Top-Down Control of Selective Amplification

选择性扩增的自上而下控制

• 批准号: 10200001
• 负责人: Gerald, Jr Kidd
• 金额: $ 65.43万
• 依托单位: BOSTON UNIVERSITY (CHARLES RIVER CAMPUS)
• 依托单位国家: 美国
• 财政年份: 2014
• 资助国家: 美国
• 起止时间: 2014-04-01 至 2024-06-30
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10200001
• 关键词: Acoustics; Affect; Algorithms; Aphasia; Attention; Attentional deficit; Attenuated; Auditory; Award; Bilateral; Brain; Cochlear Implants; Communication; Complex; Coupling; Detection; Disease; Environment; Eye; Goals; Head; Hearing; Hearing Aids; Human; Human Characteristics; Impairment; Individual; Individual Differences; Investigation; Lead; Location; Loudness; Masks; Methods; Noise; Participant; Performance; Persons; Physiological; Population; Process; Request for Applications; Sensorineural Hearing Loss; Series; Signal Transduction; Site; Sound Localization; Source; Speech; Speech Intelligibility; System; Testing; Time; Uncertainty; Vision; Visual; Visual attention; Voice; Work; base; design; expectation; experience; gaze; hearing impairment; improved; microphone; normal hearing; novel; novel strategies; preservation; selective attention; signal processing; social situation; sound; source localization

Human listeners depend on the sense of hearing to communicate effectively in everyday social situations. Although we rely heavily on our ability to selectively attend to a single voice in a noisy background, and to follow transitions in talkers during conversation, this task is quite complex and accomplishing it successfully depends on the integrity of processing at a number of physiological sites spanning the auditory periphery to the brain. It is well known that hearing loss may adversely affect a listener's ability to perceptually segregate one talker in the midst of other talkers and to understand that talker's spoken message (i.e., the "cocktail party problem"; see Middlebrooks et al., 2017, for a series of recent reviews). The most common remedy for sensorineural hearing loss (SNHL) is a hearing aid, or a pair of aids, that can boost sounds to audible levels while preserving comfortable loudness and may improve signal-to-noise ratio for certain classes of sounds via "noise reduction". However, even when listeners with SNHL wear hearing aids they often still experience extreme difficulty perceptually navigating the "auditory scene", severely limiting their ability to communicate effectively. One reason is that, from an acoustic perspective, the designation of a particular sound source as "target" versus "masker" is arbitrary because it depends on the current - and changeable - internal state of the observer. Thus, the distinction between a target talker to be attended and a masker talker to be ignored can only be made by the listener and may change from moment to moment. Although the amplification of sounds by hearing aids provides the best (often the only) option for improving communication for listeners with SNHL, current hearing aids inherently fail to solve the source selection problem because they amplify target and masker sounds indiscriminately without the ability to distinguish which source the listener has chosen as the target. Thus the challenge is to devise a hearing aid that focuses only on those sounds the listener chooses to attend and suppresses competing sounds, responding to the wishes of the listener immediately, accurately, and effectively. During the past award period, our work has demonstrated that acoustic beamforming implemented by a head worn microphone array can provide a significant advantage for listeners with SNHL in solving the cocktail party problem. Furthermore, we have found that steering the beam of amplification can be accomplished quickly and effectively by sensing eye gaze with an eye tracker and directing the acoustic look direction (ALD) of the beam accordingly. The present application requests support to continue work on this visually guided hearing aid (VGHA) and to further examine the scientific premise upon which it is based. The overall goals are to better understand how top-down control of selective amplification provides assistance to listeners with SNHL in typical social situations, to advance our understanding of auditory and auditory-visual selective attention, and to extend the potential benefits of the VGHA to new populations of listeners - users of bilateral cochlear implants and persons with aphasia - who typically experience great difficulty understanding speech in complex, multiple-talker communication situations.

人类听众依靠听觉在日常社交场合进行有效交流。 尽管我们在很大程度上依赖于我们在嘈杂的背景下选择性地关注单个声音并跟随的能力 谈话过程中谈话者的转换，这项任务相当复杂，能否成功完成取决于 跨越听觉外围到大脑的许多生理部位的处理完整性。它 众所周知，听力损失可能会对听者区分谈话者的能力产生不利影响。 在其他说话者中间并理解该说话者的口头信息（即“鸡尾酒会问题”；参见 Middlebrooks 等人，2017，最近的一系列评论）。感音神经性听力最常见的治疗方法 助听器（SNHL）是一种助听器或一对助听器，可以将声音增强到可听的水平，同时保留 舒适的响度，并可以通过“降噪”提高某些类别声音的信噪比。 然而，即使患有 SNHL 的听众佩戴助听器，他们通常仍然会遇到极大的困难 感知地导航“听觉场景”，严重限制了他们有效沟通的能力。原因之一 从声学角度来看，将特定声源指定为“目标”与“掩蔽器”是 任意的，因为它取决于观察者当前且可变的内部状态。因此，区别 要注意的目标说话者和要忽略的掩蔽说话者之间只能由听者进行，并且 可能会随时发生变化。尽管助听器放大声音提供了最好的效果 （通常是唯一的）改善 SNHL 听众沟通的选择，目前的助听器本质上是失败的 解决源选择问题，因为它们不加区别地放大目标声音和掩蔽声音，而无需 区分听众选择哪个源作为目标的能力。因此，挑战在于设计一个 助听器仅关注听者选择关注的声音并抑制竞争声音， 立即、准确、有效地回应听众的愿望。在过去的颁奖期间， 我们的工作表明，通过头戴式麦克风阵列实现的声波束形成可以 为 SNHL 的听众解决鸡尾酒会问题提供了显着的优势。此外，我们 发现通过感知眼睛可以快速有效地控制放大光束 用眼动仪注视并相应地引导光束的声学观察方向（ALD）。现在的 申请请求支持继续开发该视觉引导助听器 (VGHA) 并进一步检查 它所依据的科学前提。总体目标是更好地理解自上而下的控制 选择性放大为典型社交场合中患有 SNHL 的听众提供帮助，以促进我们的 了解听觉和听觉视觉选择性注意，并扩展听觉的潜在好处 VGHA 面向新的听众群体——双侧人工耳蜗植入者和失语症患者—— 在复杂的、多人交流的情况下，理解语音通常会遇到很大的困难。