Gentler Imaging

使用光学显微镜为活细胞成像是一项具有挑战性的工作。光学显微镜中使用的光经常破坏被研究的生物系统，我们要求的分辨率越高，需要的光就越多。这使得如果不开发新的创新解决方案，就几乎不可能在光学显微镜所能观察到的尺寸范围内快速地研究重要的生命过程。通常，一台显微镜可以执行单一的成像模式，所以想要以多种不同的分辨率/光强度成像的研究人员必须在显微镜之间移动样本。我们提出了一个通用的成像平台，可以在成像模式之间切换。这很有用，因为它允许我们用最少的光线想象一个生物系统，直到我们感兴趣的事件发生。当一些有趣的事情发生时，我们想要自动在低分辨率、低损害模式之间切换到高分辨率、高光子流模式--特别是在发生有趣事件的区域。为了自动化这个过程，我们将训练一个神经网络来学习低分辨率图像的产生，并在特定条件下切换到高分辨率模式。这将极大地减少光损伤的量，使我们能够比以前更详细地观察生物事件。我们将使用阿米巴吞噬和杀死微生物作为概念验证的例子。这与我们的免疫细胞用来清除体内病原体以保护我们免受感染的方式大致相同。阿米巴细胞为理解这一点提供了一个方便的模型，也为显微镜的发展提供了一个简单的测试系统。吞噬过程称为吞噬作用，当阿米巴或免疫细胞接触猎物并将其包裹时，就会发生吞噬过程。这将被用来作为触发器，刺激成像平台切换到更高分辨率的模式，特别是在被吞噬的微生物周围。这将防止在吞噬之前对样本进行光破坏，并使我们能够准确地在需要的时间和地点聚焦宝贵的光，以了解吞噬和杀戮是如何发生的。这种空间和时间选择性的高分辨率成像然后将与新的相机技术相结合，以确保以最佳信噪比捕获图像中的每个像素。这将使用户能够确保从每次实验中获得最大的信息。结合起来，这些技术将通过克服当前显微镜的主要限制，极大地提高我们观察一系列重要生物事件的能力。

Imaging living cells using optical microscopy is challenging. The light used in optical microscopy often damages the biological system under investigation, and the better resolution we require, the more light is needed. This makes studing the important living processes that occur both rapidly and at the size limits of what can be observed with light microscope almost impossible without the development of new, innovative solutions.Often a single microscope can perform a single imaging modality, so a researcher wanting to image at a number of different resolutions / light intensities would have to move the sample between microscopes. We propose a versatile imaging platform that can switch between imaging modalities. This is useful because it allows us to image a biological system with minimal light, until the event we are interested in occurs. At the time something interesting happens we then want to automatically switch between the low resolution low damage mode to a high resolution, higher photon flux mode - specially in the area where the interesting event is occurring. To automate this process we will train a neural network to study the low resolution images as they are produced and under specific conditions swap to the high resolution mode. This will dramatically reduce the amount of photo damage, allowing the biological event to be observed in much greater detail than previously possible.The example we will use as a proof of concept is the engulfment and killing of microbes by amoebae. This happens in much the same way that our immune cells use to clear the body of pathogens to protect us from infection. The amoeba cells provide a convenient model to understand this, as well as a simple test system for microscope development. The engulfment process, known as phagocytosis, occurs when amoeba or immune cells touch their prey and enwrap them. This will be used to act as a trigger to stimulate the imaging platform to switch to the higher resolution modality, specifically around the microbe being eaten. This will prevent photodamaging the sample prior to engulfment, and allow us to focus the precious light precisely when and where it is needed to understand how engulfment and killing occur. This spatial and temporally selective high-resolution imaging will then be combined with novel camera technology to ensure that each pixel in the image is captured at the optimal signal-to-noise. This will allow users to ensure that the maximal information is obtained from each experiment. Combined, these technologies will dramatically improve our ability to observe a host of important biological events, by overcoming the main limitation of current microscopy.