Neurons avoid talking to themselves by using 19,000 forms of one gene:

A lot of effort has gone into exploring a key aspect of wiring the brain, the ability of nerve cells to form and maintain specific connections among the huge number of potential targets within the nervous system. Just as important, however, is the opposite question: how do these neurons avoid making inappropriate connections? Perhaps the most obvious form of inappropriate connection is the equivalent of a neural short-circuit, a case where a nerve cell forms connections to itself.

We propose that this vast repertoire of Dscam recognition molecules is sufficient to provide each neuron with a unique identity and homotypic binding specificity, thereby allowing neuronal processes to distinguish between self and nonself.

Very, very interesting article. Unfortunately the full text article is not available for one reason or another. So anyways, based on the abstract it seems that neurons are able to determine if they are in contact with there own cell body. There could obviously be issues (i.e., short-circuit like scenarios) if the neuron was to excite itself instead of other nearby neurons. I find it hard to believe that only 19,000 forms of a particular gene are enough to distinguish all the billions of neurons in a person’s brain. I’d imagine that some other factor (or factors) would be playing a role in making neuron’s even more unique. Anyways, very interesting article, but I wish I could get the full text for it.