Note: to get this to work with the latest version of Java you need to add this website: - to the list of exceptions in your java control panel. This is a huge pain in the neck, I'm sory.

Here is the simulation of a fluid neural network. Active squares (blue) move and inactive squares (red), well, don't. Squares can be activated either by self activation, in which they become active during a timestep with a probability of about 1 in 200, or they can be activated by their neighbors. At low densities, such as the example above which has eight 'ants', this periodicity does not appear. One ant will become active at random, activate some, none or all of it's neighbors, and then they will stop. They will stay inactive for an interval of random length, until one becomes active at random again. Here is a look at a graph of a typical run. The y axis is the number of active 'ants' plotted against each timestep.

Definitely not periodic, however, look what happens when there are eighty ants instead of eight. [Click here to see...]