Comments on: Polychronization: Computation With Spikes

By: Chris Engels

Chris Engels — Wed, 08 Nov 2006 16:12:15 +0000

Bill, unfortunatly your link to article “An Algorithm for the Calculation of Polychronous Groups”
is dead. Would you fix it please? Thanx in ad!

By: Bill Maier

Bill Maier — Fri, 14 Apr 2006 13:45:11 +0000

The idea of polychronous groups is fascinating. At TCU we had Izhikevich come
and give two talks on the subject a couple of weeks ago. Very interesting.
However I don’t believe plasticity is a requirement for the development of
these groups. For my own algorithm on this, see my paper at

http://mysite.verizon.net/wmaier/articles.html

Polychronous groups arise naturally in networks with delays. The relation
of STDP or any learning scheme to these groups is not clear.

By: Bayle

Bayle — Fri, 06 Jan 2006 20:02:50 +0000

BTW, a friend pointed out to me that Izhikevich’s STDP rule (figure 4) is an approximation that is only correct for “isolated” spikes. When spiking frequency is more than 20hz, this is not the sort of plasticity that actually happens. So, this is one way in which we know that Izhikevich’s simulations will not accord with reality.

By: Bayle

Bayle — Fri, 06 Jan 2006 00:58:58 +0000

Thanks to Chris Chatham's great weblog Developing Intelligence for pointing me to this article. BTW, I know that the phrase "STDP plasticity" sounds silly if you expand the acronym -- but that's what happens with acronyms, go figure.