High dynamic range multiphoton microscopy for large-scale imaging

摘要 多光子显微镜是高分辨率脑功能研究的首选技术之一 由于其在厚组织中的显著深度穿透，在标准配置中， 这种成像包括在整个样品中以3D方式扫描飞秒激光焦点。的 激光功率在扫描期间是固定的，并且图像信息包含在时间依赖性中 检测到的荧光信号。这种技术可能会出现几个问题。一是在 常见的情况下，样品包含亮度的极端变化，例如 在大型胞体和更精细的树突过程之间，通常不可能捕捉到 全范围信号，在扫描亮区时不会使探测器饱和， 在扫描暗淡区域时丢失信号。第二，当对来自 如GCaMP的功能性报告基因，发生无法预测的大的亮度变化 提前，迫使用户使用低照度，以最大限度地减少检测器的可能性 饱和，从而潜在地损害SNR。第三，当执行体积扫描时， 通过扩展的深度范围，单个激光功率要么变得太弱， 深度或在浅深度太强。 我们提出了一个简单的解决方案来解决所有这些问题。解决方案包括积极 使用反馈电子器件逐像素地调节激光功率。我们已经证明 我们的技术可以将双光子显微镜的动态范围提高几个数量级 幅度为20微秒的适度快速像素时间，实现前所未有的高动态 范围（HDR）为1011：1。我们对这个项目的目标是： 1)用于视频速率HDR成像的超快反馈电子器件的开发。 2)用于大规模多区域HDR成像的切换复用技术的发展。 3)多光子HDR成像在小鼠脑解剖和功能成像中的应用。 我们的目标是实现全面的大规模多光子成像， 动态范围以简单的方式，可以很容易地实现许多实验室， 合理的成本和最小的硬件修改。

ABSTRACT Multiphoton microscopy is one of the preferred techniques for high-resolution functional brain imaging because of its remarkable depth penetration in thick tissue. In standard configurations, such imaging involves scanning a femtosecond laser focus in 3D throughout a sample. The laser power is fixed during the scan and image information is contained in the time dependence of the detected fluorescence signal. Several problems can occur with this technique. First, in common cases where the sample contains extreme variations in brightness, for example between large somas and much finer dendritic processes, it is often impossible to capture the full range of signals without either saturating the detector when scanning over bright regions, or losing signal when scanning over dim regions. Second, when imaging time-varying signals from functional reporters such as GCaMP, large brightness variations occur that cannot be predicted in advance, forcing the user to use a low illumination to minimize the possibility of detector saturation, thus potentially compromising SNR. Third, when performing volumetric scans through an extended range of depths, a single laser power becomes either too weak at large depths or too strong at shallow depths. We propose a simple solution to solve all these problems. The solution involves actively regulating the laser power pixel-by-pixel using feedback electronics. We have demonstrated that our technique can improve the dynamic range of two-photon microscopes by several orders of magnitude for moderately fast pixel times of 20s, achieving an unprecedentedly high dynamic range (HDR) of 1011:1. Our goals for this project are the: 1) Development of ultrafast feedback electronics for video-rate HDR imaging. 2) Development of switched multiplexing technique for large-scale multi-region HDR imaging. 3) Application of multiphoton HDR imaging to anatomical and functional mouse brain imaging. Our goal is to enable comprehensive large-scale multiphoton imaging with unprecedented dynamic range in a simple manner that can be readily implementable by many labs at reasonable cost and with minimal hardware modifications.