Frontoparietal Cortical Activation during VR-based Prism Adaptation in Stroke Patients with Unilateral Spatial Neglect: an fNIRS study

Technical Summary: Frontoparietal Cortical Activation during VR-based Prism Adaptation in Stroke Patients with Unilateral Spatial Neglect

Problem Statement
Unilateral spatial neglect (USN), a common non-motor impairment following right hemisphere stroke, significantly hinders functional recovery and social return. While Prism Adaptation (PA) is a well-established rehabilitation modality for USN, its clinical application is constrained by practical limitations: the requirement for physical space to mask hand trajectories, the necessity of therapist supervision, and the difficulty in quantifying adaptation effects. Virtual Reality-based Prism Adaptation (VRPA) has emerged as a potential alternative that selectively deviates the virtual hand trajectory within an immersive environment, inducing sensorimotor conflict through a mechanism distinct from the global visual field shift of conventional PA. However, while neuroimaging studies have characterized the cortical substrates of conventional PA in stroke patients, no study has yet examined how VRPA activates the brain in this specific clinical population. Understanding the neural mechanisms of VRPA is critical, as the structural difference in error signals (virtual hand vs. real hand mismatch vs. global visual shift) may engage different neuroplastic processes.

Methodology

• Participants: The study recruited six right-handed subacute stroke patients (mean age 61.17 years, mean time since onset 31.83 days) with right hemisphere damage and confirmed USN (Catherine Bergego Scale ≥ 7). One additional patient was excluded due to excessive fNIRS noise.
• Experimental Design: Participants underwent a single VRPA session consisting of four sequential phases:
  1. Pre-VRPA: Baseline pointing with aligned real and virtual hands.
  2. VRPA-10°: Virtual hand trajectory deviated 10° to the right.
  3. VRPA-20°: Virtual hand trajectory deviated 20° to the right.
  4. Post-VRPA: Return to baseline conditions (no deviation) to measure after-effects.
     Each adaptation phase involved five pointing blocks with 30-second breaks.
• VRPA System: An immersive system using an Oculus Rift DK2 HMD and a Leap Motion sensor. The system selectively shifts the virtual hand trajectory while keeping the real target and visual field undisturbed, creating a conflict between the seen virtual hand and the felt real hand.
• Neuroimaging: Functional Near-Infrared Spectroscopy (fNIRS) was used to continuously record oxygenated hemoglobin (HbO) concentrations. The setup utilized 21 channels covering bilateral Dorsolateral Prefrontal Cortex (DLPFC) and Parietal Association Cortex (PAC) regions.
• Data Analysis: Behavioral pointing errors were calculated. fNIRS data were preprocessed to remove artifacts and converted to HbO concentration changes. Statistical analysis employed the Friedman test followed by Wilcoxon signed-rank tests with Bonferroni correction to compare HbO changes across the four VRPA conditions.

Key Results

• Behavioral Adaptation: The study confirmed a characteristic adaptation pattern. Upon introducing the virtual prism, patients exhibited rightward pointing errors which progressively diminished within each VRPA phase. A leftward after-effect was observed at the onset of the Post-VRPA phase, indicating successful sensorimotor adaptation comparable to conventional PA.
• Cortical Activation Patterns:
  • Left DLPFC: HbO concentrations significantly decreased during VRPA-10°, VRPA-20°, and Post-VRPA relative to the Pre-VRPA baseline.
  • Right DLPFC: HbO concentrations significantly increased specifically during the VRPA-20° condition.
  • Bilateral PAC: Both left and right parietal association cortices showed significantly increased HbO concentrations during VRPA-20° and Post-VRPA compared to Pre-VRPA. Notably, no significant parietal activation was observed during the VRPA-10° condition.

Key Contributions

• First Neuroimaging Study of VRPA in USN: This work represents the first investigation to characterize cortical activation patterns during VRPA specifically in stroke patients with USN.
• Mechanistic Insight: It demonstrates that despite the fundamental difference in how sensorimotor conflict is generated (virtual hand deviation vs. global visual shift), VRPA engages compensatory frontoparietal circuits in the damaged brain that resemble those reported for conventional PA.
• Deviation Threshold: The findings suggest a deviation-dependent threshold for engaging compensatory circuits, as significant parietal activation only emerged at the 20° deviation level, not at 10°.

Significance and Claims
The authors claim that these findings provide preliminary neuroimaging evidence supporting the potential neuroplastic effects of VRPA. The observed pattern—bilateral parietal activation combined with right DLPFC recruitment and left DLPFC deactivation—suggests that VRPA may facilitate interhemispheric rebalancing and engage compensatory pathways similar to conventional PA. This supports the hypothesis that VRPA is a viable rehabilitative intervention for enhancing attentional recovery in stroke-induced spatial neglect.

The paper maintains a modest tone regarding its conclusions, explicitly acknowledging limitations such as the small sample size (n=6), the lack of long-term behavioral transfer measures to clinical neglect scales, and the inability of fNIRS to measure deep brain structures like the cerebellum. Consequently, the authors state that these results lay the groundwork for future randomized controlled trials with larger samples and multimodal neuroimaging approaches, rather than serving as definitive proof of clinical efficacy.