How birds outperform humans in multi-component behavior

Abstract

Recent years have witnessed an astonishing flurry of studies demonstrating that some bird species show higher-order cognitive processes on par with primates [1-3]. As birds have no neocortex, cortical processing cannot be a requirement for higher order cognition [1,4]. Although birds have more neurons than expected from their small brain weights [5], their absolute neuron count is still lower compared to cortical neuron numbers of primates. How, then, is it possible that pigeons reach performance levels in, for example, abstract numerical competence and orthographic processing, that are comparable to that of macaques [6]? While the subpallium is very similar, the organization of the pallium differs tremendously between birds and mammals [1]; moreover, the avian pallium is characterized by small, extremely tightly packed neurons [5]. It is conceivable that signal processing could be faster in such a brain as a result of a higher speed of propagation of activation between neighboring assemblies, resulting in faster switch times between neighboring networks and neuronal representations of behavioral goals. This is important, as behavioral goals in real-life situations are often achieved by a series of sub-tasks [7,8], and especially when sub-tasks supersede each other and show little overlap in processing resources, neocortical (pallial) structures are involved [7,8]. We now report that pigeons are on par with humans when a task demands simultaneous processing resources; importantly, pigeons show faster responses than humans when sub-tasks are separated such that fast switches between processes are required.