A Wireless micro-ECoG Prosthesis for Speech

用于语音的无线微型 ECoG 假肢

• 批准号: 10706320
• 负责人: Kenneth L Shepard
• 金额: $ 61.84万
• 依托单位: DUKE UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 2021
• 资助国家: 美国
• 起止时间: 2021-09-17 至 2026-08-31
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10706320
• 关键词: Algorithms; Amyotrophic Lateral Sclerosis; Area; Auditory; Auditory area; Brain; Clinical; Communication; Custom; Data; Development; Electrocorticogram; Electrodes; Electroencephalography; Environment; Eye Movements; Family; Friends; Generations; Health Personnel; Human; Implant; Infection prevention; Locked-In Syndrome; Machine Learning; Maps; Measures; Modeling; Motor Cortex; Muscular Dystrophies; Myopathy; Neuromuscular Diseases; Patients; Performance; Prosthesis; Quality of life; Reading; Research; Resolution; Sampling; Self-Help Devices; Signal Transduction; Speech; Techniques; Technology; Testing; Time; Training; Translating; Translations; Voice; Wireless Technology; Work; battery life; brain surgery; density; design; high dimensionality; improved; infection risk; integrated circuit; liquid crystal polymer; machine learning algorithm; meter; motor disorder; neural; neural prosthesis; neurotransmission; novel; power consumption; reading ability; reconstruction; signal processing; success; wireless

Project Summary / Abstract Patients who suffer from debilitating neuromuscular disorders (e.g. amyotrophic lateral sclerosis – ALS, Locked- in-Syndrome – LIS, and muscular dystrophies/myopathies) have difficulty or an inability to communicate through speech leading to a detrimental loss in quality of life. Current technology using eye movements and signals/spellers from electroencephalography (EEG) are slow and inconsistent. Neural prostheses offer an opportunity to produce fast and accurate communication for patients suffering from neuromuscular disorders, but success for regaining speech has been limited due to technological limitations: there is an inability to capture the high dimensionality of the brain and an inability to record in naturalistic conditions using fully implanted, wireless electrode arrays. To solve these challenges, we develop and optimize custom wireless micro- electrocorticographic (µECoG) arrays with over 1,000 channels to decode speech directly from the human brain. We will accomplish this by 1) Testing and optimizing the spatial resolution of µECoG to capture neural signals, 2) Fine-tune our machine learning algorithms to decode speech directly from the brain and 3) developing wireless technology to enable neural prosthetic usage in naturalistic settings. High-density, high channel-count neural interfaces will offer an unprecedented ability to decode speech from the human brain. This ability combined with wireless technology, will allow for a new generation of speech neural prostheses.

项目总结/摘要 患有衰弱性神经肌肉疾病（例如肌萎缩性侧索硬化症- ALS、闭锁性脊髓侧索硬化症- 综合征- LIS和肌营养不良症/肌病）难以或无法通过 言语导致生活质量的有害损失。目前的技术利用眼球运动和 来自脑电图（EEG）的信号/拼写是缓慢且不一致的。神经假体提供了一种 为患有神经肌肉疾病的患者提供快速准确的沟通机会， 但是由于技术的限制，恢复语言的成功是有限的： 大脑的高维度和无法在自然条件下使用完全植入的， 无线电极阵列。为了解决这些挑战，我们开发并优化了定制无线微型 具有超过1，000个通道的皮层电图（µECoG）阵列，可直接从人脑解码语音。 我们将通过1）测试和优化µECoG的空间分辨率来捕获神经信号， 2)微调我们的机器学习算法，直接从大脑解码语音，3）开发无线 技术使神经假体能够在自然环境中使用。高密度，高通道数神经 接口将提供前所未有的能力来解码人类大脑的语音。这种能力与 无线技术，将允许新一代的语音神经假体。