Ch.lab |

All-optical photoacoustic imaging for neurobiology

Mon, 01 Jan 0001 00:00:00 +0000

Light provides rich molecular contrast, but it rapidly loses direction in brain tissue. Ultrasound travels much farther with little scattering. Photoacoustic imaging combines these properties: short laser pulses generate ultrasound inside tissue, and the recorded waves reveal where light was absorbed.

Our long-term goal is to image neuronal activity non-invasively beyond one millimetre depth, with resolution and sensitivity relevant to neural circuits. The project brings together transparent ultrasound sensors, fast optical interrogation and three-dimensional reconstruction, developed as parts of a single imaging system.

Why photoacoustics can reach deeper

A nanosecond laser pulse briefly heats absorbing structures such as blood vessels or functional indicators. Their thermoelastic expansion emits broadband pressure waves. Because these waves are only weakly scattered by soft tissue, they can be measured at the surface and used to recover a three-dimensional map of optical absorption.

Photoacoustic imaging combines optical absorption contrast with acoustic propagation at depth. Temporary figure from the Institut Fresnel project page.

Our contribution: measuring sound with light

We develop transparent Fabry-Pérot ultrasound sensors made from a thin polymer layer between dielectric mirrors. An incoming pressure wave changes the layer thickness by a tiny amount and shifts its optical resonance. A focused interrogation beam converts this shift into a measurable change in reflected light.

The thin sensing layer provides broad acoustic bandwidth and high-frequency sensitivity, both of which are needed to resolve small structures. Scanning the interrogation beam across the sensor records the acoustic field over space and time; an inverse model then reconstructs the original absorber distribution.

Optical ultrasound detection and raster-scanned acquisition. Temporary figure from the Institut Fresnel project page.

Current technical work

Small thickness variations across a fabricated sensor shift its resonance from point to point. A conventional narrowband system must therefore retune its wavelength during a scan, which limits acquisition speed. We are developing several ways around this limitation: broadband illumination with fast wavelength selection, dual-comb optical interrogation that samples many wavelengths at once, and acquisition strategies that require fewer spatial measurements.

Compensating resonance inhomogeneity recovers image quality across the sensor. Temporary figure from the Institut Fresnel project page.

Fast three-dimensional reconstruction

Accurate model-based reconstruction can include detector physics, acquisition geometry, noise and prior knowledge, but repeatedly solving the full acoustic wave equation is expensive. We are developing fast forward and adjoint operators tailored to planar detection so that iterative three-dimensional reconstruction becomes practical.

From instrumentation to neurobiology

The instrumentation is being developed toward functional imaging of neuronal and vascular signals. This requires connecting sensor performance, illumination, reconstruction and biological validation rather than optimizing any one component in isolation. We work with neurobiology collaborators at INMED and INT on relevant questions and preparations.

An experimental system and an example vascular reconstruction. Temporary figure from the Institut Fresnel project page.

Working with us

This programme sits at the interface of optical instrumentation, thin-film sensing, inverse problems and neurobiology. Collaborations can contribute a strong biological question, a new functional contrast mechanism or complementary expertise in imaging and reconstruction. Students and postdoctoral researchers may work across experiment, sensor development and quantitative image formation; contact us to discuss a possible fit.

Selected sources

Institut Fresnel: All-optical photoacoustic imaging. J. Saucourt et al., Fast interrogation wavelength tuning for all-optical photoacoustic imaging, Optics Express 31, 11164–11177 (2023). E. Küçükkomürcü et al., A depth-dependent, transverse shift-invariant operator for fast iterative 3D photoacoustic tomography in planar geometry (2026 preprint).

Dual-comb optical ultrasound sensing

Mon, 01 Jan 0001 00:00:00 +0000

A Fabry-Pérot ultrasound sensor is extremely sensitive, but its resonant wavelength changes slightly across its surface. Reading a large sensor with one narrowband laser therefore requires repeated wavelength tuning. We use two optical frequency combs instead, so that many wavelengths can interrogate the sensor at the same time.

