Diffuse Optical Brain Imaging

漫反射光学成像(DOI)使我们能够获得组织中的血流动力学响应。已经证明，我们可以通过检测神经-血管耦合的影响，将其与大脑中的功能事件联系起来。DOI的现有技术受到许多因素的限制，但最严重的是缺乏可行的临床应用。最初，我们的目标是接受脑半球根治术的年轻癫痫患者，以及患有穿透性脑损伤(TBI)的退伍军人(在儿科人群中处于上端)。今年，通过我们的持续推广，我们确定了第三个关键人群--自闭症谱系障碍(ASD)患者。这些患者是低功能人群，通常是儿科人群。 在2010财年的过程中，我们发展了我们的合作，为我们的一个新仪器想法获得了专利，并在神经科学学会接受了几次演讲。我们正在实验室(通过我们在德雷克斯的合作者)和乔治敦的其他合作者实验室组装各种仪器。后者是一种与脑电相结合的DOI仪器。 临床上，我们现在开始对健康志愿者进行初步测试。我们正在等待我们通过NICHD新批准的IRB的协议号，以便在NICHD开始测试我们的原型系统。为了进行脑功能成像实验，我们在E-Prime软件中开发了一系列认知任务(事件复杂性判断任务、言语工作记忆任务和多任务)。每个刺激的呈现时间将被准确地记录下来，用于数据分析。我们还研究了从光学数据中提取脑血流动力学响应的数据处理技术。最初的实验是与乔治敦大学合作使用事件复杂性任务进行的。将主成分分析(PCA)和独立成分分析(ICA)用于去除噪声和运动伪影。相应的提取的血流动力学响应验证了该技术，因为它们与先前的功能磁共振研究结果相关联。此外，我们目前正在与NIMH的合作者讨论对ASD患者有用的合适的功能研究，一旦我们有了关于健康志愿者的初步数据，我们将对ASD患者进行一项研究，无论是年轻人还是儿童人群。 从理论上讲，我们目前正在研究各种方法来解决当前DOI技术的缺点。我们目前的主要重点是解决DOI的地图集和注册问题。这一领域是一个热门的研究话题，但与核磁共振不同，在核磁共振中，Talairach和MNI地图集已经成为事实上的标准，DOI没有这样的标准方法。部分原因是重建方法缺乏确定性，部分原因是数据联合登记的问题。作为该项目的一部分，我们一直致力于更好地量化和本地化光学信号。然而，我们目前的重点一直是通过将数据映射到不同的坐标基础来重新编排地图集和注册。我们目前正在完成一篇基于将光学数据移动到极/球基(这对其灵敏度和分辨率更自然)的原理的论文。通过这样做，我们可以基于2D表面流形的有效配准来直接在3D体积中配准数据。这种方法还将与我们正在开发的用分形维描述光学成像的灵敏度的工作相结合。这将使我们能够为DOI进化出更复杂的人脑概率功能图谱。 我们与NINDS的合作继续开发一种新的基于光纤的成像仪，以增加我们用于DOI技术比较和对比的设备选择。此外，我们正在与Drexel的合作者一起开发一个完全微型化的系统。实现了光发射机和光接收机的初步设计，并对各个模块的功能进行了实验验证。发射机采用四路垂直腔面发射激光二极管，并利用梯度折射率(GRIN)透镜技术实现了优异的光收集效率。我们目前正在努力整合这些模块。这些项目将结合在一起，生产用于近红外功能成像的完全可穿戴的DOI。我们的理论研究也将运动伪影作为一种信号而不是噪声来考察。初步结果表明，如果我们能够通过头盔接口(仪器专家正在研究的过程)模拟成像系统相对于对象的运动，我们应该能够进一步将光学信号提高到该领域前所未有的精度和量化水平。 最后，我们提交并获得了一项用于检测血肿的新型仪器设计的初步专利。它旨在协助对各种形式的脑损伤患者进行分诊。识别和分类血肿的能力将极大地提高CT和MRI等更昂贵和有限的访问系统的使用效率。目前，我们正在制作完整的理论模型来测试设计。一旦完成，这个模型将允许我们快速发展到最初的原型。该项目也对近红外建模的现状提出了挑战。初步结果表明，更多的源和更高密度的探测器更适合成像的范式正在受到挑战。这个项目的初步结果表明，在改进的仪器和由建模的计算边界引入的误差之间将会有一个很好的平衡。

Diffuse Optical Imaging (DOI) allows us access to the hemodynamic response in tissue. It has been shown that we can, by detecting the effects of neuro-vascular coupling, relate this to functional events in the brain. Existing technologies in DOI are limited by a number of factors, but most strongly by the absence of viable clinical applications where they may be applied. Originally we were targeting young epilepsy sufferers who have had radical hemispherectomies, and veterans (at the upper end of the pediatric population) with penetrating Traumatic Brain Injury (TBI). This year we have, through our continuing outreach, identified a third key population, Autistic Spectrum Disorder (ASD) patients. These patients are low-functioning and typically pediatric populations. Over the course of the 2010 fiscal year we have developed our collaborations, secured a patent for one of our novel instrument ideas, and had several presentations accepted at the Society for Neuroscience. We are assembling a variety of instruments both in the lab (through our collaborators at Drexel) and at our other collaborators lab in Georgetown. The latter is a DOI instruments combined with EEG. Clinically we are now commencing initial testing with healthy volunteers. We are awaiting the protocol number for our newly approved IRB through NICHD to begin testing on our prototype systems here at NICHD. To conduct functional brain imaging experiments, we have developed a number of cognitive tasks (event complexity judgment task, a verbal working memory task, and a multi-task) in E-prime software. The timing of the presentation of each stimulus will be accurately recorded which will be used in data analysis. We have also worked on data processing techniques to extract the brain hemodynamic response from the optical data. Initial experiments were performed using the event complexity task in collaboration with Georgetown. Principal component analysis (PCA) and independent component analysis (ICA) have been applied to the data to remove the noise and motion