The line for state 0, age 0, contemplates up to eight neighbors, various combinations of which can cause activation or infection. Having a single neighbor initiate the reaction may be more realistic epidemiologically, but it also leads to a high level of overall excitation.
Requiring a pair of neighbors to initiate an infection, it is harder generate activity with low densities of concentration, since the occurrence of pairs is much rarer than the occurrence of singlets. Once activation occurs, enough neighbors are usually excited all at once to become self sustaining.
Note that there is a difference between having an exact number of neighbors which will initiate an infection, and a threshold by which any larger number will also suffice. Epidemiologically the latter would probably be more realistic; the rule may be chosen to suit either alternative.