Reconstructed rat neural and astroglial circuitry
[youtube=http://www.youtube.com/watch?v=FZT6c0V8fW4]
This 6.5 min video was posted online in Oct '10, drawing on the research from this paper, which used serial section transmission em and did 3d reconstruction in small volumes of the rat hippocampus. The intro scene explains that, in their schema:
dendrites = yellow
postsynaptic densities = red
endoplasmic reticula = cyan
mitochondria = magenta
axons = green
dendritic spines = yellow (looks kind of gold-ish, though)
synapses = copper
astrocytes (key components of the tripartite synapse) = blue
The video builds up the components from the ground up and "zooms" through at different angles, giving perspectives that can't be seen from just one 2d shot.
The map is static, but of course in real time all of these components can and do change. For example, endoplasmic reticula have a fast turnover in dendritic spines with a half life of > 1 day (source). Mitochondria move quickly along microtubules (~1 um/s) and have a longer half life of ~ 20 days (source). That doesn't mean that these sort of maps don't have tremendous value! But it's useful to include the motion in your mental models.
The authors discuss in the paper how many synapses are missed by non-em approaches, in part because contact between a dendritic spine and an axon only sometimes leads to a synapse. Still, going forward, it will be interesting to see to what degree these kind of maps can be approximated probabilistically. And if so, how might one represent that uncertainty in 3d space?
References
Mishchenko Y et al. 2010 Ultrastructural Analysis of Hippocampal Neuropil from the Connectomics Perspective. doi: 10.1016/j.neuron.2010.08.014
Toresson H, et al. 2005 Dynamic distribution of endoplasmic reticulum in hippocampal neuron dendritic spines. doi: 10.1111/j.1460-9568.2005.04342.x
Perkins and Ellisman, "Brain energetics: integration of molecular and cellular processes." In Brain energetics: integration of molecular and cellular processes, Ch 4.2, p. 279.