An enhancement to nGoggle: a VR Goggle for Structural and Functional assessment in Glaucoma

项目摘要 视神经的功能丧失和结构损伤的评估是典型的评估， 在青光眼的诊断和监测中形成。本方案旨在开发硬件系统 这将允许使用nGoggle（一种便携式脑机接口（BCI））采集眼底照片 它集成了一个可穿戴的无线脑电图（EEG）系统和一个头戴式显示器， 监测与视野刺激相关的脑电活动。在以前的项目中，我们开发了 nGoggle便携式系统用于评估视功能缺陷，以解决标准的局限性 自动视野检查（SAP），一种传统但主观的视野检查。我们比较了 nGoggle和SAP在区分青光眼患者和健康受试者中的作用，并发现BCI 即使不比SAP更好，至少也能表现得一样好，而且具有可移植性和客观性的额外优势。 随着智能手机摄像头的快速发展，现在可以开发紧凑的自动对焦 眼底摄像头并将其集成到我们的神经监测nGoggle中。最近我们开发了一种方法 名为机器对机器（M2M）的深度学习算法，经过训练可以分析眼底照片， 预测由光谱提供的视网膜神经纤维层厚度和神经视网膜边缘的定量测量 域光学相干断层扫描（SDOCT），与原始SDOCT具有高度相关性和一致性 意见。这种方法开辟了使用简单眼底照片提取定量的可能性， 青光眼的神经损伤通过结合目前由 nGoggle具有结构评估能力，这种新系统将导致一种非常独特的设备， 便携式结构和功能评估的视神经损伤的多种眼睛的条件，不限于 青光眼，如年龄相关性黄斑变性、视网膜变性和非青光眼性视神经病变， 同情心 第一阶段SBIR项目的具体目标是：（1）将高分辨率自动聚焦 眼底相机到nGoggle头戴式显示器完全符合眼科仪器安全标准- ards和眼底照相机功能标准，以及（2）展示nGoggle眼底照相机的能力 从模型眼睛中捕捉高质量的视觉神经图像。该系统的成功开发将使 容易且廉价地进行视网膜状况的结构和功能评估， 便携式设备。

PROJECT SUMMARY Assessments of functional loss and of structural damage to the optic nerve are typical evaluations per- formed in the diagnosis and monitoring of glaucoma. This proposal aims at developing the hardware system that will allow acquisition of fundus photographs using the nGoggle, a portable Brain-Computer Interface (BCI) which integrates a wearable, wireless, electroencephalogram (EEG) system and a head-mounted display to monitor electrical brain activities associated with visual field stimulation. In a previous project, we developed the nGoggle portable system for the assessment of visual function deficits in order to address limitations of standard automated perimetry (SAP), a traditional but subjective visual field test. We compared the diagnostic accuracies of the nGoggle and SAP in discriminating patients with glaucoma from healthy subjects and found that the BCI performed at least as well, if not better than SAP, with the additional advantage of portability and objectivity. With rapid advancements in smartphone cameras, it is now possible to develop compact auto-focusing fundus cameras and integrate them into our neuro-monitoring nGoggle. Recently we developed an approach named Machine-to-Machine (M2M), a deep learning algorithm that was trained to analyze fundus photos and predict quantitative measurements of retinal nerve fiber layer thickness and neuroretinal rim provided by spectral domain-optical coherence tomography (SDOCT), with high correlation and agreement with the original SDOCT observations. This approach opens up the possibility of using simple fundus photographs to extract quantitative information about neural damage in glaucoma. By combining functional assessment currently provided by the nGoggle with structural assessment capability, this new system would result in a very unique device capable of portable structural and functional assessment of optic nerve damage for a multitude of eye conditions, not limited to glaucoma, such as age-related macular degeneration, retinal degenerations and non-glaucomatous optic neu- ropathies. The specific aims of this Phase I SBIR project are (1) to incorporate a high-resolution auto-focusing fundus camera to the nGoggle head-mounted display in full compliance with ophthalmic instrument safety stand- ards and fundus camera functional standards, and (2) to demonstrate the ability of the nGoggle fundus camera to capture high-quality optical nerve images from a model eye. Successful development of this system will enable both structural and functional assessments of retinal conditions to be performed easily and inexpensively using a portable device.