IEEE Future Directions
Related Articles
The human brain is being studied by big consorzia in project funded by governments in Asia, Europe and US. The Human Connectome is working to develop a complete map of the connectivity network among our brain’s neurones, a seemingly impossible feat.
Smaller initiatives are popping up in many places and industries focusing on brains that are easier to manage, having much smaller size. So far we have seen the complete mapping of neuronal pathways of worms, involving a few hundred neurones.
Now Google has released a almost complete connectome of a fruit-fly brain, 25,000 neurones and 20 million synapses. It is a gigantic step although an infinitesimal one on the path to a complete human brain conncectome (compare those figures with the 100 billion neurones and the 10s of trillions synapses in a human brain).
You can see the rendering of the fruit-fly brain in the image and take a look at some if its neuronal pathways in the clip.
Even such a small brain would have required (Google estimate) some 250 people working for 2 years to stitch the various neurones by observing the brain slices of the fruit-fly brain each 20µ thick (actually thin!).
The analyses of the slices and their stitching and 3D rendering was performed automatically by AI based software. The Google Connectomic Group developed the tools to make this possible and have now released them for free. You can get more info on their approach and see more clips on the technology here.
I find this result fascinating, and the availability of the tools coupled with almost unlimited processing power now available will make quick advances possible.
Notice that having a complete connectome does not tell us how the brain works, but surely provides some hints that may help in progressing our understanding. Finding out what are those connections representing, what are they being used for and what kind of meaning they generate remains a challenge. This is why the clip showing neural circuits is so interesting. It has yet to be proven complete and exhaustive and that will take more time. As an example, Google scientists believe they have identified the circuitry that control the fly … fly and these seems to be composed by a relative small number of neurones (some 5,000). By comparison we are using billions of transistors to take care of the flight of a commercial airplane. Nature seems to be much more efficient than we are in this job (true enough Nature doesn’t care if a million flies die because of a glitch in the circuit, as long as many survive and procreate a new generation…).
