Intelligent Wearable for Dysarthria Recovery Post-Stroke (www.nature.com)

A groundbreaking advance in assistive technology has emerged with the introduction of an AI-driven intelligent throat (IT) system designed to aid in the recovery of dysarthria for stroke patients. This wearable device leverages ultrasensitive textile strain sensors to monitor throat muscle vibrations and carotid pulse signals, enabling real-time continuous speech decoding through integration with large language models (LLMs). Unlike traditional systems that operate on fixed time slots, the IT system employs token-level processing, allowing seamless communication without the delays typical of existing technologies. In tests with five dysarthric patients, it achieved impressive accuracy metrics, including a 4.2% word error rate, a 2.9% sentence error rate, and a 55% increase in user satisfaction. The significance of this development lies in its potential to transform communication for individuals with neurological disorders by providing a portable, intuitive solution that accommodates emotional expression and contextual nuances. The system intelligently expands user-generated silent speech into coherent, personalized sentences, effectively restoring both efficiency and naturalness in communication. Furthermore, the technical architecture—featuring continuous strain monitoring and advanced machine learning techniques—establishes a promising foundation for future applications, potentially extending beyond dysarthria to a broader range of neurological conditions and languages.