EAGER: Efficient coding and the evolution of animal signals

University Of Maryland Baltimore County, Baltimore MD

Investigators

Abstract

An outstanding question in evolutionary biology is why many animals use complex and exaggerated colors and sounds during courtship. These signals are intriguing not only because they may attract the attention of predators, but also because they come in many forms, with closely-related species differing in important ways. The efficient coding hypothesis (EC) is a new explanation that states that the human visual system is very good at making sense of the world around us, and when something is easy to understand, it is preferred. EC therefore predicts that male courtship signals will evolve to match their surroundings so that they are easily understood by female receivers. This prediction is in stark contrast to other current theories which predict that courtship signals should stand out from their backgrounds so that they are more easily detected. Here the researchers develop statistical and computational methods to test EC, and apply them to a group of colorful fish found in the eastern United States. Their main objectives are to develop the toolkit to apply EC with many animals and to gather the information they need to design manipulative experiments that will test EC directly. Along the way, they will also provide international training opportunities for student researchers and after-school activities for school children in Baltimore. This project explores the role of efficient coding in the diversification of complex animal signals. Efficient coding is grounded in information theory and describes how sensory systems maximize information from the environment while minimizing redundancies, thus limiting the number of neurons required to encode relevant information. The efficient coding hypothesis (EC) is applied here to animal signal design to ask whether visual systems adapted to distinct habitats can contribute to the diversification of complex mating signals. EC goes beyond the current signal-detection framework of sensory drive, which predicts that animal signals maximize attractiveness by maximizing detectability. EC predicts that attractive signals also should match statistical parameters of the environment that sensory systems have evolved to decode. The main objectives are to establish the validity of EC in non-humans and to adapt methods developed for humans in terrestrial environments into a toolkit for understanding complex visual signal evolution across animals. The focal study system is the North American freshwater fish genus Etheostoma, called darters. Male darters are characterized by a diversity of vivid, complex, species-specific nuptial color patterns. Although closely-related darter species occupy similar habitats, more distantly-related species have distinct habitat preferences. The central hypothesis is that the adaptation of darter visual systems to encode the statistical regularities of species-specific habitats has contributed to the diversification of complex color patterns across the genus. Three predictions of that hypothesis are tested.