Computational method to improve signal detection in NIRS instruments

描述（由申请人提供）：近红外光谱（NIRS）是一种有前途的和快速发展的技术，具有几个独特的特点，如便携式，低成本和多功能。它是唯一对组织血流动力学（慢信号）和神经元活动（快信号）敏感的光谱技术。限制其应用的问题之一是在记录的信号中的表面组织层的贡献以及被称为“全局干扰”的其他伪影和系统生理信号。这些不良影响在测量光信号强度的连续波NIRS仪器中尤其突出。我们的提案是对RFA-RR-09-001的响应，旨在通过使用基于独立成分分析（伊卡）的先进信号处理技术优化数据采集，提高NIRS仪器的灵敏度，选择性和信噪比。它包括1）信号分解成统计上独立的分量，2）识别伪分量和3）通过信号恢复去除伪分量。我们的初步数据表明，在近红外光谱信号的功能相关的变化的信噪比可以显着提高通过ICA为基础的信号处理。结果，可以以毫秒范围内的时间分辨率可靠地记录光学信号中的功能相关的瞬态变化。该项目的目标是：1）通过基于ICA的算法开发改进的仪器和数据采集方法，以可靠地检测来自感兴趣区域组织深层的光谱信号;（二）为了证明伊卡方法在神经生理学实验中的有效性，健康受试者在快速物体识别的认知任务期间测量血液动力学和快速光学信号，视觉和听觉模态;和3）将伊卡方法作为软件工具箱来实施，该软件工具箱被用作NIRS仪器的组成部分，允许研究者在实验期间优化数据采集设置。我们将开发、测试、验证和表征拟议的计算算法在提高非侵入性NIRS技术功能能力方面的有效性。该研究将建立该方法的潜在效用，用于改善健康和疾病中生物组织的非侵入性评估和监测。 公共卫生相关性（由申请人提供）：非侵入性光学成像技术，如近红外光谱（NIRS），提供了许多潜在的优势，超过现有的技术，如MRI或PET监测生物组织。NIRS技术对受试者体内金属物体的存在不敏感（MRI研究的一个禁止因素），对受试者运动不太敏感，并且相对便宜。NIRS应用中的限制因素之一是组织表层的影响，其掩盖了深层的贡献，并可能产生不期望的噪声成分（伪影）。在拟议的研究中，我们将开发一种计算方法，以提高一种新的信号处理技术（独立成分分析或伊卡）的基础上的NIRS仪器的灵敏度和选择性。去噪和优化算法的成功应用将鼓励低成本NIRS技术在基础研究和临床实践中的广泛应用。我们的项目将为进一步开发计算工具奠定基础，改进近红外技术，并增加其在各种功能条件下监测生物组织的有用性。随着灵敏度和选择性的提高，NIRS技术将在床边环境中的许多病理类型的基础研究和临床研究中找到许多应用。

DESCRIPTION (provided by applicant): Near-infrared spectroscopy (NIRS) is a promising and rapidly developing technology with several unique features such as portability, low cost and multi-functionality. It is the only spectroscopic technique which is sensitive to tissue hemodynamics (slow signal) and neuronal activity (fast signal). One of the problems limiting its application is the contribution of superficial tissue layers in the recorded signal as well as other artifacts and systemic physiological signals termed "global interference". Those undesirable effects are especially prominent in continuous-wave NIRS instruments, which measure the intensity of the optical signal. Our proposal is a response to RFA-RR-09-001 and aims to improve the sensitivity, selectivity and signal-to-noise ratio of the NIRS instruments through optimization of data acquisition using the advanced signal processing technique based on Independent Component Analysis (ICA). It includes 1) signal decomposition into the statistically independent components, 2) identification of artifactual components and 3) their removal through signal restoration. Our preliminary data show that the signal-to-noise ratio of the functionally relevant changes in the NIRS signal can be significantly improved through the ICA-based signal processing. As a result, functionally relevant transient changes in the optical signal can be reliably recorded with a temporal resolution in the millisecond range. The aims of the proposed project are: 1) to develop improved instrumentation and data acquisition methods through ICA-based algorithms to reliably detect optical spectroscopic signals from the deeper layers of tissue in the area of interest; 2) to demonstrate the effectiveness of the ICA method in neurophysiological experiments with healthy subjects measuring the hemodynamic and fast optical signals during cognitive tasks of rapid object recognition in visual and auditory modality; and 3) to implement the ICA method as a software toolbox to be used as an integral part of the NIRS instruments allowing the investigator to optimize the data acquisition setup during the experiment. We will develop, test, validate and characterize the proposed computational algorithm in terms of its effectiveness in increasing the functional capabilities of non-invasive NIRS technology. The study will establish the potential utility of the method for improved noninvasive assessment and monitoring of biological tissue in health and disease. PUBLIC HEALTH RELEVANCE (provided by the applicant): Noninvasive optical imaging techniques, such as near-infrared spectroscopy (NIRS), offer a number of potential advantages over existing techniques such as MRI or PET for monitoring biological tissue. NIRS technology is insensitive to the presence of metal objects in the subject's body (a prohibiting factor for MRI studies), less sensitive to subject movement and relatively inexpensive. One of the limiting factors in the application of NIRS is the effect of superficial layers of tissue which obscure the contribution of the deeper layers and may create undesirable noisy components (artifacts). In the proposed studies, we will develop a computational method to improve the sensitivity and selectivity of the NIRS instruments based on a novel signal processing technique (Independent Component Analysis or ICA). The demonstrated successful application of de-noising and optimization algorithms would encourage further application of low-cost NIRS technology in a wide variety of possible applications in basic research and clinical practice. Our project will set a stage for further development of computational tools improving the NIRS technology and increasing its usefulness for monitoring biological tissue during various functional conditions. With increased sensitivity and selectivity, the NIRS technology will find numerous applications in the basic research as well as clinical studies of many types of pathology in a bedside setting.