In our first demonstration, three optical channels recorded a 10 MHz acoustic wave in a single shot with 20 ns temporal resolution. We are now developing this principle into a spatially resolved readout for faster, high-resolution all-optical photoacoustic imaging.

Why two frequency combs?

A frequency comb is a set of evenly spaced, mutually coherent optical frequencies. When one or more comb lines lie on the slope of a Fabry-Pérot resonance, a pressure-induced resonance shift modulates their reflected intensity.

A second comb with a slightly different spacing converts the optical signals into distinguishable beat notes at radio frequencies. Each useful pair of comb lines then acts as a separate measurement channel. The system can retain the fast response required for ultrasound while covering a much broader optical range than a single narrowband probe.

Comb-line pairs that overlap a resonance slope carry the acoustic modulation. Diagram from the project presentation dofc.pptx.

What has been demonstrated

Our electro-optic combs are generated from two narrow-linewidth lasers using intensity and phase modulators. In the first experiment, their repetition rates were 10 GHz and 9.95 GHz. Approximately thirty lines covered a 300 GHz optical span, chosen to accommodate realistic resonance variations across a thin polymer sensor.

The combined light interrogated a Fabry-Pérot cavity while a transducer generated a 10 MHz acoustic burst. A balanced photodetector recorded the dual-comb interferogram, whose radio-frequency components were separated by time-frequency analysis.

Experimental arrangement used for the proof-of-principle measurement. Diagram from dofc.pptx, based on Chatterjee et al. (2025).

This experiment provided a single-shot measurement of a 10 MHz acoustic wave through three simultaneous optical channels, with a temporal resolution of 20 ns. The recovered waveform agreed with a conventional continuous-wave reference measurement. It establishes that sensor resonances separated in wavelength can be monitored together, without retuning the probe between measurements.

The optical interferogram is mapped to radio-frequency channels, from which the acoustic waveform is recovered. Results from dofc.pptx, based on Chatterjee et al. (2025).

Current direction

The present result demonstrates the spectro-temporal principle at one position on the sensor. We are now working to improve sensitivity, refine the balance between comb spacing and acoustic bandwidth, and integrate the method into spatially resolved photoacoustic acquisition. Parallel or compressed interrogation could ultimately reduce the time needed to map a complete ultrasound field.

Working with us

The project connects electro-optic frequency-comb generation, low-noise optical detection, Fabry-Pérot sensor physics and fast acquisition. It is suited to collaborators and prospective researchers interested in taking a precise spectroscopic result toward a complete imaging instrument. Contact us if your expertise or application could help test that transition.

Reference

D. Chatterjee et al., Real-time electro-optic dual comb detection of ultrasound waves, Journal of Physics B: Atomic, Molecular and Optical Physics 58, 153501 (2025).

Imaging neural circuits in Danionella

Mon, 01 Jan 0001 00:00:00 +0000

Understanding a neural circuit requires observing its activity while the animal senses and responds to its environment. Danionella offers an unusual opportunity to do this in an adult vertebrate: the fish is only about 12 mm long, remains transparent throughout life and has no skull roof above the brain.

Unlike many transparent larval models, adult Danionella displays mature social and sensory behaviours. Its small brain and optical accessibility make it possible to study these behaviours while retaining a view of neural activity across a substantial fraction of the brain. Our question is how distributed neural activity represents the physical cues that allow the animal to localize sound.

Why Danionella

Adult female and male Danionella translucida. Image extracted from Schulze et al. (2018).

Schulze et al. described an adult brain of approximately 0.6 mm³ containing an estimated 650,000 neurons. They established genetic manipulation with CRISPR–Cas9 and Tol2 transgenesis, and demonstrated calcium imaging in transgenic fish expressing GCaMP6f. Two-photon imaging reached about 300 µm into the brain—nearly half its volume—while the animal remained capable of adult behaviours such as courtship, shoaling and acoustic communication.

