High-speed hyperspectral microscopy: Fourier transform spectroscopy vs. liquid crystal tunable filters

Hyperspectral fluorescence microscopy holds a great promise for imaging of highly multiplexed, living biological specimens. However, conventional spectral filters such as liquid crystal tunable filters (LCTFs) and acousto-optic tunable filters (AOTFs) cause severe photon loss by rejecting out-of-band photons. To overcome this limitation, we test Fourier transform spectroscopy (FTS) as an enabling platform for high-speed hyperspectral microscopy at a low photon budget. In a head-to-head comparison, we assessed both the LCTF and FTS platforms using two identical camera units and a shared optical train under matching exposure constraints. Using a panel of twelve distinct CD marker-fluorophore bead conjugates, we compare performance across varying signal strengths. By capturing all emission wavelengths simultaneously, the FTS produces a significantly lower noise floor and spectral uncertainty than the LCTF, producing a significantly higher signal-to-noise ratio (SNR). The SNR advantage of the FTS enables clear imaging of dim targets, which are invisible in the LCTF imaging. Across all twelve biomarker combinations, we measured that FTS produced a 42% higher SNR than the LCTF for equivalent peak intensities-defined as the intensity at zero optical path difference for FTS and at the brightest wavelength for the LCTF-and a 2.5-fold to 8.5-fold SNR boost for the same targets. These findings demonstrate that FTS offers a superior architecture for high-speed hyperspectral microscopy.