Abstract
A major area of investigation in evolutionary biology focuses on understanding how intelligent behaviors first evolved. We have evidence in the fossil record that demonstrates an apparent increase in the upper bounds of organismal complexity over time, but the levels of intelligence displayed by those organisms is less clear. For example, the progression of behaviors registered in trace and other fossils from the Ediacaran period have inspired intense speculation as to the cognitive capacity of animals leading up to the Cambrian Explosion. While it is challenging to get a more detailed window into what actually transpired hundreds of millions of years ago, computational Artificial Life techniques allow us to conduct empirical studies under analogous conditions and examine the patterns by which intelligent behaviors arise. In a series of experiments using the Avida platform, we evolved digital organisms with simple sensory and locomotory potential that were capable of increasingly complex cognitive abilities, spanning from efficient patch harvesting to associative learning and nonelemental learning. The patterns of the evolutionary sequences of these organisms are reminiscent of those found in Precambrian fossils, and allow us to start refining our ideas about the evolutionary origins of intelligence.