skip to main content


Title: Deep-learning-assisted Fourier transform imaging spectroscopy for hyperspectral fluorescence imaging
Abstract Hyperspectral fluorescence imaging is widely used when multiple fluorescent probes with close emission peaks are required. In particular, Fourier transform imaging spectroscopy (FTIS) provides unrivaled spectral resolution; however, the imaging throughput is very low due to the amount of interferogram sampling required. In this work, we apply deep learning to FTIS and show that the interferogram sampling can be drastically reduced by an order of magnitude without noticeable degradation in the image quality. For the demonstration, we use bovine pulmonary artery endothelial cells stained with three fluorescent dyes and 10 types of fluorescent beads with close emission peaks. Further, we show that the deep learning approach is more robust to the translation stage error and environmental vibrations. Thereby, the He-Ne correction, which is typically required for FTIS, can be bypassed, thus reducing the cost, size, and complexity of the FTIS system. Finally, we construct neural network models using Hyperband, an automatic hyperparameter selection algorithm, and compare the performance with our manually-optimized model.  more » « less
Award ID(s):
1808331
NSF-PAR ID:
10319104
Author(s) / Creator(s):
; ; ;
Date Published:
Journal Name:
Scientific Reports
Volume:
12
Issue:
1
ISSN:
2045-2322
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. Fluorescence imaging microscopy has traditionally been used because of the high specificity that is achievable through fluorescence labeling techniques and optical filtering. When combined with spectral imaging technologies, fluorescence microscopy can allow for quantitative identification of multiple fluorescent labels. We are working to develop a new approach for spectral imaging that samples the fluorescence excitation spectrum and may provide increased signal strength. The enhanced signal strength may be used to provide increased spectral sensitivity and spectral, spatial, and temporal sampling capabilities. A proof of concept excitation scanning system has shown over 10-fold increase in signal to noise ratio compared to emission scanning hyperspectral imaging. Traditional hyperspectral imaging fluorescence microscopy methods often require minutes of acquisition time. We are developing a new configuration that utilizes solid state LEDs to combine multiple illumination wavelengths in a 2-mirror assembly to overcome the temporal limitations of traditional hyperspectral imaging. We have previously reported on the theoretical performance of some of the aspects of this system by using optical ray trace modeling. Here, we present results from prototyping and benchtop testing of the system, including assembly, optical characterization, and data collection. This work required the assembly and characterization of a novel excitation scanning hyperspectral microscopy system, containing 12 LEDs ranging from 365- 425 nm, 12 lenses, a spherical mirror, and a flat mirror. This unique approach may reduce the long image acquisition times seen in traditional hyperspectral imaging while maintaining high specificity and sensitivity for multilabel identification and autofluorescence imaging in real time. 
    more » « less
  2. Point scanning imaging systems (e.g. scanning electron or laser scanning confocal microscopes) are perhaps the most widely used tools for high resolution cellular and tissue imaging. Like all other imaging modalities, the resolution, speed, sample preservation, and signal-to-noise ratio (SNR) of point scanning systems are difficult to optimize simultaneously. In particular, point scanning systems are uniquely constrained by an inverse relationship between imaging speed and pixel resolution. Here we show these limitations can be miti gated via the use of deep learning-based super-sampling of undersampled images acquired on a point-scanning system, which we termed point -scanning super-resolution (PSSR) imaging. Oversampled ground truth images acquired on scanning electron or Airyscan laser scanning confocal microscopes were used to generate semi-synthetictrain ing data for PSSR models that were then used to restore undersampled images. Remarkably, our EM PSSR model was able to restore undersampled images acquired with different optics, detectors, samples, or sample preparation methods in other labs . PSSR enabled previously unattainable xy resolution images with our serial block face scanning electron microscope system. For fluorescence, we show that undersampled confocal images combined with a multiframe PSSR model trained on Airyscan timelapses facilitates Airyscan-equivalent spati al resolution and SNR with ~100x lower laser dose and 16x higher frame rates than corresponding high-resolution acquisitions. In conclusion, PSSR facilitates point-scanning image acquisition with otherwise unattainable resolution, speed, and sensitivity. 
    more » « less
  3. Hyperspectral imaging technologies have shown great promise for biomedical applications. These techniques have been especially useful for detection of molecular events and characterization of cell, tissue, and biomaterial composition. Unfortunately, hyperspectral imaging technologies have been slow to translate to clinical devices – likely due to increased cost and complexity of the technology as well as long acquisition times often required to sample a spectral image. We have demonstrated that hyperspectral imaging approaches which scan the fluorescence excitation spectrum can provide increased signal strength and faster imaging, compared to traditional emission-scanning approaches. We have also demonstrated that excitation-scanning approaches may be able to detect spectral differences between colonic adenomas and adenocarcinomas and normal mucosa in flash-frozen tissues. Here, we report feasibility results from using excitation-scanning hyperspectral imaging to screen pairs of fresh tumoral and nontumoral colorectal tissues. Tissues were imaged using a novel hyperspectral imaging fluorescence excitation scanning microscope, sampling a wavelength range of 360-550 nm, at 5 nm increments. Image data were corrected to achieve a NIST-traceable flat spectral response. Image data were then analyzed using a range of supervised and unsupervised classification approaches within ENVI software (Harris Geospatial Solutions). Supervised classification resulted in >99% accuracy for single-patient image data, but only 64% accuracy for multi-patient classification (n=9 to date), with the drop in accuracy due to increased false-positive detection rates. Hence, initial data indicate that this approach may be a viable detection approach, but that larger patient sample sizes need to be evaluated and the effects of inter-patient variability studied. 
    more » « less
  4. Abstract

