Only 1/10,000 of the total volume need be imaged [1] - A mammalian brain has highly repetitive neural circuitry (compared to C. Elegans) and only a tiny fraction of a mammalian brain’s neurons and connections would need to be traced in order to elucidate the full statistics of its regions, pathways, and synaptic circuits. Unfortunately, this volume is relatively evenly distributed across the brain’s total volume and so demands full-volume slicing and a sparse-directed imaging protocol.

Reduce total number of slices needing to be cut and handled by cutting thicker slices (100nm to 1mm) and imaging using electron tomography – A major hurdle to synapse-resolution imaging is the fragileness (and shear numbers) of ultrathin tissue slices. Electron tomography (often using higher voltages than standard 100kV TEM’s) tilts the tissue relative to the electron beam in order to gain information about the structures of the tissue perpendicular to the slice. This can allow 10x better z-resolution than the slice thickness, thus reducing the total number of slices. In addition, thicker slices are more robust to handling and easier to cut. 

Automate microtome slice cutting via lathe-like spiral cutting and block-face taping – Traditional microtomes disengage and reengage the block for each successive slice. A lathe-microtome can cut a continual spiral of tissue in one smooth, continuous motion. Block-face taping is a technique in microtomy in which an adhesive tape is bonded to the tissue block face before cutting. This has the effect making a thin slice as robust as a thicker slice would be, and at the same time provides a holding mechanism for the newly sliced tissue. Note: Adhesives interfere with TEM imaging so more specific modifications to standard block-face taping must be made (along with significant process development research) to make this technique viable.

Automate microtome slice retrieval by collecting cut slices on a tape during continuous cutting procedure – Standard microtomy work involves the collection of newly cut tissue slices from a water bath surrounding the knife. This is an intrinsically manual process. Using a tape (like the block-face tape above) to collect the slices nicely automates this collection process and provides a medium for tissue storage and random access imaging.

Automate TEM loading of multiple samples (and allow random access) using tape cassette of tissue slices – TEM’s are high vacuum instruments. In order to load a new sample for imaging requires the breaking of vacuum and subsequent pump-down. This is incredibly time consuming. A “tape cassette” like device can prevent the need for breaking vacuum and can allow fast random access to any slice on a single tissue tape. The tape is feed through the TEM such that it passes through the path of the electron beam.  

Use complete MRI and LM (light microscope) atlases of same brain to complement and help register TEM sparse atlas – All cell bodies and myelinated axons can be imaged directly using only a 1um resolution LM atlas. TEM imaging need only be performed on dendritic processes and axonal arbors where synapses occur. This further reduces the total amount of TEM images that must be performed to map out a particular neural circuit, and it allows the precise slice needing to be imaged to be determined in the LM atlas before TEM loading and imaging begins, obviating the need for scanning large areas in the TEM.

Use voxel-based automatic neural tracing algorithms on the LM and TEM images to preclude hand-tracing and to automatically steer TEM imaging – Automating the 3D reconstruction process is a difficult proposition and significant research has been put toward this goal in the past with some success. Automating tracing is even more important when using sparse-directed imaging since some experiments require steering the TEM imaging, in order to follow a particular neural process, without the need for human intervention. (see Types of Experiments that can be performed using a Single Brain Physical Slice Library)  

Use the Single Brain Physical Slice Library (SBPSL) paradigm to invite collaboration (and cost sharing) of neuroanatomical specialists in an online collective brain mapping experiment – A significant barrier to the successful creation and use of a mammalian SSRBC Atlas is the need for many neuroanatomical specialists (who are experts in the know neural circuitry of particular brain regions) to collaborate together in order to intelligently guide the sparse imaging process. Given the technological slicing, imaging, and tracing infrastructure described above, the whole process could be made available online for remote operation (a SBPSL). This would allow researchers with specific brain connectivity questions to logon and create batch-imaging jobs specific to their questions. The most exciting prospect is that one researcher could follow up a previous researcher’s mapping experiment by imaging and tracing processes connected to the exact same neurons as the previous study.