Comparing Performance of Spectral Image Analysis Approaches for Detection of Cellular Signals in Time-Lapse...

Area of Science:

• Biophotonics and Imaging
• Cellular and Molecular Biology
• Spectroscopy

Background:

• Hyperspectral imaging (HSI) is versatile, used from remote sensing to medicine.
• Traditional fluorescence microscopy HSI faces signal limitations due to light dispersion.
• Sampling fluorescence excitation spectra offers a promising alternative for stronger signals.

Purpose of the Study:

• To investigate excitation-scanning HSI for dynamic cell signaling studies.
• To analyze calcium (Ca2+) signaling in human airway smooth muscle cells (HASMCs).
• To compare the performance of four spectral analysis algorithms for this application.

Main Methods:

• Acquisition of time-lapse excitation-scanning HSI data for Ca2+ signals in HASMCs.
• Analysis using linear unmixing (LU), spectral angle mapper (SAM), constrained energy minimization (CEM), and matched filter (MF).
• Application of a theoretical sensitivity framework for pixel filtering based on a minimum detectable limit.

Main Results:

• Linear unmixing (LU) and matched filter (MF) algorithms showed linear responses to increasing Ca2+.
• Both LU and MF proved effective for excitation-scanning HSI analysis.
• Pixel filtering, guided by sensitivity analysis, revealed subtle kinetic features and improved signal extraction.

Conclusions:

• Excitation-scanning HSI is suitable for kinetic measurements of cellular signals and dynamic events.
• Algorithm selection and pixel filtering are crucial for quantitative signal extraction.
• This approach is particularly valuable for overcoming challenges posed by cellular autofluorescence and competing signals.