Measurement and Applications of Speech-evoked Auditory Brainstem Responses (sABRs)

当声音到达耳朵时，微小的电脉冲通过大脑中不同的听觉中心传播，产生听觉感觉。虽然这个过程的很多方面仍然是个谜，但通过在头皮上贴上表面电极，我们可以非侵入性地记录这些听觉中心的电活动。特别是，记录听觉脑干反应是可能的，这是在大脑深处产生的。这些反应使临床医生能够诊断听力问题，并确定其起源于大脑。临床医生在这项测试中最常用的声音是短暂的响亮的滴答声，其好处是它是一种简单的刺激，具有明确的诱发反应。然而，点击并不能提供足够的信息来评估我们听到语音的能力。因此，人们对开发记录大脑对语音的电反应的测试越来越感兴趣。最近，我们的团队记录了自然和合成语音刺激下的语音诱发听性脑干反应(SABR)，我们已经证明这些反应可以为了解正常和听力受损个体的语音听觉处理提供一个重要的窗口。例如，我们已经证明SABR与元音音高的细微变化密切相关。事实上，这种大脑信号与语音波形非常相似，如果将其转换为声音并回放，它就会被感知为可理解的语音。这意味着它让我们能够“监听”大脑中语言的内部表示。 在这一领域取得进展的一个障碍是往往需要过多的记录时间。提出的研究计划旨在开发更有效的方法来检测SABR，这些方法的灵感来自于为噪声通信环境开发的语音增强算法。此外，由于SABR是在有限的一组语音刺激下录制的，因此将进行实验和理论工作，以澄清在不同的语音样本和不同的声学环境下响应的基本属性。SABR的测量还将用于确定助听器对语音信号的内部表示的影响。这有望为改进助听器配件提供一个强大的客观工具，这对于优化助听器的性能是很重要的。 拟议的研究计划预计将在生物医学工程和听力学领域产生重大影响。将开发评估和治疗听力障碍的新技术--这是加拿大的一个重要健康问题--包括改进的人类语音听觉处理模型、语音诱发反应的新信号处理算法，以及测量这些反应并将其应用于助听器配戴的仪器。此外，博士生和硕士研究生将接受跨越工程学和听力学的跨学科研究培训，并将培养对加拿大生物医学和助听器行业、医疗保健以及听力和语音研究社区有价值的技能。

When sound reaches the ear, tiny electrical pulses propagate through the different auditory centers in the brain, resulting in the sensation of hearing. Although much about this process remains a mystery, by taping surface electrodes to the scalp we can non-invasively record the electrical activity in these auditory centers. In particular, it is possible to record auditory brainstem responses, which are generated deep within the brain. These responses allow clinicians to diagnose hearing problems and locate their origin in the brain. The sound that clinicians most commonly use for this test is a brief loud click, which has the benefit of being a simple stimulus with a well-defined evoked response. However, a click does not provide sufficient information to allow us to assess the ability to hear speech. As a result, there is increasing interest in developing tests that record the brain's electrical response to speech. Recently, our group has recorded Speech-evoked Auditory Brainstem Responses (sABRs) with natural and synthetic speech stimuli, and we have shown that these responses can provide an important window into auditory processing of speech in normal and hearing impaired individuals. For example, we have shown that the sABR closely follows the fine variations in the pitch of a vowel. In fact, this brain signal is so similar to the speech waveform, if it is converted to a sound and played back, it is perceived as intelligible speech. This means that it allows us to "listen in" on the internal representation of speech within the brain. One impediment to progress in this field has been that excessive recording times are often required. The proposed research program aims to develop more efficient methods to detect the sABR that are inspired by speech enhancement algorithms developed for noisy communication environments. Moreover, because the sABR has been recorded in response to a limited set of speech stimuli, experimental and theoretical work will be performed to clarify the basic properties of the responses, with a diverse set of speech samples and in different acoustic environments. The measurement of sABRs will also be used to determine the impact of hearing aids on the internal representation of the speech signal. This is expected to provide a powerful objective tool for improving hearing aid fitting, which is important for optimising hearing aid performance. The proposed research program is expected to have a significant impact in the fields of Biomedical Engineering and Audiology. Novel techniques for the assessment and treatment of hearing impairment - an important health problem in Canada - will be developed, including improved models of human auditory processing of speech, new signal processing algorithms for speech-evoked responses, and instrumentation for measuring these responses and applying them for hearing aid fitting. Moreover, Doctoral and Master’s students will be trained in interdisciplinary research spanning engineering and audiology, and will develop skills of value to the Canadian biomedical and hearing assistive devices industries, healthcare, and the hearing and speech research communities.