Becoming a Visual Expert

After 17 years of experimental research we have proposed a new and comprehensive theory of perception and categorization in pigeons (Columba livia) (Huber and Aust 2006). We believe that pigeons are able to quickly and flexibly extract the object- or class-specific features and use them for the formation of representations of complex objects or categories. It remained, however, unclear whether they can also represent (configural) relationships between those features. We assume that they would need a great amount of individual experience with the objects or classes, acquired during their ontogeny. With this project we want to examine in pigeons the role of developmental and training-specific experience in the recognition of static, configurally complex, natural (e.g. human faces) and artificial (Greebles) objects, as well as in biological motion perception. As pigeons are certainly not "naturally born experts" for human face recognition and at the same are themselves a highly mobile bird species, they are well suited for both directions of research. There is evidence that in humans the mechanisms mediating face recognition differ from those underlying recognition of other objects in the degree to which they depend on configural or holistic representations rather than on featural or decompositional analysis. However, that these mechanisms have become distinctive not through natural but through developmental selection operating on domain-general recognition mechanisms has been suggested from dog expert and Greeble (artificial forms) studies. Most interestingly, configural processing is also reported to be involved in biological motion perception. Given the extraordinary visual capacities of pigeons and their surprising performance in categorizing and memorizing of complex stimuli, we expect that pigeons may also become experts for faces, Greebles and complex movement patterns if appropriately trained. The role of object- specific expertise will be addressed by comparing groups of pigeons that differ in their amount of visual experience with humans, Greebles and specific movement patterns (e.g. walking legs) acquired during early ontogeny and by comparing groups of pigeons that differ in the amount of training and the level of categorization (basic vs. subordinate). Furthermore we will pursue in-depth analyses of the mathematical features and the statistical potentials of a newly established multiple-matching paradigm (Huber et al., 2005) that we will use (besides a more conventional go/no- go method and matching-to-sample) as core method for training and testing pigeons in this project. By modeling the experimental behavior of subjects by means of game theory and several bio-mathematical concepts, we aim at integrating our research into a broader scientific framework. The procedural and statistical potentials of the multiple-matching paradigm will be addressed by meticulous statistical analyses of pigeons` performances and by mathematical modeling and computer simulations.

After 17 years of experimental research we have proposed a new and comprehensive theory of perception and categorization in pigeons (Columba livia) (Huber and Aust 2006). We believe that pigeons are able to quickly and flexibly extract the object- or class-specific features and use them for the formation of representations of complex objects or categories. It remained, however, unclear whether they can also represent (configural) relationships between those features. We assume that they would need a great amount of individual experience with the objects or classes, acquired during their ontogeny. With this project we want to examine in pigeons the role of developmental and training-specific experience in the recognition of static, configurally complex, natural (e.g. human faces) and artificial (Greebles) objects, as well as in biological motion perception. As pigeons are certainly not "naturally born experts" for human face recognition and at the same are themselves a highly mobile bird species, they are well suited for both directions of research. There is evidence that in humans the mechanisms mediating face recognition differ from those underlying recognition of other objects in the degree to which they depend on configural or holistic representations rather than on featural or decompositional analysis. However, that these mechanisms have become distinctive not through natural but through developmental selection operating on domain-general recognition mechanisms has been suggested from dog expert and Greeble (artificial forms) studies. Most interestingly, configural processing is also reported to be involved in biological motion perception. Given the extraordinary visual capacities of pigeons and their surprising performance in categorizing and memorizing of complex stimuli, we expect that pigeons may also become experts for faces, Greebles and complex movement patterns if appropriately trained. The role of object- specific expertise will be addressed by comparing groups of pigeons that differ in their amount of visual experience with humans, Greebles and specific movement patterns (e.g. walking legs) acquired during early ontogeny and by comparing groups of pigeons that differ in the amount of training and the level of categorization (basic vs. subordinate). Furthermore we will pursue in-depth analyses of the mathematical features and the statistical potentials of a newly established multiple-matching paradigm (Huber et al., 2005) that we will use (besides a more conventional go/no- go method and matching-to-sample) as core method for training and testing pigeons in this project. By modeling the experimental behavior of subjects by means of game theory and several bio-mathematical concepts, we aim at integrating our research into a broader scientific framework. The procedural and statistical potentials of the multiple-matching paradigm will be addressed by meticulous statistical analyses of pigeons` performances and by mathematical modeling and computer simulations.