Action Potential bandwidth

Bill Skaggs skaggs at bns.pitt.edu
Sun Aug 20 09:50:45 EST 2000


Well, here's a shot at it:  There are, by the most current estimates
I've seen, about 5x10^10 neurons in the human brain.  Of these, about
4x10^10 are tiny granule cells in the cerebellum, which is a sort of
"peripheral device" for the brain.  The firing rates of granule cells
are not known with any accuracy, because they are very small and very
closely crowded together and hence difficult to record from.  My
guess, based on functional considerations and comparison with other,
better studied types of neurons, is that the firing rates are low,
probably in the neighborhood of 0.1 action potentials per second.
If I am right, this gives around 4x10^9 spikes per second.

Next, the human neocortex contains about 10^10 pyramidal cells, which
are much more thoroughly understood.  At least half of them are
"regular spiking" cells, having average firing rates in the
neighborhood of 3-10 spikes per second.  The remaining cells have much
lower average firing rates.  This yields, say, 5x10^10 spikes per
second from the neocortex.

Other types of neuronss in the brain are less numerous by several
orders of magnitude, and probably don't contribute significantly to
the total number of spikes per second.

So all in all, I would say that your guess of 10^11 firings per second
is right on the mark.

	-- Bill






More information about the Neur-sci mailing list