The transparent head and absence of a skull roof provide direct optical access to the adult brain. Image extracted from Schulze et al. (2018).

A tractable question: directional hearing

Sound localization is particularly interesting underwater. The pressure difference between the two sides of a small fish is too weak to provide the cue used by terrestrial vertebrates, yet fish can still determine where a sound originates.

Work in Danionella cerebrum showed that pressure and particle motion are both required for directional hearing. The swim bladder and Weberian ossicles provide an indirect pressure pathway to the inner ear, while the otolith organs sense particle motion directly. Their relative phase carries directional information. This establishes a well-defined sensory computation that can now be followed from the mechanics of the ear toward its neural representation.

Research direction

For our group, Danionella provides a compact biological system in which new optical and photoacoustic methods can be developed against identifiable sensory signals and behaviour. Our aim is to connect measurements of brain activity with the physical sound cues available to the animal, and to determine how distributed neural circuits represent and use directional information.

Working with us

This project brings together optical imaging, underwater acoustics, behaviour and systems neuroscience. It is particularly relevant to collaborators with Danionella genetics or neurobiology expertise, and to prospective researchers who want to connect instrument development with a precise biological question. Contact us to discuss complementary approaches or a possible project.

References

L. Schulze et al., Transparent Danionella translucida as a genetically tractable vertebrate brain model, Nature Methods 15, 977–983 (2018). J. Veith, T. Chaigne et al., The mechanism for directional hearing in fish, Nature (2024).

Optical pressure mapping in microfluidic channels

Mon, 01 Jan 0001 00:00:00 +0000

Pressure is central to microfluidic operation, but measuring it locally often requires an embedded sensor that complicates fabrication or changes the device itself. Our approach uses the channel as its own sensing element: pressure deforms the transparent structure, and quantitative phase imaging measures this deformation without contact.

The first experiments resolved relative changes in channel radius below 0.1 µm, corresponding to an estimated pressure sensitivity of about 5 mbar in the deliberately compliant test device. The next step is to extend this calibrated measurement to more realistic microfluidic networks.

Measurement principle

Light accumulates a different optical delay when it passes through water and through polydimethylsiloxane (PDMS). Internal pressure expands a compliant channel, changing both its geometry and the optical path through the device. Quadrilateral lateral shearing interferometry records the resulting wavefront distortion as an optical-path-difference map.

Pressure changes the channel cross-section and therefore the transmitted optical wavefront. Measurement principle from Acharya et al. (2026).

For a channel with known geometry and material properties, transverse phase profiles can be fitted with a mechanical model. The fit provides the channel radius and refractive-index contrast, from which pressure can be estimated after calibration.

What has been demonstrated

The phase images below show the same compliant channel with no applied pressure and at 1 bar. The wider optical-path profile at high pressure directly reveals the expansion of the channel.

Optical-path-difference map at 0 bar, extracted from Acharya et al. (2026).

Optical-path-difference map at 1 bar, extracted from Acharya et al. (2026).

In the proof-of-principle device, pressure ramps from 0 to 1000 mbar produced model-consistent changes in channel shape. The deliberately compliant PDMS channel reached more than 40% strain at the highest pressure, while relative variations in radius were resolved below 0.1 µm. Under those experimental conditions this corresponded to an estimated relative pressure sensitivity of about 5 mbar. The achromatic phase method also worked with ordinary white-light microscope illumination, and time-resolved measurements revealed the asymmetric response to positive and negative pressure steps.

These values describe the test device rather than a universal sensor specification. The response depends on channel geometry, material properties, imaging conditions and calibration.

Where the method could be useful

A non-contact, image-based measurement could map pressure at several positions in a soft microfluidic network without integrating a dedicated sensor at every point. This is relevant to lab-on-chip systems, organ- and tissue-on-chip experiments, nonlinear microfluidic flows and the characterization of soft transparent materials.

