Accuracy of rats in discriminating visual objects Is explained by the complexity of their perceptual strategy

Despite their growing popularity as models of visual functions, it is widely assumed that rodents deploy perceptual strategies not nearly as advanced as those of primates, when processing visual objects. Such belief is fostered by the conflicting findings about the complexity of rodent pattern vision, which appears to range from mere detection of overall object luminance to view-invariant processing of discriminant shape features. Here, we sought to clarify how refined object vision is in rodents, by measuring how well a group of rats discriminated a reference object from eleven distractors, spanning a spectrum of image-level similarity with the reference. We also presented the animals with random variations of the reference, and we processed their responses to these stimuli to obtain subject-specific models of rat perceptual choices. These models captured very well the highly variable discrimination performance observed across subjects and object conditions. In particular, they revealed how the animals that succeeded with the more challenging distractors were those that integrated the wider variety of discriminant features into their perceptual strategy. Critically, these features remained highly subject-specific and largely invariant under changes in object appearance (e.g., size variation), although they were properly reformatted (e.g., rescaled) to deal with the specific transformations the objects underwent. Overall, these findings show that rat object vision, far from being poorly developed, can be characterized as a feature-based filtering process (iterated across multiple scales, positions, etc.), similar to the one that is at work in primates and state-of-the-art machine vision systems, such as convolutional neural networks