Multispectral fluorescence imaging for odorant discrimination and visualization

Abstract

A concept based on multispectral fluorescence imaging was proposed in this paper for the discrimination and visualization of odorants. Fluorescent dyes with different excitation/emission spectra were mixed into agarose gel to prepare a multiple probe sensing film. Odorants remained in environment were recorded on the sensing film via a process called odorant exposure. The odorant-induced fluorescence change of the film under various excitation lights was captured by a CCD camera to obtain multispectral images. It was demonstrated that the use of multiple fluorescence probes provided discrete emission bands, which increased the dimensions of vector space of the multispectral images. Complicated interactions between probes and probes, probes and odorants resulted in the diversiform fluorescence change patterns of the images. Combined with principal component analysis (PCA), different odorants could be discriminated and clustered in the principal component spaces in association with their molecular structures. A hand-shape odorant mark with region-segmented components was visualized with high spatial resolution. Additionally, the technique also succeeded in the visualized demonstration of an airflow containing mixed odorants. Compared with the existing gas and odor sensing technologies, the multispectral fluorescence imaging can be used not only to discriminate different odorants, but also to visualize their time-averaged spatial distribution in environment. Due to its novelty and high information acquisition ability, it can be expected as a new and powerful tool in odor sensing.