Limitations and current work

Quantitative pressure estimates depend strongly on mechanical calibration. PDMS stiffens with age, and fabrication variability, bonding, alignment and channel geometry all influence the deformation response. We are therefore working toward wider-field measurements, more complex channel networks and calibration strategies that separate pressure changes from changes in the opto-mechanical properties of the chip.

Working with us

Useful collaborations would bring demanding microfluidic applications, independent pressure measurements or expertise in soft-material mechanics. The project also offers student and postdoctoral work combining quantitative phase imaging, modelling and device experiments. Contact us if you have a system in which local pressure is important but difficult to measure.

Reference

K. Acharya, S. Monneret, M. Brandenbourger and T. Chaigne, Real time monitoring of pressure-induced deformation of PDMS to evaluate pressure distribution in microfluidic channels (2026 preprint).

Photoacoustic-guided wavefront shaping

Mon, 01 Jan 0001 00:00:00 +0000

Coherent light becomes a granular speckle pattern after propagating through a strongly scattering material. Although the original beam appears to have lost its direction, the propagation remains deterministic over a short time. By controlling the phase of the incident light, it is therefore possible to make the scattered waves interfere constructively at a chosen position.

This principle is well established when the target can be observed with a camera. Focusing inside biological tissue is harder because the target is hidden and an internal source of feedback is required. We investigate whether a photoacoustic signal—generated only where light is absorbed—can provide that guide from within the tissue.

Photoacoustics as an internal guide star

Photoacoustics provides feedback from absorbing structures below the surface. A spatial light modulator changes the incident optical wavefront, the absorber converts the local optical energy into an ultrasonic signal, and an ultrasound detector measures the result. The phase pattern can then be optimized to increase the signal from a selected acoustic region.

A spatial light modulator controls the incident field, while the photoacoustic signal provides feedback from inside the scattering medium. Adapted from Xia et al. (2024).

An alternative is to measure a photoacoustic transmission matrix. In this approach, the elements of a photoacoustic image take the place of camera pixels and record how a set of input optical modes influences absorbing targets. Once measured, the matrix can be used to calculate the wavefront that concentrates light on a selected target.

What has been demonstrated

Early experiments demonstrated this approach through scattering samples, including 0.5 mm of chicken tissue, and extended it to two-dimensional photoacoustic imaging with a linear ultrasound array. These results established non-invasive acoustic feedback as a practical route to optical control behind a scattering layer.

Why deep tissue remains difficult

The remaining difficulty is the large difference between optical and acoustic length scales. A single acoustically resolved volume may contain thousands of independent optical speckles. Increasing the total photoacoustic signal from that volume does not automatically produce one sharply focused optical spot, and the useful modulation can be very small compared with the background signal.

Deep biological tissue also decorrelates as it moves, so the wavefront must be measured and updated before the scattering pattern changes. Detector sensitivity, laser stability, dynamic range and the speed and number of modes of the spatial light modulator all become limiting factors. As reviewed by Xia et al. (2024), concentrating light deep inside living tissue has therefore not yet been demonstrated in the same way as focusing through static scattering layers.

Current direction

We study how photoacoustic feedback can be made more selective and more rapidly acquired, and how optical modulation, acoustic detection and reconstruction should be designed together. The aim is both to approach optical focusing at depth and to develop less stringent forms of light redistribution that can improve photoacoustic excitation or fluorescence measurements in scattering tissue.

Working with us

Progress depends on combining wavefront control, high-sensitivity ultrasound detection, fast optimization and realistic biological samples. The project is well suited to collaborators, students and postdoctoral researchers interested in the boundary between optical physics, acoustics and computational imaging. Contact us if you would like to explore a shared experiment or methodological question.

References

F. Xia, I. Leite, R. Prevedel and T. Chaigne, Optical wavefront shaping in deep tissue using photoacoustic feedback, Journal of Physics: Photonics (2024). T. Chaigne, Control of scattered coherent light and photoacoustic imaging: toward light focusing in deep tissue and enhanced, sub-acoustic resolution photoacoustic imaging, doctoral thesis, Université Paris 6 (2016).