Two-photon microscopy through scattering media harnessing speckle autocorrelation Journal: opticajournal Publication type: Research Article

Lei Zhu, 1 Bernhard Rauer, 1 Hilton B. de Aguiar, 1 and Sylvain Gigan 1,* · 2025

Read Two-photon microscopy through scattering media harnessing speckle autocorrelation Journal: opticajournal Publication type: Research Article by Lei Zhu, 1 Bernhard Rauer, 1 Hilton B. de Aguiar, 1 and Sylvain Gigan 1,* (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-13747.txt.

Introduction

Lei Zhu and three co-authors found in 2025 a way to run two-photon microscopy through scattering layers that normally blur it away. The method leans on the optical memory effect, a quirk where speckle patterns shift predictably with angle. By reading the grainy pattern's self-similarity and running a phase retrieval routine, they rebuilt pictures of glowing objects hidden behind scatterers. No wavefront shaping hardware was needed.

Two-photon imaging is prized for deep tissue work, so peering through scatter matters for biology. The authors report clear reconstructions where standard imaging fails. Their approach trades heavy optics for computation after capture. They see it extending the depth reach of a workhorse microscope. The study appears in an optics journal as a research article. Further tests in living tissue will show how far the trick carries.

Catalog record

CREATORLei Zhu, 1 Bernhard Rauer, 1 Hilton B. de Aguiar, 1 and Sylvain Gigan 1,*
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE17 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-13747

