Background

Quantitative biological imaging requires Nyquist sampling in space and time. However, due to the limited acquisition rate of laser scanning and camera-based microscopes, many biological processes occur too rapidly to be observed in 3D. This is partly because volumes are usually acquired sequentially with tens-to-hundreds of focal planes. However, imaging rates could be several orders of magnitude faster if information spanning multiple focal planes could be integrated into a single raster scan or camera exposure. While multiple projection imaging methods exist, they have common hindrances, including that they typically deteriorate lateral resolution; that they require specialized setups, and that they provide only projections along the optical axis. In their work, Chang et al (1) introduce a cost-effective and easy-to-implement scan unit which enables any camera-based microscope to perform projection imaging from diverse angles. By imaging the sample from one or multiple perspectives, their novel method enables visualization of rapid biological processes, real-time stereoscopic imaging and 3D particle localization throughout a cellular volume from just two images.

 

Key findings and developments

In their work, Chang et al introduce a simple scan unit that converts any camera-based microscope into a projection imaging system which can integrate information from various viewing angles. Instead of requiring sample rotation, the method uses optical shearing to provide various viewing directions of the sample. Moreover, the method can be combined with Lattice Light-sheet Microscopy (LLSM) and Oblique Plane Microscopy (OPM).  Importantly, while both of these imaging methods require computational shearing of the image data for visualization and quantification, the projection method developed by Chang and co-authors performs optical shearing during the acquisition of a single camera frame, resulting in a more intuitive instantaneous 3D rendering of the sample. The method involves the development of a simple galvanometer-based module that attaches to the microscope camera and shears the data optically instead of computationally. The scan unit is swept synchronously with the acquisition of a z-stack, resulting in the projection of high-contrast, high-resolution volumetric data onto a single camera frame. The magnitude of the scan sweep can be changed in order to obtain projections from different viewing perspectives in a way that is analogous to a shear-warp transform.

As proof of principle, the authors imaged mammalian cells using LLSM-Pro (i.e. LLSM combined with the projection mode here developed), and OPM-Pro. In both cases, the authors were able to switch between projection and volumetric modes. Computationally projected data acquired with a LLSM was indistinguishable from data acquired with a LLSM-Pro, with the difference that LLSM-Pro acquired the data in a single image frame, and thus reduced the imaging time and data overhead significantly. The same was true using OPM-Pro for both simple and complex samples. Another feature explored, was the speed of image acquisition. The authors were able to image rapid cellular dynamics such as the formation and retraction of pressure-based blebs with LLSM-Pro or calcium waves in cultured neurons with OPM-Pro. They note that in LLSM-Pro, the imaging speed is limited by the speed of the piezo stage where the sample is mounted on. In OPM-Pro, which uses much faster laser scanning optics than LLSM, the imaging speed is limited only by the camera frame rate or the available fluorescence signal. However, the authors also state that because fluorophores are only illuminated for a fraction of the camera exposure, the sensitivity of the method decreases as the imaged volume grows larger.

Besides the acquisition speed, the authors note that the possibility of acquiring data from multiple perspectives opens the door to a number of exciting opportunities, including rendering images stereoscopically. Equally, images acquired from different projection perspectives encode volumetric information otherwise difficult to acquire in a timely manner, like the distribution of nanoparticles throughout a 3D volume. Also, by adding a second camera to their microscope, they were able to simultaneously visualize the rapidly beating heart of a 3-day post-fertilization zebrafish embryo from orthogonal imaging perspectives with high spatial resolution and no apparent distortions. The authors further emphasize that unlike other methods, their method does not interfere with normal microscope operation. They also point out that their projection technique can cover large volumes, and this is only limited by the axial scan range while maintaining an invariant PSF. Lastly, because LLSM and OPM require computational shearing (a.k.a. deskewing) of their raw data before it can be intuitively visualized, exploring biological samples can be particularly challenging.  In contrast, the projection method developed here allows viewing of deskewed and projected volumes in real-time and as such greatly simplifies user interaction with such microscopes.

The authors conclude that their novel method exploits the shear-warp transform, which obtains rotated projections at intermediate angles without physically rotating the sample or imaging it through multiple detection objectives. Moreover, they explore multiple applications where this exciting approach would be very advantageous and propose that this tool will allow the transformation of cutting-edge imaging technologies into user-friendly and interactive data-acquisition and hypothesis-testing machines.

 

What I like about this preprint

 I like novel technologies that address methodological gaps. I think addressing those gaps will be key for the advancement of scientific exploration. Having consistently used intravital microscopy, and the hurdles that in this tool we must overcome, it seems this tool hereby proposed will be of great use for many scientific fields requiring fast imaging and/or volumetric imaging.

 

References

1. Chang et al, Real-time multi-angle projection imaging of biological dynamics, bioRxiv, 2020

 

doi: https://doi.org/10.1242/prelights.26531

Read preprint (No Ratings Yet)