📄 bioengineering

Dual-Mode Bio-CM2: Multimodal Computational Miniature Mesoscope for Fluorescence and Label-Free Imaging

The paper presents Dual-Mode Bio-CM2, a scalable computational miniature mesoscope that utilizes a shared optical architecture to simultaneously capture co-registered fluorescence and label-free reflectance images, enabling comprehensive multimodal analysis of biological function across diverse freely moving and head-fixed specimens.

Original authors: Deng, Q., Hu, G., Rauscher, B. C., Villafuerte, M., Weinberg, B., Chen, Z., Feng, H., Devor, A., Thunemann, M., Tian, L.

Published 2026-09-14

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Deng, Q., Hu, G., Rauscher, B. C., Villafuerte, M., Weinberg, B., Chen, Z., Feng, H., Devor, A., Thunemann, M., Tian, L.

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

To understand how a living creature works, scientists often need to look at two different things at once. One thing is the specific molecular machinery inside cells, such as proteins that light up when a neuron fires. The other is the broader context: the shape of the animal's body, how it moves, or how blood flows through its tissues. For decades, researchers have struggled to capture both views simultaneously in a single, tiny microscope that can be worn by a freely moving animal. Traditional microscopes are large and heavy, and when scientists try to shrink them, they usually have to choose between seeing a wide area or seeing fine details, and they often lose the ability to see more than one type of signal at a time. This trade-off has made it difficult to connect what is happening inside a cell with what the animal is actually doing in the real world.

A team of researchers at Boston University has now built a solution to this problem. They created a new device called dual-mode Bio-CM2, a miniature microscope small enough to be mounted on a mouse, yet powerful enough to see a large area of the brain with high clarity. Unlike previous versions that could only see fluorescent tags, this new device captures two distinct types of images at the same time: one that shows specific glowing molecules and another that shows the natural structure and movement of the tissue without any labels. By combining these two views, the researchers can watch how an animal behaves while simultaneously seeing the activity of its cells, all within a single, compact instrument.

The device works by using a clever arrangement of four tiny lenses that work together to cover a large area, roughly the size of a small postage stamp, which is much wider than what a standard miniature microscope can see. Instead of using complex mirrors or beam splitters to separate different colors of light, the system takes pictures in rapid succession, switching between a blue light that makes fluorescent proteins glow and a green light that bounces off the tissue to reveal its shape. Because the switching happens so quickly, the two images are effectively captured at the same moment, allowing the researchers to line them up perfectly. This design means the microscope does not need to be bulky to get a wide view; it can remain small enough to be carried by a mouse while still providing a detailed map of the entire top surface of the brain.

The researchers tested this new microscope on three very different types of animals to prove it works across the board. First, they watched tiny roundworms swimming freely. The microscope showed the glowing clumps of protein inside the worm's muscles, which are linked to a disease model, while the second view showed the worm's entire body bending and twisting as it moved. This allowed the scientists to see exactly where the protein clumps were located in relation to the worm's posture, something that was impossible to do with a microscope that only saw the glowing spots. Next, they observed a baby zebrafish swimming in a tank. The device highlighted the heart beating with a specific glow, while the other view tracked the fish's whole body as it darted through the water. This combination let the researchers link the activity of the heart directly to the fish's swimming behavior.

Finally, the team turned to the most complex subject: the brain of a mouse. They placed the microscope on the head of a mouse that had been genetically modified so that its brain cells would glow when active. The device captured a wide view of the top of the brain, showing both the electrical activity of the neurons and the flow of blood through the vessels. This dual view was crucial because the blood vessels can sometimes interfere with the signal from the neurons, making it look like the brain is active when it is actually just the blood moving. By seeing the blood flow in the second image, the researchers could mathematically remove this interference, revealing the true activity of the neurons with much greater clarity. This ability to see both the cells and the blood flow at the same time, over such a large area of the brain, provides a much more complete picture of how the brain functions than previous tools could offer.

The success of this project demonstrates that it is possible to expand what a miniature microscope can do without making it larger or more complicated. By integrating two different ways of seeing into one shared system, the researchers have created a tool that can tell a fuller story about life in motion. Whether watching a worm navigate its world, a fish hunt for food, or a mouse process a thought, this new device allows scientists to see the molecular details and the physical behavior as a single, connected event. The work suggests that future versions of this technology could be made even smaller and lighter, potentially allowing for even more detailed studies of how the brain and body work together in freely moving animals.

Technical Summary: Dual-Mode Bio-CM2

Problem Statement
Understanding biological function often requires simultaneous acquisition of complementary information regarding molecular identity, morphology, behavior, and physiology from the same specimen. While fluorescence microscopy offers molecular specificity and label-free reflectance reveals structural and physiological context, integrating both modalities into miniature microscopes (miniscopes) for freely behaving animals remains challenging. Traditional approaches typically increase optical complexity, compromise the field of view (FOV), reduce spatial resolution, or enlarge the form factor. Existing dual-mode systems often lack the scalability to capture large tissue areas (e.g., the entire mouse dorsal cortex) with cellular resolution while maintaining a compact footprint.

