"Grid-Tape": A High-Throughput Platform for Brain Connectomics and Nanoscale Structural Analysis

脑回路的线路图是现代科学的基础和基本问题之一， 神经科学由于这些电路的连接性严重影响它们的功能，因此了解 这些网络的结构将对了解它们在健康和疾病方面的作用产生重大影响。的 主要的挑战是以突触分辨率提取电路大小的体积，这需要大规模的电子 显微镜成像数千个脑组织切片。到目前为止，组织切片的缓慢速度和 成像一直是连接组学领域的主要限制，需要多年来获取 这个尺寸。在第一阶段开发的GridTape™通过允许快速自动成像， 在透射电子显微镜（TEM）内的连续带上的数千个切片。新的卷轴- 带卷样品基底利用基于相机的TEM的固有速度和分辨率优势， 与扫描电子显微镜（SEM）替代品相比，具有更高的数据采集速率和更低的成本。在第一阶段， GridTape™通过成功获取立方毫米的数据，实现了连接组学的长期目标。 体积在4纳米分辨率，产生超过2 PB的数据集。数据跨越26，500个串行 切片在GridTape™的七个卷轴之间分开，并且用五个TEM显微镜成像， 以500 Mpixel/s的有效成像速率并行六个月。虽然最好的多光束扫描电镜替代品 使用GridTape™可以实现3倍于单个TEM的成像速率，成本大约高出20倍（400 - 600万美元）， 目前世界上只有少数几个多波束SEM;这大大推动了体积EM的经济性， 支持使用多个更便宜的TEM与GridTape™并行的方法。明确可行性 在第二阶段，Luxel将与哈佛合作，改善一些数据质量问题， GridTape™.技术挑战包括减少来自本征结构的背景图像噪声， 支撑膜降低了薄膜涂层的成本，减轻了由组织裂纹引起的图像伪影， 和折叠，并扩展膜覆盖的槽区域，以允许更大的组织样本而不断裂。我们计划 为了实现这些目标，开发替代的低噪音聚合物薄膜， 包括更快的衬底去除的可缩放性益处。我们将与连接组学研究人员合作， 哈佛和艾伦研究所确定最佳的组织块制备配方和方法，以保持 薄膜张力，减轻组织裂纹和褶皱。

The wiring diagram of brain circuits is one of the foundational and fundamental questions of modern neuroscience. Since the connectivity of these circuits critically influences their function, understanding the structure of these networks will have a major impact on understanding their role in health and disease. The primary challenge is extracting a circuit-sized volume at synaptic resolution, which requires large-scale electron microscopy imaging of thousands of sections of brain tissue. Until now, the slow speed of tissue sectioning and imaging has been a major limitation to the field of connectomics, requiring many years to acquire a dataset of this size. GridTape™, developed in Phase I, removes this bottleneck by allowing fast automated imaging of thousands of sections on a continuous tape inside a transmission electron microscope (TEM). This new reel- to-reel sample substrate leverages inherent speed and resolution advantages of camera-based TEM to provide higher data acquisition rates with lower cost than scanning electron microscope (SEM) alternatives. In Phase I, GridTape™ achieved a longstanding goal for connectomics with the successful acquisition of a cubic mm volume at 4 nm resolution, yielding a dataset of more than 2 petabytes. The data spanned 26,500 serial sections split between seven reels of GridTape™, and was imaged with five TEM microscopes running in parallel for six months at an effective imaging rate of 500 Mpixel/s. While the best multi-beam SEM alternative can achieve 3X the imaging rate of a single TEM with GridTape™, the cost is roughly 20X higher ($4-6M) and only a few multi-beam SEMs currently exist in the world; this drives the economics of volume-EM heavily in favor of an approach using multiple cheaper TEMs in parallel with GridTape™. With feasibility clearly demonstrated, Luxel will partner with Harvard in Phase II to improve a number of data quality issues for GridTape™. The technical challenges include reducing the background image noise from intrinsic structure in the support films, reducing costs for the thin film coating, mitigating image artifacts that arise from tissue cracks and folds, and expanding the film-covered slot areas to allow larger tissue samples without breakage. We plan to achieve these goals by developing alternative low-noise polymer films that also offer manufacturing scalability benefits including faster substrate removal. We will partner with connectomics researchers at Harvard and the Allen Institute to determine optimal tissue block preparation formulas and methods to maintain film tension that mitigate tissue cracks and folds.