移参数，分别为：睁眼宽站（EOWF）、闭眼宽站（ECWF）、睁眼窄站（EONF）、闭眼窄站（ECNF）。
在T0（基线）与T1（随访）两个时间点报告以下指标：晃动路径长度、95%置信椭圆面积（Area CE）、压力中心在中外侧（ML）与前后侧（AP）方向的平均速度（Vel ML、Vel AP），以及jerk指标。
2. 儿童运动评估量表第二版（Movement Assessment Battery for Children-2, MABC-2）：
报告所有受试者在T0与T1时间点的手工灵活性、瞄准与接球能力、平衡得分及总分。
- 次要结局指标
1. 家长报告版发育性协调障碍问卷（Developmental Coordination Disorder Questionnaire, DCDQ）：
报告所有受试者在T0与T1时间点的运动控制、精细运动、整体协调性得分及总分。
2. 地面平衡评估（工作表"COM"）：
报告所有受试者在四种站姿条件下，于T0与T1时间点的质心（Center of Mass, COM）晃动路径长度。
3. 反应性平衡评估（工作表"Reactive CoP"）：
通过主动扰动阶段及后续恢复阶段的压力中心位移范围、压力中心峰值、达峰时间（TTP）及恢复时间（TOR），评估受试者对可控扰动的反应性姿势平衡能力。所有受试者在T0与T1时间点的相关特征均已收录。
4. 步态参数（工作表"Gait parameters (Mean values)"与"Gait parameters (std values)"）：
报告所有受试者在T0与T1时间点的步态时空参数、运动学参数（踝关节、膝关节及髋关节屈曲活动范围）与动力学参数（踝关节、膝关节及髋关节功率）。
本研究通过采集试验过程中的多个步态周期计算得到每个步态参数的均值与标准差，并以采集特征的标准差作为步态模式变异性的评估指标。

若使用本数据集，请引用以下论文：Falivene, A.; Scaccabarozzi, G.; Busti Ceccarelli, S.; Molteni, M.; Klingels, K.; Verbecque, E.; Storm, F.A.; Biffi, E.; Crippa, A. Virtual Reality-Based Postural Balance Training in Autistic Children: A Pilot Randomized Controlled Trial. J. Clin. Med. 2025, 14, 5616. https://doi.org/10.3390/jcm14165616