Methodology
The authors present dual-mode Bio-CM2, a multimodal computational miniature mesoscope that integrates fluorescence and label-free reflectance imaging within a shared optical architecture based on distributed computational optics.

Optical Architecture: The system utilizes a $2 \times 2$ array of four miniature imaging modules that simultaneously acquire overlapping subregions of the sample. These sub-images are computationally stitched to form a continuous large-field image. This architecture partitions optical complexity across subfields rather than the entire FOV, preserving a compact form factor.
Shared Detection Path: Both modalities share the same collection optics, a single emission filter (ET525/50m, 500–550 nm passband), and a single image sensor. This eliminates the need for beam splitters, duplicated detection paths, or post-acquisition image registration, ensuring intrinsic spatial co-registration.
Illumination and Spectral Configuration:
- Fluorescence: Excited by 470 nm blue LEDs in an epi-illumination geometry.
- Reflectance: Illuminated by 530 nm green LEDs at a steep off-axis angle. This geometry directs specular reflections from planar samples outside the collection aperture (dark-field configuration) while collecting scattered light across the full FOV.
- Spectral Separation: The 530 nm reflectance wavelength lies within the fluorescence emission passband but is spectrally separated from excitation. The 530 nm wavelength was specifically chosen as it lies near the isosbestic point of oxy- and deoxyhemoglobin, making diffuse reflectance sensitive to total hemoglobin (HbT) and cerebral blood volume changes.
Acquisition Strategy: The system employs frame-interleaved acquisition. Blue and green LED pulses are synchronized with the camera's rolling shutter timing. LEDs are switched on only during the interval when all sensor rows integrate simultaneously (pseudo-global shutter) and switched off during readout. This produces alternating fluorescence and reflectance frames at 10 fps per modality (20 fps total), minimizing channel cross-talk and rolling-shutter distortion.
Image Processing: Raw data undergoes background correction, geometric distortion correction, and field-dependent deconvolution using experimentally measured point spread functions (PSFs). The sub-images are then stitched into a composite image. For mouse cortex imaging, a reflectance-based ratiometric correction is applied to the fluorescence signal to suppress hemodynamic contamination caused by hemoglobin absorption.

Key Contributions

Scalable Multimodal Framework: The paper extends the distributed computational optics framework (previously demonstrated for single-contrast fluorescence in Bio-CM2) to support simultaneous fluorescence and label-free reflectance imaging.
Performance Metrics: The system achieves a FOV of approximately $7.5 \times 10 \text{ mm}^2$ with a lateral resolution of $\sim 6 \text{ \textmu m}$ (FWHM), enabling cellular-resolution imaging across centimeter-scale specimens without mechanical scanning.
Intrinsic Co-registration: By sharing the optical path and using frame-interleaving, the system provides intrinsically co-registered molecular and structural data without the misalignment issues common in multi-camera or beam-splitter systems.

Results
The versatility of dual-mode Bio-CM2 was demonstrated across three distinct biological systems:

Freely Moving C. elegans: In transgenic worms expressing YFP-labeled polyglutamine aggregates (Huntington's disease model), the fluorescence channel visualized discrete protein aggregates, while the reflectance channel captured the complete body contour and bending posture. This allowed the localization of sparse molecular signals within the context of the animal's locomotor behavior.
Freely Swimming Larval Zebrafish: In transgenic larvae expressing EGFP in the heart and neural crest cells, fluorescence localized cardiac tissue, while reflectance captured whole-body morphology and swimming trajectories. This linked organ-specific physiology to whole-animal behavior within a single recording.
Mouse Dorsal Cortex: In head-fixed mice expressing GCaMP6s in layer 5 pyramidal neurons, the system simultaneously mapped neuronal calcium activity (fluorescence) and intrinsic hemodynamic signals (reflectance) across nearly the entire dorsal cortex.
- The reflectance signal at 530 nm provided a label-free measure of blood volume changes.
- The authors demonstrated that hemodynamic fluctuations in the raw GCaMP signal could be suppressed using pixel-wise ratiometric correction based on the simultaneous reflectance data, revealing clearer neuronal activity contrast.
- Cross-correlation analysis confirmed a $\sim 1.1$ s delay of the hemodynamic signal relative to neuronal activity, consistent with neurovascular coupling.

Significance and Claims
The paper claims that dual-mode Bio-CM2 successfully addresses the trade-off between imaging capability and form factor in miniature microscopy. By integrating complementary fluorescence and reflectance contrasts into a shared distributed computational optics architecture, the system extends the platform's utility beyond single-contrast imaging.

The authors state that this work establishes distributed computational optics as a scalable platform for multimodal miniature imaging. The system enables the interpretation of molecularly specific signals within their anatomical, behavioral, and physiological context simultaneously. While the current prototype was designed as a proof of concept (with cortical imaging performed under head-fixed conditions), the modular architecture supports future extensions, such as additional fluorescence channels or multi-wavelength reflectance for quantitative oxygenation measurements, without fundamentally altering the underlying imaging architecture.

Technical Summary: Dual-Mode Bio-CM2

More like this