Dynamic neural coding of spectro-temporal sound features during free movement

自由运动时谱时声音特征的动态神经编码

• 批准号: 10656110
• 负责人: Stephen V David
• 金额: $ 22.52万
• 依托单位: OREGON HEALTH & SCIENCE UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 2023
• 资助国家: 美国
• 起止时间: 2023-05-01 至 2025-04-30
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10656110
• 关键词: Acoustics; Address; Affect; Algorithms; Anatomy; Animal Model; Animals; Architecture; Arousal; Attention; Auditory; Auditory Prosthesis; Auditory area; Auditory system; Behavior; Behavioral; Brain; Calibration; Chronic; Code; Compensation; Complex; Computational Technique; Computer Models; Cues; Data; Discrimination; Ear; Engineering; Environment; Equipment; Feedback; Ferrets; Head; Hearing; Hearing problem; Implant; Individual; Location; Loudness; Machine Learning; Measurement; Measures; Methods; Microelectrodes; Modeling; Motion; Motor; Motor Activity; Movement; Neurons; Patients; Population; Positioning Attribute; Posture; Process; Property; Research; Role; Rotation; Sensory; Shapes; Signal Transduction; Source; Specificity; Speech; Stimulus; Study models; System; Technology; Testing; Update; Visualization; Visualization software; Work; artificial neural network; auditory processing; computer framework; convolutional neural network; deaf; experimental study; hard of hearing; hearing impairment; improved; insight; nervous system disorder; neural; neurophysiology; normal hearing; receptive field; response; sensory input; signal processing; sound; tool

Project Summary Despite ongoing advances in auditory prostheses, patients with hearing loss often have difficulty understanding speech and other important sounds in noisy environments. This is due, in part, to degraded spatial and spectral sound information, which is leveraged by normal-hearing listeners to parse concurrent sounds in the real world. Current understanding of spatial processing is drawn primarily from studies in which the subject is head-fixed relative to the sound sources. Despite this dominant experimental paradigm, listeners in real-world conditions typically move through space while orienting their head to improve their ability to understand auditory signals. The existence of neural connections between the vestibular, motor, and auditory systems suggests that a listener's movement and body posture provide substantial input to the auditory system to facilitate hearing. A better understanding of how the healthy auditory system operates while moving through an acoustic environment will support new treatments for auditory disorders. The current study will investigate how information about a listener's motion and body/head posture influence sound processing in the auditory cortex. Historically, studies in free-moving subjects have been limited by the difficulty of precisely measuring auditory input during unconstrained movement through a complex sound field. Recent advances in computing, machine learning, and neural recording technology now make this problem tractable. There are two specific aims. The first is to simultaneously record from large numbers of auditory cortex neurons during free movement through a calibrated sound field. These experiments will develop the equipment, experimental approach, and computational techniques needed to accurately track the sound input to each ear during movement through an auditory scene. The second aim will evaluate how the position and self-motion impact sound coding in auditory cortex. Recently developed methods use artificial neural networks to predict the activity in single neurons evoked by complex natural sounds. These algorithms will be updated to include body posture and self-motion as inputs, allowing measurement of how response properties may change based on these variables. By characterizing dynamic sound coding in free-moving animals, these studies will provide new insight into how the auditory system processes sound under more natural conditions and can support improved signal processing algorithms for auditory prostheses.

项目摘要 尽管听觉假体持续进展，但听力损失的患者通常很难理解 语音和其他重要的声音在嘈杂的环境中。这部分原因是降级空间和 光谱声音信息是由普通听力听众在解析同时发生的声音中利用的 现实世界。当前对空间处理的理解主要是从受试者是的研究中得出的 相对于声源固定。尽管有这种主导的实验范式，但在现实世界中的听众 条件通常会在空间中移动，同时定向头部以提高其理解能力 听觉信号。前庭，运动和听觉系统之间的神经连接存在 表明听众的运动和身体姿势为听觉系统提供了大量输入 促进听力。更好地理解健康的听觉系统的运作方式 声学环境将支持有关听觉疾病的新疗法。 当前的研究将调查有关听众的运动和身体/头部姿势影响的信息 听觉皮层中的声音处理。从历史上看，自由移动主题的研究受到了 在通过复杂的声场不受约束的运动过程中精确测量听觉输入的难度。 现在的计算，机器学习和神经记录技术的最新进展现在使这个问题 可处理。有两个具体的目标。首先是同时记录大量听觉 自由运动过程中的皮层神经元通过校准的声场。这些实验将发展 设备，实验方法和计算技术需要准确跟踪声音输入 通过听觉场景运动期间的每个耳朵。第二个目标将评估位置和 听觉皮层中的自动撞击声音编码。最近开发的方法使用人工神经网络 预测复杂自然声音引起的单个神经元的活性。这些算法将更新为 包括身体姿势和自我运动作为输入，允许测量响应属性 根据这些变量进行更改。通过表征自由移动动物中的动态声音编码，这些 研究将提供有关听觉系统在更自然条件下听起来的新见解 并可以支持改进的听觉假体信号处理算法。