Dr. rer. nat. Daniel R. Berger
Research Associate (MIT/HHMI)
Office: 43 Vassar Street, Room 46-5065
EMail: dberger (at) mit.edu
Connectomics, Image alignment and stitching, Electron microscopy, 3D reconstruction, Cortical synaptic connectivity
The emerging field of Connectomics pushes the automatization of precise cutting and imaging of biological tissue samples at electron-microscopic resolution. This allows neuroscientists for the first time to analyze large volumes of brain tissue in 3D, to study the actual connectivity of networks of neurons.
To this end, I am collaborating with the lab of Jeff Lichtman at Harvard University, where methods of automated ultramicrotome tissue cutting, slice collection and imaging in electron microscopes are developed. Slices are typically thinner than 30 nm, and stacks of thousands of such slices can be cut and collected reliably. Very large electron-microscopic images are then taken at very high resolution (i.e., 32k*32k pixels at 0.8 nm pixel size).
The Seung Lab specializes in automating the analysis of such image stacks. This involves aligning the images in the stacks to produce a consistent 3D volume, segmenting the volume into individual neurons and pieces of neurons, finding synapses, and analyzing the connectivity.
My work so far involved the alignment and stitching of very large image stacks, manual labeling of neurites and synapses in those data sets, 3D rendering of those objects, and analysis of local connectivity. This work was introduced in my PI's TED talk, "I am my connectome". I also wrote software to help the acquisition of large electron-microscopic image stacks in the Lichtman Lab, and I have recently also done tissue preparations for electron microscopy as well as ultra-thin cutting.