Development of Voice Neural Prosthesis Technology

描述（由申请人提供）：大约一半接受全喉切除术治疗晚期喉癌的患者依赖手持式人工喉器（EL）进行通信。不幸的是，使用一只手来控制EL的需求在物理上是有限的，并且这些设备产生的语音听起来不像人类，降低了可理解性和响度，并吸引了用户不必要的注意力。本项目的主要目的是开发一种改进的人工电喉通信系统，使其更接近于正常的声音和言语产生。这将在喉切除术患者中实现，通过开发一个神经接口来提供对改进的EL声源的免提控制，即语音神经假体。我们将进行一项前瞻性的手术方案，以保留正常神经支配的带状肌肉和由喉返神经再支配的带状肌肉的发音相关信号。信号将使用表面肌电电极检测，并将由基于dsp的系统进行处理，该系统将控制线性、免提EL换能器，以产生更人性化的语音。我们将采用电生理和影像学方法对患者进行喉切除术后的系统随访，以更好地了解潜在的解剖和生理因素如何影响神经义声的发展。喉神经形态学将研究手术标本，以帮助解释恢复肌电图信号。将采用术后培训方案，以确保对新系统的最大熟练程度的发展，并正式评估培训在使用语音神经假体中的作用。主要目的是正式比较使用两种不同的神经支配源和传统的人工控制的语音通信。此外，正在进行的迭代开发工作也将尝试使用更先进的信号处理方法来进一步改善/优化系统性能接近个体患者正常语音产生的程度。从长远来看，一旦基本方法建立起来，我们计划将这项新技术的各个方面用于更广泛的喉切除术人群，并作为未来努力开发其他手术-修复/神经修复方法的基础，用于语音和语言康复。

DESCRIPTION (provided by applicant): About half of the patients who undergo total laryngectomy to treat advanced laryngeal cancer rely on a handheld artificial electrolarynx (EL) for communication. Unfortunately, the need to use one hand to control an EL is physically limiting, and these devices produce speech that sounds non-human, has reduced intelligibility and loudness, and draws undesirable attention to the user. The primary aim of this project is to develop an improved artificial electrolarynx communication system that more closely approximates normal voice and speech production. This will be accomplished in laryngectomy patients by developing a neural interface to provide hands-free control of an improved EL sound source, i.e., a voice neural prosthesis. We will carry out a prospective surgical protocol to preserve phonation-related signals from both normally innervated strap muscles and from strap muscles that have been reinnervated by the recurrent laryngeal nerve. The signals will be detected using surface EMG electrodes and will be processed by a DSP-based system that will control a linear, hands-free EL transducer to produce more human-sounding speech. Patients will be systematically followed post-laryngectomy using electrophysiological and imaging approaches to better understand how underlying anatomical and physiological factors influence the development of neural prosthetic voice. Laryngeal nerve morphology will be studied in surgical specimens to help interpret recovered EMG signals. A post-surgical training protocol will be employed to ensure development of maximum proficiency with the new system, and to formally assess the role of training in the use of a voice neural prosthesis. A main goal is to formally compare speech communication using the two different innervation sources and conventional manual control. In addition, an ongoing iterative development effort will also attempt to use more advanced signal processing approaches to further improve/optimize the degree to which system performance approximates normal speech production in individual patients. In the long-term, once the fundamental approaches have been established, we plan to utilize aspects of this new technology for a broader laryngectomy population, and as a foundation for future efforts to develop additional surgical-prosthetic/neuroprosthetic approaches to voice and speech rehabilitation.