A Simple Artificial Life Model Explains Irrational Behavior in Human Decision-Making

• Grant number: 12/05211-7
• Support Opportunities: Regular Research Grants - Publications - Scientific article
• Start date: June 01, 2012
• End date: November 30, 2012
• Field of knowledge: Biological Sciences - Physiology - Physiology of Organs and Systems
• Principal Investigator: Marcus Vinícius Chrysóstomo Baldo
• Grantee: Marcus Vinícius Chrysóstomo Baldo
• Host Institution: Instituto de Ciências Biomédicas (ICB). Universidade de São Paulo (USP). São Paulo , SP, Brazil
• City of the host institution: São Paulo

Abstract

Although praised for their rationality, humans often make poordecisions, even in simple situations. In the repeated binary choiceexperiment, an individual has to choose repeatedly between the sametwo alternatives, where a reward is assigned to one of them with fixedprobability. The optimal strategy is to perseverate with choosing thealternative with the best expected return. Whereas many speciesperseverate, humans tend to match the frequencies of their choices tothe frequencies of the alternatives, a sub-optimal strategy known asprobability matching. Our goal was to find the primary cognitiveconstraints under which a set of simple evolutionary rules can lead tosuch contrasting behaviors. We simulated the evolution of artificialpopulations, wherein the fitness of each animat (artificial animal)depended on its ability to predict the next element of a sequence madeup of a repeating binary string of varying size. When the string wasshort relative to the animats' neural capacity, they could learn itand correctly predict the next element of the sequence. When it waslong, they could not learn it, turning to the next best option: toperseverate. Animats from the last generation then performed the taskof predicting the next element of a non-periodical binary sequence. Wefound that, whereas animats with smaller neural capacity keptperseverating with the best alternative as before, animats with largerneural capacity, which had previously been able to learn the patternof repeating strings, adopted probability matching, being outperformedby the perseverating animats. Our results demonstrate how the abilityto make predictions in an environment endowed with regular patternsmay lead to probability matching under less structured conditions.They point to probability matching as a likely by-product of adaptivecognitive strategies that were crucial in human evolution, but maylead to sub-optimal performances in other environments. (AU)