Ultimately the development of the brain is guided genetically, by instructions from the DNA. However once the basic wiring pattern is in place, the brain continues to organize itself on the basis of the information flowing through it. The data-driven organization is essential for memory as well as feature extraction and a host of other important human capabilities.

Much of our knowledge of neural self-organization comes from studies in machine learning. Machines give us a friendly environment in which to test mathematical theories of data organization, because we can experiment with various algorithms quickly and efficiently without the complex preparation involved in electrode penetration. Still, there is more to self-organization than just the modification of synaptic weights.

Currently, we understand only a little about human memory. Most of what we know about self organization pertains to the geometrical and statistical development of network wiring. What we know for sure, is that the Hebbian models of synaptic plasticity are not adequate to describe the human memory system. Memory is complex, different brain pathways subserve different functions and even the nature of episodic storage is poorly understood. In this section, we'll begin discovering how computational modeling can help us, by constructing some simple models. We'll look at elementary forms of synaptic plasticity and topographic wiring, and we'll start visualizing the neural information timeline.