Computational methods for fast nonlinear optical microscopy of label-free metabolic dynamics in live organisms

Due to constraints in signal-to-noise ratio (SNR) and instrumentation bandwidth, many advanced label-free nonlinear optical microscopy techniques suffer from slow acquisition, restricting applications that require high throughput and/or fast dynamic imaging. To overcome this, computational methods can be used to reconstruct temporal and spectral optical signatures acquired using well-defined and calibrated optics and electronics. High-speed two-photon NAD(P)H and FAD autofluorescence intensity and lifetime is achieved using computational photon counting, bypassing the traditionally slow analog electronics to perform photon counting on directly digitized data. This method enables imaging over 2× faster than the fastest previously demonstrated methods, and 40× faster than standard methods. Coherent anti-Stokes Raman scattering (CARS) microscopes generally use a slow x-y-λ acquisition scheme, and often use dimensionality reduction methods for data analysis in postprocessing. To speed up the acquisition and analysis pipeline, a pulse shaper with phase and amplitude control can be used to selectively determine excitation envelopes to excite customized hyperspectral components to maximize spectral contrast between different chemical species, thus integrating the dimensionality reduction into the acquisition process. These novel methods enable imaging of rapid metabolic (autofluorescence) and biochemical (CARS) dynamics in diverse living biological specimens. Using these new capabilities, the static and dynamic nonlinear optical metabolic signatures of bacteria and biofilms are explored. Due to the lack of previous research on nonlinear optical microscopy of bacteria, baseline characteristics needed to be established for photobleaching, segmentation, single-colony heterogeneity, and the effect of plating conditions. Additionally, differences between planktonic bacteria and biofilms were examined, with biofilms showing a significant shift towards higher protein and lipid content in CARS spectra. Furthermore, the dynamic optical metabolic signatures associated with resistant, bactericidal, and bacteriostatic antibiotic response were explored. Results indicate that FAD intensity and lifetime are immediately indicative of antibiotic response. Overall, this work opens the door to many new studies by enabling faster and higher-throughput imaging, and suggests that nonlinear optical microscopy is a promising method for bacteria characterization.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Bioengineering

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2024

Author dc:creator

• Sorrells, Janet E.

Contributors dc:contributor

• Boppart, Stephen A
• Bhargava, Rohit
• Maslov, Sergei
• Smith, Andrew

• Nonlinear Optics
• Computational Imaging
• Optical Microscopy
• Fluorescence Lifetime
• Coherent Raman

Statement dc:rights

• © 2024 Janet E. Sorrells

Language dc:language

en, eng

Handle dc:identifier

https://hdl.handle.net/2142/125685

OAI record oai:identifier

https://www.ideals.illinois.edu/oai-pmh?verb=GetRecord&metadataPrefix=etdms&identifier=oai%3Awww.ideals.illinois.edu%3A2142%2F125685