    DNA‐templated silver nanoclusters (AgNC@DNA) are a novel type of nanomaterial with advantageous optical properties. Only a few atoms in size, the fluorescence of nanoclusters can be tuned using DNA overhangs. In this study, we explored the properties of AgNCs manufactured on a short single‐stranded (dC)12when adjacent G‐rich sequences (dGN, withN = 3–15) were added. The ‘red’ emission of AgNC@dC12with λMAX = 660 nm dramatically changed upon the addition of a G‐rich overhang with NG = 15. The pattern of the emission–excitation matrix (EEM) suggested the emergence of two new emissive states at λMAX = 575 nm and λMAX = 710 nm. The appearance of these peaks provides an effective way to design biosensors capable of detecting specific nucleic acid sequences with low fluorescence backgrounds. We used this property to construct an NA‐based switch that brings AgNC and the G overhang near one another, turning ‘ON’ the new fluorescence peaks only when a specific miRNA sequence is present. Next, we tested this detection switch on miR‐371, which is overexpressed in prostate cancer. The results presented provide evidence that this novel fluorescent switch is both sensitive and specific with a limit of detection close to 22 picomoles of the target miR‐371 molecule.

     
    more » « less
  5. In many situations, sample data is obtained from a noisy or imperfect source. In order to address such corruptions, this paper introduces the concept of a sampling corrector. Such algorithms use structure that the distribution is purported to have, in order to allow one to make “on-the-fly” corrections to samples drawn from probability distributions. These algorithms then act as filters between the noisy data and the end user. We show connections between sampling correctors, distribution learning algorithms, and distribution property testing algorithms. We show that these connections can be utilized to expand the applicability of known distribution learning and property testing algorithms as well as to achieve improved algorithms for those tasks. As a first step, we show how to design sampling correctors using proper learning algorithms. We then focus on the question of whether algorithms for sampling correctors can be more efficient in terms of sample complexity than learning algorithms for the analogous families of distributions. When correcting monotonicity, we show that this is indeed the case when also granted query access to the cumulative distribution function. We also obtain sampling correctors for monotonicity even without this stronger type of access, provided that the distribution be originally very close to monotone (namely, at a distance $O(1/\log^2 n)$). In addition to that, we consider a restricted error model that aims at capturing “missing data” corruptions. In this model, we show that distributions that are close to monotone have sampling correctors that are significantly more efficient than achievable by the learning approach. We consider the question of whether an additional source of independent random bits is required by sampling correctors to implement the correction process. We show that for correcting close-to-uniform distributions and close-to-monotone distributions, no additional source of random bits is required, as the samples from the input source itself can be used to produce this randomness. 
    more » « less