Ulrike Aust

The role of item- and category-specific information in the discrimination of people versus nonpeople images by pigeons

Herrnstein and Loveland (1964, pp. 549–551) successfully trained pigeons to discriminate pictures showing humans from pictures that did not. In the present study, a go/no-go procedure was employed to replicate and extend their findings, the primary focus of concern being to reevaluate the role of item- and category-specific information. The pigeons readily acquired the discrimination and were also able to generalize to novel instances of the two classes (Experiment 1). Classification of scrambled versions of the stimuli was based on small and local features, rather than on configural and global features (Experiment 2). The presentation of gray-scale stimuli indicated that color was important for classifying novel stimuli and recognizing familiar ones (Experiments 1 and 2). Finally, the control that could possibly be exerted by irrelevant background features was investigated by presenting the pigeons with images of persons contained in former person-absent pictures (Experiment 3). Classification was found to be controlled by both item- and category- specific features, but only in pigeons that were reinforced on person-present pictures was the latter type of information given precedence over the former.

Natural categorization through multiple feature learning in pigeons.

Recently (Troje, Huber, Loidolt, Aust, & Fieder 1999), we found that pigeons discriminated between large sets of photorealistic frontal images of human faces on the basis of sex. This ability was predominantly based on information contained in the visual texture of those images rather than in their configural properties. The pigeons could learn the distinction even when differences of shape and average intensity were completely removed. Here, we proved more specifically the pigeons’ flexibility and efficiency to utilize the class-distinguishing information contained in complex natural classes. First, we used principal component as well as discriminant function analysis in order to determine which aspects of the male and female images could support successful categorization. We then conducted various tests involving systematic transformations and reduction of the feature content to examine whether or not the pigeons’ categorization behaviour comes under the control of categorylevel feature dimensions—that is, those stimulus aspects that most accurately divide the stimulus classes into the experimenter-defined categories of “Male” and “Female”. Enhanced classification ability in the presence of impoverished test faces that varied only along one of the first three principal components provided evidence that the pigeons used these class-distinguishing stimulus aspects as a basis for generalization to new instances.

Visual artificial grammar learning: comparative research on humans, kea (Nestor notabilis) and pigeons (Columba livia)

Artificial grammar learning (AGL) provides a useful tool for exploring rule learning strategies linked to general purpose pattern perception. To be able to directly compare performance of humans with other species with different memory capacities, we developed an AGL task in the visual domain. Presenting entire visual patterns simultaneously instead of sequentially minimizes the amount of required working memory. This approach allowed us to evaluate performance levels of two bird species, kea (Nestor notabilis) and pigeons (Columba livia), in direct comparison to human participants. After being trained to discriminate between two types of visual patterns generated by rules at different levels of computational complexity and presented on a computer screen, birds and humans received further training with a series of novel stimuli that followed the same rules, but differed in various visual features from the training stimuli. Most avian and all human subjects continued to perform well above chance during this initial generalization phase, suggesting that they were able to generalize learned rules to novel stimuli. However, detailed testing with stimuli that violated the intended rules regarding the exact number of stimulus elements indicates that neither bird species was able to successfully acquire the intended pattern rule. Our data suggest that, in contrast to humans, these birds were unable to master a simple rule above the finite-state level, even with simultaneous item presentation and despite intensive training.

Target-defining features in a "people-present/people-absent" discrimination task by pigeons.

Aust and Huber (2001) demonstrated that in apeople-present/people-absentdiscrimination task, pigeons actually attended to properties of the target (i.e., the human figure). The aim of the present effort was to specify what sort of information contained in the target was used for classification, as well as to investigate whether and in what way the target-defining features interacted. Six pigeons were trained in a go/no-go procedure to discriminate between color photographs characterized by the presence or absence of people. They were then presented with various types of test stimuli that contained some category-relevant features but lacked others. The results showed that properties related to target size and internal structure played an important role and that human silhouettes were insufficient for eliciting a people-present response. Furthermore, some properties of the human figure (e.g., hands/arms) made good predictors of the people-present category, whereas others (e.g., feet/legs or skin color) did not. Responses to test stimuli that belonged to the people-absent category but nevertheless contained some features normally typical for humans (e.g., nonhuman primates) provided evidence that various category-relevant features contributed to classification in an additive way. Taken together, the results suggest that the pigeons made use of a polymorphous class rule involving collections of differently weighted target features.

