| 摘要: | 在螢光顯微影像技術的早期應用中，研究者主要依賴螢光分子的光譜特性差異，搭配適當的濾光片波段，來區分不同種類的螢光分子。然而，此種辨識方式高度依賴螢光強度，而螢光強度容易受到激發光功率、螢光分子濃度等因素的影響，進而降低分析準確性。特別是在多重標定或複雜生物樣本中，若螢光分子的發射光譜高度重疊，僅透過光譜訊息將難以有效區分不同螢光分子。為克服這些限制，後續研究人員發現螢光生命週期（Fluorescence Lifetime）是一項穩定且具辨識力的參數。螢光生命週期定義為螢光分子在被激發後返回基態發光的時間長度，其不受激發光強度以及分子濃度影響，故本研究使用螢光生命週期結合掃描影像顯微術（Fluorescence Lifetime Imaging Microscopy, FLIM）， 並使用頻域(frequency domain)量測方法，提供除螢光光譜以外的時間維度資訊，以提升螢光分子的判別能力。在頻域量測中，透過分析螢光訊號與激發光訊號之間的調制深度差與相位差來推算螢光生命週期，其中調制深度量測是藉由快速傅立葉轉換(Fast Fourier transform, FFT)，分析其直流項與交流項，相位量測部分採用零差檢測（homodyne detection）方式，將參考訊號調制成弦波訊號並改變八個不同的相位，接著將螢光訊號與參考訊號相乘後取其直流項重建出螢光相位 。本研究成功透過螢光生命週期影像區分不同的螢光分子，並進一步透過模擬與實驗數據，比較由相位差與調制深度所推算出的螢光生命週期在抗雜訊能力方面的表現。結果顯示，相較於調制深度法，基於相位資訊所計算出的螢光生命週期在雜訊環境下具有更高的穩定性與準確性。此外，由於本系統的取樣頻率為100MHz，而所選用的調制頻率為20 MHz與40 MHz，等於一個週期僅對應5個取樣點與2.5個取樣點，無法完整還原真實的訊號，進而影響螢光生命週期的準確性。為考慮此問題，進行另一個模擬實驗，在有雜訊情況下探討不同取樣率對螢光壽命量測結果的影響。結果顯示，即使在低訊號雜訊比（SNR）情況下，較高的取樣率能夠提高訊號的還原完整度與螢光生命週期的計算準確性。;In the early applications of fluorescence microscopy, researchers primarily relied on the spectral characteristics of fluorophores, using appropriate filter sets to distinguish between different types of fluorescent molecules. However, this method is highly dependent on fluorescence intensity, which can be significantly affected by factors such as excitation light power and fluorophore concentration, thereby reducing analytical accuracy. This limitation becomes particularly problematic in multiplexed labeling or complex biological samples, where significant spectral overlap among fluorophores makes it difficult to distinguish them based solely on spectral information.
To overcome these limitations, subsequent studies identified fluorescence lifetime as a stable and distinguishable parameter. Fluorescence lifetime is defined as the time a fluorophore remains in the excited state before returning to the ground state and emitting fluorescence. Since fluorescence lifetime is independent of excitation intensity and fluorophore concentration, this study employs Fluorescence Lifetime Imaging Microscopy (FLIM), using a frequency-domain measurement approach to provide additional temporal information beyond fluorescence spectra, thereby enhancing the ability to differentiate fluorophores.
In frequency-domain measurements, fluorescence lifetime is derived from the modulation depth and phase shift between the fluorescence signal and the excitation light. Modulation depth is obtained through Fast Fourier Transform (FFT) by analyzing the DC and AC components of the signal. Phase measurement is performed using homodyne detection, in which the reference signal is modulated into a sinusoidal waveform with eight different phase shifts. The fluorescence signal is then multiplied by the reference signal, and its DC component is extracted to reconstruct the fluorescence phase.
This study successfully distinguishes different fluorophores based on their fluorescence lifetimes. Furthermore, through both simulations and experimental data, we compare the noise robustness of fluorescence lifetimes derived from phase shifts versus modulation depth. The results show that fluorescence lifetime values calculated from phase information exhibit greater stability and accuracy under noisy conditions compared to those calculated from modulation depth.
In addition, considering that the system′s sampling rate is 100 MHz and the chosen modulation frequencies are 20 MHz and 40 MHz—corresponding to only 5 and 2.5 sampling points per cycle, respectively—accurate reconstruction of the true signal becomes challenging, which in turn affects the accuracy of the fluorescence lifetime measurement. To address this issue, an additional simulation was conducted to investigate the impact of different sampling rates on fluorescence lifetime measurement under noisy conditions. The results demonstrate that, even under low signal-to-noise ratio (SNR) conditions, a higher sampling rate improves signal fidelity and enhances the accuracy of fluorescence lifetime calculations. |