artifact components. Corresponding extracted hemodynamic responses validate the technique as they correlate with the results of previous fMRI studies. In addition, with collaborators at NIMH, we are currently discussing suitable functional studies useful to ASD patients, and once we have initial data on healthy volunteers, we will produce a study on ASD patients, either in a young adult or pediatric population. Theoretically we are currently working on a variety of approaches for handling the shortcomings of current DOI techniques. Our primary focus at the moment is addressing the issue of atlasing and registration for DOI. This area is a hot topic of research but unlike MRI where Talairach and MNI atlases have become de facto standards DOI has no such standard approach. Part of this is the absence of certainty in reconstruction methodologies and part is the issue of co-registering data. We have been working on better quantitation and localization of optical signals as part of the project. Our current focus though has been on readdressing atlasing and registration by mapping data to a different co-ordinate basis. We are currently finishing a paper based on the principle of moving optical data into a polar/spherical basis (which is more natural to its sensitivity and resolution). By doing so we can directly register data in the 3D volume based upon an effective registration of the 2D surface manifold. This method will also combine with work we are developing on describing the sensitivity of optical imaging in fractal dimensions. This will combine to allow us to evolve much more sophisticated probabilistic functional atlases of the human brain for DOI. Our collaboration with NINDS continues to develop a new fiber based imager to add to our selection of devices for comparison and contrast of DOI techniques. Also we have a fully miniaturized system under development in association with our collaborators at Drexel. Initial designs for optical transmitter and receiver have been implemented and the functionality of each module has been verified experimentally. Quad vertical cavity surface emitting laser diodes are used in the transmitter, and the system takes advantage of Gradient-Index (GRIN) lens technology to achieve excellent optical collection efficiency. We are currently working on the integration of the modules. These projects will combine to produce fully wearable DOI for Near Infrared functional imaging. Our theoretical research is also examining motion artifact as a signal instead of noise. Initial results suggest that if we can model the motion of our imaging system relative to the subject via the helmet interface (a process being worked on by our instrumentalists), we should further be able to enhance optical signals to an unprecedented level of accuracy and quantitation in the field. Finally we have submitted and received an initial patent to a novel instrument design intended to detect hematomas. It is designed to assist in the triage of patients with all forms of TBI. The ability to identify and triage for the presence of hematoma would greatly increase efficiency of use of more expensive and limited access systems such as CT and MRI. Currently we are working on the full theoretical model to test the design. Once complete this model will allow us to progress rapidly to an initial prototype. The current status-quo in NIR modeling is also being challenged by this project. Initial results indicate that the paradigm that more sources and detectors at higher density is better for imaging is being challenged. Initial results from this project suggest there will be a fine balance between improved instrumentation and error introduced by the computational boundaries of modeling.