When I read (and wrote about) the recent controversy between Modha and Markram, I had this inescapable déjà vu feeling….weird, where did I hear that somebody already simulated a “brain” of the scale of the human brain? Of course!…. Eugene Izhikevich, a very bright (and VERY funny) neuroscientist that, in 2007, visited our center CELEST. During that visit, he showed what at that time (in 2005, and may be up to today) was one of the “largest scale” neural simulation.

Here is an excerpt from his web page:

“I develop large-scale models of the brain having microcircuitry of the mammalian thalamo-cortical system.
On October 27, 2005 I finished simulation of a model that has the size of the human brain (bold added by Max…). The model has 100,000,000,000 neurons (hundred billion or 10^11) and almost 1,000,000,000,000,000 (one quadrillion or 10^15) synapses. It represents 300×300 mm^2 of mammalian thalamo-cortical surface, specific, non-specific, and reticular thalamic nuclei, and spiking neurons with firing properties corresponding to those recorded in the mammalian brain. The model exhibited alpha and gamma rhythms, moving clusters of neurons in up- and down-states, and other interesting phenomena. One second of simulation took 50 days on a beowulf cluster of 27 processors (3GHz each).

Well, here we go… Eugene has done a more detailed simulation, and a more thorough analysis of it from the scientific point of view, than the “simple cat” built at IBM… and wait a second: he used a simple workstation! Only later he used a cluster, just to speed it up a bit! Eugene, though, was very careful in exactly defining what the simulation is, which, if I recalled correctly, was: “this is the largest simulation of a system with the same number of neurons as a human brain”. Bravo Eugene! He was really skilled in saying exactly what this was, stopping the critiques before they could even be formulated.

One thing to keep in mind is that the size of the simulation in terms of neurons implemented is not necessarily the most crucial aspect: you can simulate a neuron with one, or with a million differential equations. What is larger scale, a network of 1/2 millions neurons, each of which is simulated by 2 differential equations, of a 1-neuron network, simulated with 1 million differential equations? If the metrics is the number of equations…well, you got the point.

Bottom line: you might be reading this blog in a machine that could be able to simulate your brain, now! If Eugene did it, you can do it as well!

Anyway, the biggest challenge that lies ahead the field, and SyNAPSE as well, is not to build the largest size simulation, but the one that exhibit intelligent behavior, which eventually has to be judged on a completely different scale. So, forget about the metrics called “number of neurons”: more exciting, and surely challenging, simulations will be needed to claim we achieved even the intelligence of an insect!

Max Versace