RAEM: random-access electron microscopy for revisitable 3D imaging.

Area of Science:

• Neuroscience
• Cell Biology
• Microscopy

Background:

• Volume electron microscopy (EM) provides crucial 3D context for biological structures.
• Imaging large specimens at nanometer resolution is often infeasible, forcing a trade-off between anatomical breadth and ultrastructural detail.

Purpose of the Study:

• To introduce a novel framework, random-access electron microscopy (RAEM), for scalable, multi-resolution imaging of biological specimens.
• To overcome the limitations of exhaustive single-resolution EM by enabling targeted high-resolution imaging within large volumes.

Main Methods:

• RAEM involves an initial low-resolution 3D survey of the specimen.
• Knowledge derived from the survey (human or AI) guides the microscope to specific locations for subsequent high-resolution imaging.
• Reconstructed 3D coordinates precisely map to electron beam positions for targeted analysis.

Main Results:

• Demonstrated vesicle-resolved imaging of synaptic boutons in human cortex.
• Achieved targeted imaging of over one million human cortical mitochondria.
• Enabled hierarchical imaging of a nematode nervous system and retrospective targeting of a petabyte-scale dataset.

Conclusions:

• RAEM transforms serial-section EM into a query-driven, multi-resolution approach.
• This method facilitates scalable biomedical discovery by integrating broad anatomical context with ultrastructural detail.
• RAEM allows for efficient and targeted investigation of nanoscale structures within large biological samples.