New technologies for in vivo spectral resolved high speed multiphoton microscopsy

 描述(申请人提供)：光谱分辨成像在众多生物学研究中普遍存在，从绘制突触动力学图，到监测细胞内信号，以及研究蛋白质-蛋白质相互作用。能够独立监测细胞结构的寿命和动力学，如突触、细胞核、蛋白质运输小泡和各种其他多组分复合体，对于揭示它们的细胞功能以及它们的组装和拆卸是至关重要的。虽然使用共聚焦显微镜对同一细胞中3-4种不同蛋白质的光谱分辨可视化是相当常规的 在固定的脑切片或细胞培养中，体内动态多蛋白质成像仍然是一个挑战，然而许多细胞内和细胞间的相互作用依赖于完整组织的背景。我们的目标是开发和实施与高通量多光子显微镜兼容的光谱分辨技术，以允许对多组分亚细胞结构进行大体积的在体成像。在前两个目标中，我们建议测试两种新的光谱方法，用于大容量、高速成像，具有各自的优势和劣势，可以根据不同的成像需求进行定制。在第三个目标中，我们将开发一种高效的基于小波的泊松去噪光谱分解算法，该算法可以通过对信噪比低得多的图像进行准确分析来潜在地增强这两种方法。目的1：设计一种色散光谱仪(DS)，用于多焦多光子显微镜(MMM)系统中的高光谱成像。目的2：设计一种傅里叶变换光谱仪(FTS)，使其能够在MMM和广域多光子显微镜(WFMM)系统中进行高光谱成像。目的3：提出一种形态引导的泊松去噪最大似然(MLE)谱分解算法，以降低原始图像的信噪比要求。

 DESCRIPTION (provided by applicant): Spectrally-resolved imaging is ubiquitous in numerous biological studies ranging from mapping synapse dynamics, to monitoring of intracellular signaling, and studying protein-protein interactions. The ability to independently monitor the lifetime and dynamics of cellular structures, such as the synapse, the nucleus, protein trafficking vesicles, and various other multicomponent complexes is critical to revealing their cellular function as well as their assembly and disassembly. While spectrally resolved visualization of 3-4 different proteins in the same cell is quite routine using confocal microscopy in fixed brain sections or in cell culture, dynamic multi-protein imaging in vivo remains a challenge, yet many intra- and inter- cellular interactions are dependent on the context of an intact tissue. Our goal is to develop and implement spectrally resolved technologies that are compatible with high throughput multiphoton microscopy to allow large volume, in vivo imaging of multicomponent subcellular structures. In the first two aims we propose testing two novel spectrometric approaches for large volume, high-speed imaging, with respective strengths and weaknesses, that can be tailored to tackle different imaging needs. In the third aim, we will develop a highly efficient wavelet based Poisson denoised spectral un-mixing algorithm that can potentially enhance both approaches by allowing accurate analysis of images with far lower SNR. Aim 1: Design a dispersive spectrometer (DS) to enable hyperspectral imaging in a multifocal multiphoton microscope (MMM) system utilizing multianode PMTs (MAPMT). Aim 2: Design a Fourier transform spectrometer (FTS) to enable hyperspectral imaging in MMM and wide-field multiphoton microscopy (WFMM) systems. Aim 3: Develop a morphology-guided, Poisson-denoised maximum likelihood (MLE) spectral decomposition algorithm to reduce SNR requirement of raw images.