AI Microscope Eliminates Sample Transfer to Protect Sensitive Biological Context in Mechanical Systems

Technical Summary ⚙️

The new MOSAIC microscope was developed at Lawrence Berkeley National Laboratory to overcome the challenges of traditional microscopy in studying sensitive biological samples. This system integrates more than ten imaging techniques into a single reconfigurable device, using adaptive optics to correct distortion caused by living tissues. MOSAIC provides massive processing capabilities of up to 4 terabytes of data per hour, supported by high-performance computing tools, opening new horizons for monitoring complex biological processes with high temporal and spatial precision and contributing to faster scientific discoveries through the integration of artificial intelligence.

🔧 Introduction: Traditional Challenges in Biological Microscopic Imaging

Studying complex biological systems is one of the biggest challenges in biomechanical engineering and imaging. Molecular interactions occur on the nanometer scale and within milliseconds, while changes associated with diseases such as Alzheimer’s expand over large ranges reaching millimeters. The challenge lies in observing these diverse phenomena at the same time in the same sample.

Traditionally, samples had to be transferred between different specialized imaging devices, which leads to damage to the biological context and the loss of important data, in addition to slowing research progress. Moreover, the higher the microscope magnification on living tissues, the greater the degradation in image clarity and the greater the complexity of analyzing the resulting data.

🔥 What Is MOSAIC and How Does It Work?

MOSAIC (Multimodal Optical Scope with Adaptive Imaging Correction) is a unique technical achievement and represents a leap in the field of compound microscopes. The project emerged after the development of a light microscope based on a light-wave system and adaptive-optical lattice light-sheet microscope technology in 2018.

MOSAIC was redesigned to bring together more than ten different imaging methods in a single compact device that can be rapidly reconfigured in a time ranging between two and five seconds, while reducing the device footprint compared with previous designs that occupied a large space reaching 3×1.2 meters.

⚙️ Core Technical Innovations in MOSAIC:

• An intelligent optical system based on programmed light switching using devices such as lasers, mirrors, and cameras, to provide multiple imaging modes using the same hardware.
• Adaptive optics used to enable clear viewing through living tissue, similar to the technology that enhances viewing of stars through the atmosphere.
• The use of fluorescent molecules that allow the molecular and cellular structures in living tissue to be distinguished effectively.
• The ability to transfer and process massive data reaching 4 terabytes per hour, supported by a high-performance computing environment that may include use of the Perlmutter supercomputer at NERSC.

🚗 How Did MOSAIC Change the Understanding of Biological Tissues?

The most notable feature of MOSAIC is the monitoring of cellular dynamics and complex structures inside living organisms without the need to destroy the tissues or remove them from their natural context. During multiple experiments, the device was able to track single molecules in living cells, observe organelle movements in fish embryos, and map neural structures in an enlarged human brain tissue associated with Alzheimer’s disease.

What is striking in live neural imaging experiments using MOSAIC is that adaptive optical corrections revealed a 2.5-fold increase in the detection of neural calcium activity compared with traditional imaging, indicating that previous microscope devices underestimated the true size of neural activity.

✍️ Case Study: Volumetric Imaging via Photochemical Sectioning (VIPS)

MOSAIC technology was used in the VIPS study published in 2025, where the team managed to image the entire olfactory peduncle in mice, producing about a terabyte of data at nanometer resolution over two weeks, with analysis that took two years. This highlights the vast difference between the ability to collect data and the ability to process and understand that data so far.

🏭 Big Data and Artificial Intelligence: The Future of Biological Microscopes

To overcome the barrier of processing enormous data volumes, the Berkeley team developed the open-source PetaKit5D program to handle the data streaming from MOSAIC quickly and efficiently, reducing processing costs by more than tenfold.

By leveraging Perlmutter supercomputer computing, two MOSAIC devices are currently operating at the University of California, Berkeley around the clock, collecting five-dimensional data that include the three spatial coordinates, time, and molecular identity. This data is used to train advanced artificial intelligence models to map and interpret complex biological processes.

📈 Future Capabilities:

• A visual language model that directly handles biological data, linking molecular information, experimental data, and prior knowledge.
• Integration of automation processes into laboratory workflows, from microscopy and sample processing to biological interference systems.
• The possibility of developing self-operating biological laboratories, accelerating discovery and innovation in biomechanical engineering and medical industries.

🔥 Conclusion: The Impact of MOSAIC on the Field of Biomechanical Engineering

The MOSAIC microscope represents a qualitative leap in the field of advanced mechanical systems used in biological imaging, as it combines multimodal imaging, adaptive optical correction technologies, and high-performance computing. This combination enables the study of complex vital processes in precise depth, while preserving sample integrity, and opening the door to accelerating research and discovering new treatments.

Thanks to the partnership between mechanical engineering, applied physics, and computer science, researchers now have an advanced tool that may change the concept of biological research in an independent and sustainable way, translating technological breakthroughs into tangible industrial and scientific achievements.

📚 Related articles

Servomex Enhances Reliability in Mechanical Systems with SERVOTOUGH SpectraExact 2500F

2026/09/08

⚡ Understanding the Concept of Interconnection for Large Loads

2026/09/08

Shapiro Sets Binding GRID Requirements in Pennsylvania with Focus on Data Center Energy Costs

2026/09/08