STTR Phase I: Microscope-based Technology For Automatic Brain Cell Counts Using Unbiased Methods

这个小型企业技术转移(STTR)第一阶段项目的更广泛的影响/商业潜力是自动化无偏立体学的过程，这是生命科学中用于组织切片上染色细胞计数的最新方法。无偏见的体视学使神经科学家能够准确地分析脑细胞的大小和数量，这些细胞在许多神经疾病和精神疾病中都会发生变化。由于目前尚不清楚的原因，阿尔茨海默氏症、帕金森氏病和肌萎缩侧索硬化症都与脑细胞进行性丧失有关。相比之下，自闭症儿童生来就有太多的脑细胞，这导致在处理复杂的信息流方面存在终生问题。体视学在研究许多影响大脑的疾病和评估可能的治疗方法的有效性和安全性方面发挥着重要作用。虽然这项拟议的技术最初将以了解和治疗所有神经系统疾病的研究为目标，但它可以用于对所有组织中的细胞进行自动评估，包括癌症筛查和活检诊断。拟议中的项目将开发和优化一种算法，帮助大脑科学家确定神经疾病和精神疾病的原因和治疗方法。这项拟议的技术将使用深度学习(人工智能)系统来自动识别、计数组织切片上的脑细胞并调整其大小。这项技术的一个重要用途将是分析小鼠和大鼠的大脑和神经组织，这些小鼠和大鼠在治疗后表现出与人类相似的神经疾病。这些动物模型为测试治愈人类脑部疾病的治疗方法提供了一个强大的工具。目前，对这些动物组织的体视学研究需要一名训练有素的技术人员坐在电脑屏幕前，对成百上千个微观细胞进行繁琐的手工计数。这种过时的方法未能利用为拟议的软件提出的强大的深度学习方法，这些方法可以以10倍的速度完成这些任务，并且错误和人为偏见更少。因此，拟议的技术将利用深度学习技术来加快美国的基础研究和药物开发，从而建立一个长期的经济引擎，并通过科学突破和医学发现为社会带来重大利益。

The broader impact/commercial potential of this Small Business Technology Transfer (STTR) Phase I project is in automating the process of unbiased stereology, the state-of-the-method used in the life sciences for counting stained cells on tissue sections. Unbiased stereology allows neuroscientists to accurately analyze the size and number of brain cells, which are altered in many neurological disorders and mental illnesses. For reasons that are currently unknown, Alzheimer's disease, Parkinson's disease and Amyotrophic Lateral Sclerosis are all associated with a progressive loss of brain cells. In contrast, children with autism are born with too many brain cells, which leads to life-long problems in processing complex streams of information. Stereology plays an important role in investigating many conditions affecting the brain and assessing the efficacy and safety of possible treatments. Though the proposed technology will initially target research studies to understand and treat all neurological conditions, it can be useful for automatic assessments of cells in all tissues, including cancer screening and diagnosis from biopsies. The proposed project will develop and optimize an algorithm to help brain scientists identify causes and treatments for neurological disease and mental illness. The proposed technology will use deep learning (artificial intelligence) systems to automatically recognize, count and size brain cells on tissue sections. An important use of this technology will be to analyze brains and nerve tissue from mice and rats treated to show similar neurological diseases as those found in humans. These animal models provide a powerful tool for testing treatments to cure brain disease in humans. Currently stereology studies of tissues from these animals require a trained technician to sit before a computer screen making tedious manual counts of hundreds and thousands of microscopic cells. This outdated approach fails to take advantage of powerful deep learning methods proposed for the proposed software that could complete these tasks 10 times faster and with fewer errors and human biases. Therefore, the proposed technology will use deep learning technology to accelerate basic research and drug development in the U.S., thereby establishing a long-term economic engine and bringing significant benefits to society through scientific breakthroughs and medical discoveries.