Read on Twitter
We are excited to share our preprint on uniaxial permittivity tensor imaging (uPTI) that measures density and anisotropy of transparent organelles, cells, tissues, and materials with confocal-like resolution.
https://www.biorxiv.org/content/10.1101/2020.12.15.422951v1
@LiHao_Yeh et al.
(1/n)
https://www.biorxiv.org/content/10.1101/2020.12.15.422951v1
@LiHao_Yeh et al.
(1/n)
@LiHao_Yeh has been an amazing partner in bringing this idea to fruition. My team @czbiohub ( @ieivanov1 , @symguo , @BryantChhun , @janie00) and our collaborators @GladstoneInst ( @Juan_PerBer, Coklin lab) and @Stanford (Ezzat, Han lab) provided many key inputs. Thank you!! (2/n)
Label-free measurements with phase and polarization are becoming mainstream, because they a) report physical properties that complement molecular properties accessible with fluorecent markers, and b) can reveal architecture of specimens that are hard to label (3/n)
Density (optical path length) and anisotropy (retardance) are isotropic and anisotropic components of the permittivity tensor. Existing methods typically measure one or the other, and polarization methods don’t account for diffraction, limiting accuracy and resolution (4/n)
uPTI makes invisible components of permittivity tensor visible using illumination-diversity, polarization-diversity, and focus-diversity. The physical properties are decoded with a new vector diffraction model and multi-channel convex optimization. (5/n)
Measurements of phase, principal retardance, 3D orientation, and optic sign with uPTI provide a fascinating view in tissue and cell architecture at multiple scales (cm-250nm). Checkout this multi-scale fly-through around the corpus callosum in the mouse brain tissue. (6/n)
The color in the above movie is the 3D orientation of axon bundles mapped to the top hemisphere. The flat color legend in above movie is rendered as hemispehre here. High-res imaging of 3D orientation across volume has been an outstanding gap in microscopy that uPTI fills (7/n).
Our optical design and vector diffraction model allow use of high-NA oil immersion lenses that provide confocal-like resolution - as in these XY, XZ, YZ fly-throughs of a volume of a mouse brain section. We resolve cross-sections of axons in the network! (8/n)
uPTI is a module that is easily added to a widefield fluorescence microscope. Multiplexed imaging of iPSC-derived cardiomyoctes with uPTI/fluor deconvolution shows their rich architecture- myofibrils, bands of sarcomere, troponin bands, nuclei, membranous organelles, … (9/n)
The resolution and accuracy allow us to characterize anisotropic glass (5D crystal) written with femtosecond laser from Kazansky lab @orctweets. We can image nature of laser-induced defects in 3D. The technology is promising for metrology too. (10/n)
Why the name? Permittivity tensor imaging is a spatio-angular imaging approach analogous to diffusion tensor imaging, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2041910/). DTI measures angular distribution of water diffusion, PTI measures the angular distributions of electrons bound to dielectrics. (11/n)
Our modeling and reconstruction approach is analogous to optical diffraction tomography, which represents the specimen as scalar scattering potential. We extend the idea to distribution of a scattering potential tensor. (12/n)
Assuming uniaxial symmetry, single scattering, and weak absorption allows linearization of the forward model, which can be inverted with multi-channel deconvolution. These OTFs relate recorded volumes (rows) and scatterinng potential tensor (columns) in Fourier space. (13/n)
@LiHao_Yeh and I also understood the effect of "missing cone of orientations” in angular dimensions during an intense brainstorm. The effect is analogous to the “missing cone of spatial frequencies” in spatial dimensions in single-view fluorescence microscopes. (14/n)
Python code for wave optical simulations and reconstructions is available at https://github.com/mehta-lab/waveorder. The repo is useful also for measurement of subsets of uPTI measurements, e.g., QLIPP ( https://elifesciences.org/articles/55502 ) and 3D DPC( https://doi.org/10.1364/BOE.7.003940). (15/n)
We are eager to collaborate on 3D rendering of this rich data to make it accessible and useful. Any @napari_imaging aficionados want to collaborate on rendering this multi-channel vectorial volumes ( https://twitter.com/mattersOfLight/status/1340082680684146688?s=20) ? (16/n)
We have exciting improvements to the technology and collaborations in biology on deck. If you have read this far and read the preprint, do share your critique and questions with @LiHao_Yeh and me. We are keen to see rapid refinement and adoption. (end)
