Cerebral cortical activation and functional connectivity characteristics of children with unilateral cerebral palsy during a single-session virtual reality throwing motor task: a functional near-infrared spectroscopy study.

OBJECTIVE

This cross-sectional study used functional near-infrared spectroscopy (fNIRS) to explore cerebral cortical activation and functional connectivity in children with Unilateral Cerebral Palsy (UCP) at resting state, during traditional non-virtual reality (non-VR) and single-session VR throwing motor tasks, aiming to clarify the differences in neural responses between VR and non-VR tasks, and provide preliminary neuroimaging evidence for VR application in pediatric cerebral palsy rehabilitation.

METHODS

A total of 35 children with UCP (18 males, 17 females, mean age 9.67 ± 2.4 years; 16 with left hemiplegia, 19 with right hemiplegia; 3 graded MACS Level Ⅰ, 13 Level Ⅱ, 19 Level Ⅲ) were enrolled, with the affected hemisphere standardized to the left. All participants completed fNIRS data acquisition in a randomized order under three conditions: 6-minute resting state (seated quietly with eyes closed), a single-session semi-immersive VR throwing task (8 s per movement cycle, with real-time audio-visual reward feedback for successful hits on horizontally moving virtual targets via ToF 3D motion capture), and a traditional non-VR throwing task with matched movement rhythm but no targets or feedback. fNIRS was used to monitor oxygenated hemoglobin (HbO₂) concentration in 6 regions of interest (ROIs): bilateral prefrontal (LPFC/RPFC), premotor (LPMC/RPMC) and sensorimotor cortices (LSMC/RSMC). One-way repeated measures ANOVA with FDR correction was applied for statistical analysis, which was verified by a professional biostatistician.

RESULTS

Cerebral functional connectivity strength analysis showed that after FDR correction, there were significant main effects of task condition on the FCS of LSMC-LPMC [F(2,68) = 16.956, p < 0.001, ηₚ2 = 0.333], LPMC-RPFC [F(2,68) = 16.256, p < 0.001, ηₚ2 = 0.323], and LPMC-LPFC [F(2,68) = 22.725, p < 0.001, ηₚ2 = 0.401], all with large effect sizes (ηₚ2 ≥ 0.14). The FCS of these ROI pairs showed a consistent trend: resting state > single-session VR throwing motor task > non-VR throwing motor task, with significant differences between any two conditions (PFDR < 0.05). No statistically significant differences were observed in FCS between other ROIs (PFDR > 0.05). Cerebral activation level analysis showed that after FDR correction, there were significant main effects of task condition on the activation levels of the affected-side LSMC [F(2,68) = 15.952, p < 0.001, ηₚ2 = 0.319] and affected-side LPFC [F(2,68) = 13.606, p < 0.001, ηₚ2 = 0.286], both with large effect sizes (ηₚ2 ≥ 0.14). Specifically, the activation level of LSMC followed the order: single-session VR throwing motor task > non-VR throwing motor task > resting state; the activation level of LPFC followed the order: single-session VR throwing motor task > resting state > non-VR throwing motor task, with significant differences between any two conditions (PFDR < 0.05). No statistically significant differences were detected in activation levels of other brain regions (PFDR > 0.05).

DISCUSSION: Based on functional near-infrared spectroscopy (fNIRS) technology, this cross-sectional study confirmed that children with unilateral cerebral palsy (UCP) exhibit distinct cerebral cortical neural activity patterns during a single-session VR throwing motor task, compared with the resting state and traditional non-VR throwing motor task. Specifically, the single-session VR motor task was associated with significantly higher activation levels in the left sensorimotor cortex (LSMC) and left lateral prefrontal cortex (LPFC) - the two regions in the affected hemisphere uniformly standardized to the left for all subjects - and modulated the functional connectivity between the premotor cortex and its functionally related brain regions. Compared with the traditional non-VR throwing motor task, the single-session VR throwing motor task elicited greater activation in the above-mentioned motor-related and higher-order cognitive brain regions, and induced a significantly different neural response pattern of the motor-cognitive network. The findings of this study reveal the cortical neural response characteristics of children with UCP during a single-session VR motor task, and provide preliminary neuroimaging evidence for subsequent studies exploring the neural mechanisms of VR-based rehabilitation in children with cerebral palsy.