Integration of adaptive optics on a lattice light-sheet microscope for super-resolution imaging in depth

The lattice light-sheet microscopy (LLSM) allows fluorescence imaging inside thick samples down to about 30µm, with very high spatial and temporal resolution and very low phototoxicity. This microscope, initially designed by E. Betzig (Nobel 2014) and his team, was successfully developed and modified at Bordeaux Imaging Center (BIC).

To achieve nanometric resolutions within thick samples, we decided to integrate single molecule localization microscopy (SMLM) methods, and adaptive optics (AO) into the LLSM. The SMLM bypass the phenomenon of light diffraction, which naturally limits the spatial resolution.  The AO correct in real time optical aberrations that distort the image and cause a decrease in spatial resolution. The integration of AO and SMLM techniques on LLSM is the subject of the CIFRE thesis of Maxime Malivert funded by Imagine Optic and supervised by Mathieu Ducros. The first step, which consists in combining AO correction with LLSM in classic resolution, was accomplished over last months. The detection path of the microscope was modified to integrate a deformable mirror (DM) and a wavefront sensor (WFS). A software, using 3N optimization, has also been developed to suppress aberrations defined with Zernike functions. This sensorless optimization, applied on sample’s images, is appropriate to optimize rapidly the wavefront without the need for point sources inside the sample.

We study the influences of the metric, the range, the number of modes chosen and the structure of the sample to determinate the best parameters for image optimization. In parallel, we determine the weight of each aberration mode in the deformation of the wavefront with an indirect measurement.

Figure1: Images of a neuron, with a detail on a single spine at approximatively 40 µm under the surface of the sample, imaged with LLSM, before and after 3N optimization with AO. We demonstrate the incidence of each Zernike mode as a function of depth with an indirect method, and the resolution gain in depth with 3N optimization by the FWHM measure of dendritic spine.