基于神经网络原理的荧光寿命成像研究

荧光寿命成像技术(fhlorescence lifetime imaging，FLIM)是一种新颖且功能强大的、能用于复杂生物组织和细胞结构与功能分析的生物组织成像技术。传统的时域荧光寿命成像数据分析方法，由于没有考虑荧光分子团之间以及他们与周围环境的相互作用，可能导致复杂的连续分布荧光寿命这一实际情况，因此对生物组织中自发荧光发光强度衰减过程的实验数据拟合效果欠佳。文章提出利用人工神经网络(artificial neural network，ANN)原理拟合算法来计算生物荧光分子团衰减动力过程，该方法能有效地建立生物荧光分子团衰减动力过程的非线性模型，并且具有处理非线性模型能力强、鲁棒性好、拟合精度高和所需计算时间少等优点。通过计算证明，相对于单参量指数与多参量指数衰减函数，这种数据拟合方法对于某些荧光分子团的多槽基面效价测定样品(multi-well plate assays)的数据有更好的一致性和更小的计算量。同时在文章中讨论了将该拟合算法应用于荧光寿命成像的前景。

Fluorescence Lifetime Imaging Based on Neural Network Algorithm

Fluorescence lifetime imaging is a rather effective and powerful method that can be used to analyze complex biological tissues and molecules. However, the traditional approaches of data analysis did not provide a good fit data of auto fluorescence decay data due to the lack of consideration of a continuous distribution of fluorescence lifetime generated from interactions either among fluorophores or between the fluorophores and their environments. In the paper, it was put forward that a neural network algorithm was likely to provide a truer representation of the underlying fluorescence dynamics. The nonlinear model of fluorescence dynamics can be established effectively by using this method. It has advantages of robustness (the initial value of fit is free), stronger ability to treat nonlinear model, better fitting precision and much less processing time. As compared with those of single exponential and multi-exponential decay functions, the novel model can yield the better goodness of fit and more effective calculation using the data from multi-well plate assays of interesting fluorophores chemically and biologically. In the same time, the potential application of neural network algorithm to fluorescence lifetime imaging was discussed.