The Vienna comparative cognition technology (VCCT): An innovative operant conditioning system for various species and experimental procedures

This article describes a laboratory system for running learning experiments in operant chambers with various species. It is based on a modern version of a classical learning chamber for operant conditioning, the so-called “Skinner box”. Rather than constituting a stand-alone unit, as is usually the case, it is an integrated part of a comprehensive technical solution, thereby eliminating a number of practical problems that are frequently encountered in research on animal learning and behavior. The Vienna comparative cognition technology combines modern computer, stimulus presentation, and reinforcement technology with flexibility and user-friendliness, which allows for efficient, widely automatized across-species experimentation, and thus makes the system appropriate for use in a broad range of learning tasks.

Elemental versus configural perception in a people-present/people-absent discrimination task by pigeons

In the present experiment, we investigated whether pigeons rely exclusively on elemental information or whether they are also able to exploit configural information in apeople-present/people-absent discrimination task. Six pigeons were trained in a go/no-go procedure to discriminate between 800 color photographs characterized by the presence or absence of people. Thepeople-present stimuli were designated as positive, and thepeople-absent stimuli were designated as negative. After training and a subsequent generalization test, the pigeons were presented with both familiar and novel people-present stimuli containing human figures that were distorted in one of seven different ways. All the pigeons learned the initial discrimination and also showed generalization to novel stimuli. In the subsequent test, performance on all types of distorted stimuli was diminished in comparison with that on the intact original pictures from which they had been derived. At the same time, however, peck rates clearly exceeded the level of responding found for regular people-absent stimuli. This result strongly suggests that responding was controlled by both the constituting target components and their spatial relations and, therefore, points to the dual importance of elemental and configural information.

Representational insight in pigeons: comparing subjects with and without real-life experience

Understanding the relation between objects and their pictures at a level beyond mere feature discrimination is by no means a trivial cognitive ability, and support of this is still weak in nonhuman species. Here, we report evidence of representational insight in pigeons. Responding to pictures of human body parts was compared in birds that had extensive pre-experience with live humans and in birds that had never seen any human heads. In a two-alternative forced-choice procedure the pigeons were trained to discriminate between pictures of either handless or headless humans and nonhumans. On test, the birds had to choose (i) between body parts they had already seen in training and the parts that had been missing, (ii) between previously seen parts and arbitrary skin patches, and (iii) between previously missing parts and skin patches. Only the pigeons that lacked experience with real heads and were trained with pictures of headless humans failed to show a significant preference for pictures of missing parts (i.e., heads) over arbitrary skin patches. This demonstrates the importance of individual experience with the real 3D-referents of pictures for classification of the latter and is thereby evidence of representational insight.

Does the Use of Natural Stimuli Facilitate Amodal Completion in Pigeons

Three experiments were carried out to investigate whether amodal completion in pigeons can be facilitated by the use of colour photographs instead of highly artificial stimuli such as geometrical shapes. Ten pigeons were trained in a go/no-go procedure to discriminate between photographs of complete and of incomplete pigeon figures. In the subsequent test, the birds classified pictures of partly occluded pigeons as though they were complete (experiment 1). However, we found evidence that classification was based on spurious stimulus features that paralleled the intended class rule of figural completeness versus incompleteness. In particular, classification was shown to be guided by white background gaps that separated the parts of the fragmented pigeon figures (experiment 2), as well as by cues related to overall Gestalt (experiment 3). In summary, the present results indicate that the use of more natural stimuli such as photographs instead of geometrical shapes is insufficient for providing amodal completion in pigeons. It is suggested that a combination of various cues, including, eg, 3-D information and common motion in addition to surface and contour properties, may be required to induce a perceptual bias favouring visual completion of occluded portions.

Pigeons use item-specific and category-level information in the identification and categorization of human faces.

Three experiments examined the pigeons ability to adapt to the specific tasks of identification and categorization of complex visual stimuli by attending to information at different levels of abstraction. In Experiment 1, pigeons were successively trained to discriminate first between pairs of human faces from the same sex and then between male and female faces. A final test revealed that the associations formed in the first training phase did not survive categorization training, indicating that the pigeons used item-specific information to discriminate faces of the same sex, but shifted their attention to category-level information for discriminating between the male and female categories. Two further experiments disproved alternative explanations. The results support the feature-learning account of open-ended categorization in animals.

More than one way to see it: Individual heuristics in avian visual computation

Graphical abstract