Signal decoding and receiver evolution - An analysis using an artificial neural network

We use a connectionist model, a recurrent artificial neural network, to inv estigate the evolution of species recognition in sympatric taxa, We address ed three questions: (1) Does the accuracy of artificial neural networks in discriminating between conspecifics and other sympatric heterospecifics dep end on whether the networks were trained only to recognize conspecifics, as opposed to being trained to discriminate between conspecifics and sympatri c heterospecifics? (2) Do artificial neural networks weight most heavily th ose signal features that differ most between conspecifics and sympatric het erospecifics, or those features that vary less within conspecifics! (3) Doe s selection for species recognition generate sexual selection! We find that : (1) Neural networks trained only on self recognition do not classify spec ies as accurately as networks trained to discriminate between conspecifics and heterospecifics, (2) Neural networks weight signal features in a manner suggesting that the total sound environment as opposed to the relative var iation of signals within the species is more important in the evolution of recognition mechanisms. (3) Selection for species recognition generates sub stantial variation in the relative attractiveness of signals within the spe cies and thus can result in sexual selection, Copyright (C) 2000 S. Karger AG, Basel.