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Two-photon microscopy through scattering media harnessing speckle autocorrelation Journal: opticajournal Publication type: Research Article Lei Zhu, 1 Bernhard Rauer, 1 Hilton B. de Aguiar, 1 and Sylvain Gigan 1,* Abstract Two-photon (2P) microscopy is a powerful technique for deep-tissue fluorescence imaging; however, tissue scattering limits its effectiveness for depth imaging using conventional approaches. Despite typical strategies having been put forward to extend depth imaging capabilities based on wave-front shaping (WFS), computationally recovering images remains a significant challenge using 2P signal. In this work, we demonstrate the successful reconstruction of fluorescent objects behind scattering layers using 2P microscopy, utilizing the optical memory effect (ME) along with the speckle autocorrelation technique and a phase retrieval algorithm. Our results highlight the effectiveness of this method, offering significant potential for improving depth imaging capabilities in 2P microscopy through scattering media. 1 Introduction Fluorescence-based microscopy techniques have evolved from a custom tool to a broadly available imaging modality in the life sciences. However, the transport of light through an inhomogeneous media, such as biological tissues, leads to the scattering phenomenon [ 1 , 2 ] . In an imaging system, this scattering effect causes photons to deviate from the focal point when passing through a scattering medium. This effect significantly reduces the energy at the focal plane and introduces aberrations into the imaging system. Modern techniques, such as optical coherence tomography (OCT) [ 3 , 4 ] , actively reject scattered light while preserving ballistic light but are limited to thin or semi-transparent samples due to exponential signal loss with depth. Another approach is the WFS technique, which compensates for scattering effects by focusing the scattered light to a diffraction-limited point [ 5 , 6 , 7 ] . This is accomplished by optimizing the wavefront with a spatial light modulator (SLM), using an appropriate feedback signal from the system. However, this technique requires access to both sides of the scattering medium [ 5 ] or the presence of a guide star (or a known object) [ 6 , 7 , 8 ] . Measuring the optical transmission matrix (TM) [ 9 ] enhances the ability to manipulate the output optical field by establishing a relationship between the input and output of the medium. Nevertheless, in practical biological applications, the use of the TM is limited by time-consuming acquisition and physical access constraints. Recently, WFS combined with computational optimization methods was shown to enable fluorescence imaging through opaque samples. This has been achieved either by computationally retrieving the incoming and outgoing TMs [ 10 , 11 , 12 ] , or by optimizing the different feedback signals to apply a physical correction to the excitation or detection path [ 13 , 14 ] . Alternatively, retrieving either the incoming or outgoing TM allows focusing and imaging behind a complex medium [ 15 , 16 ] . In non-linear optical fluorescence imaging scenarios, computational methods [ 17 , 18 , 19 ] have been proposed to address this challenge. These methods [ 17 , 18 , 19 ] effectively achieve focusing through scattering samples, followed by raster scanning to image within the range of optical ME [ 1 , 2 ] . However, refocusing through opaque samples using a spatial light modulator (SLM) remains relatively slow, constrained by both the computational speed of the algorithm and the modulation rate of the SLM itself. Due to the presence of the optical ME, computational recovery of hidden objects behind complex samples has become widely used in fluorescence-based image reconstruction [ 20 , 21 ] . Here, we present a method for non-invasive imaging of hidden objects in a 2P fluorescence scenario using unknown random illuminations, without the need for a SLM. Leveraging the optical ME, the illumination laser pattern remains highly correlated during raster scanning, allowing the captured 2P fluorescence image to be expressed as the convolution of the scanning pattern and the hidden object. By applying an autocorrelation technique, the Fourier amplitude of the hidden object is preserved, enabling image reconstruction through an iterative phase retrieval algorithm. However, in 2P scenarios, the limited number of effective speckle grains and illumination power makes a single illumination realization insufficient for autocorrelation-based recovery. To overcome this limitation, multiple random speckle illumination realizations are used, ensuring a high signal-to-noise ratio (SNR) in the autocorrelation process. Finally, we experimentally demonstrate the applicability of our technique for imaging a hidden object through a random scattering medium in 2P microscopy. 2 Principle In laser scanning microscopy, an image, I , is acquired by raster scanning a focused beam across the sample, O . The spatial profile of the laser beam, S , act as the point spread function (PSF) of the imaging system. The recorded image in 1-photon, I_{1P} , can be expressed as: I_{1P}(\theta)=\int{S(r)O(r-\theta d)d^{2}r}=[S\ast O](\theta) (1) where \ast refers to the convolution operator, \theta denotes the scanning angle, d indicates the focus length of the imaging system, and r represents the spatial coordinate. Figure 1: Imgaing through scattering media using 2P signal based on optical ME: concept and numerical example. a , The laser beam passes through a rotating diffuser (RD), which randomly modulates the beam. The modulated beam is then delivered through the galvo mirror (GM) unit, which performs raster scanning, before reaching the back focal plane of the objective (OB). Then, the hidden object behind the scattering medium (SM) is excited. Upon excitation, the fluorescence passes through a second OB and is collected by a photomultiplier tube (PMT). b , A stack of speckle images is recorded by the PMT. c , The autocorrelation of each speckle. d , The averaged autocorrelation from (c). e , The final reconstruction is retrieved using an iterative phase-retrieval algorithm. SL: scanning lens, O: object, DM: dichroic mirror, TL: tube lens, F: filter, Cam.: camera, and Recon.: reconstruction. An ideal imaging system has a PSF that is a diffraction-limited Airy function, with its energy concentrated in the central lobe. On the contrary, a scattering media will scramble the phase, and distort the diffraction-limited spot, creating a random speckle pattern. Despite being spatially random, this speckle pattern still preserves high correlation while shifting within the optical ME range. Thus, the speckle pattern can still function as a PSF, and the convolution model remains valid in the speckle scenario within the optical ME range. However, 2P microscopy owes its advantages of nonlinear excitation, where the PSF changes to |S|^{2} , which leads to the suppression of weak speckle grains and to an enhanced contrast. In the 2P scattering imaging system, the captured 2P image, I_{2P} , can be denoted as I_{2P}(\theta)=\int{|S(r)|^{2}O(r-\theta d)d^{2}r}=[|S|^{2}\ast O](\theta) (2) Taking the autocorrelation of I_{2P} and using the convolution theorem yields I_{2P}\star I_{2P}=[|S|^{2}\ast O]\star[|S|^{2}\ast O]=[O\star O]\ast[|S|^{2}\star|S|^{2}] (3) where \star is the autocorrelation operator. Within the ME range, the autocorrelation of |S|^{2} is a sharply peaked function and its width is equal to the average speckle grain size. Then, the autocorrelation of the hidden object can be retrieved from the autocorrelation of the I_{2P} : O\star O\approx(I_{2P}\star I_{2P})+C (4) C is an additional constant background term. The autocorrelation of the hidden object, located behind various scattering medium, can be retrieved. Importantly, compared with 1P fluorescence experiments, due to the 2P effect, I_{2P} doesn’t

Read the complete text