Assessment and treatment of visuospatial neglect using active learning with Gaussian processes regression

Area of Science:

• Neuroscience
• Artificial Intelligence
• Rehabilitation Medicine

Background:

• Visuospatial neglect is a neurological disorder impacting spatial awareness, often following stroke, significantly affecting daily life.
• Current assessment methods for visuospatial neglect are limited, often paper-based, and fail to capture real-world complexity.
• Existing treatment options for visuospatial neglect are sparse, with modest reported improvements.

Purpose of the Study:

• To introduce an AI-driven solution for accurate 3D assessment of visuospatial neglect.
• To utilize active learning with Gaussian process regression for efficient patient assessment.
• To explore the application of this AI model in personalized treatment, gamification, and tele-rehabilitation.

Main Methods:

• Development of an AI-based virtual reality (VR) assessment module for visuospatial neglect.
• Implementation of an active learning method using Gaussian process regression to optimize assessment efficiency.
• Clinical validation through real-world trials comparing AI assessment with conventional tests.

Main Results:

• The AI-based VR assessment demonstrated higher sensitivity compared to conventional visuospatial neglect tests.
• High intra-rater reliability was maintained in the AI-driven assessment module.
• The study confirmed the accuracy and reliability of the AI model for diagnosing and monitoring visuospatial neglect.

Conclusions:

• The AI-powered VR tool offers a more sensitive and reliable method for assessing visuospatial neglect.
• This technology facilitates personalized healthcare, gamification, and tele-rehabilitation, potentially improving patient engagement and outcomes.
• The developed module represents a significant advancement in the diagnosis and management of visuospatial neglect.