Identifying and investigating emergent behaviour in neural systems

The concept of emergence is key to understanding much of neurophysiology, neurodevelopment, and neurodegeneration. Many brain functions are underpinned by behaviour that emerges from interactions of components such as neurons, synapses, and glia. Perturbations such as ion channel mutations or trauma can lead to new types of, often detrimental, emergent behaviour. Typical approaches to neuroscience involve studying neural behaviour in a range of in vitro, in vivo, and in silico models designed to capture the key components of the brain region under investigation. Translating from these models back to the brain requires that the component interactions within the models can be mapped onto interactions in the brain. However, in many cases, this may not hold, leading to emergent behaviour that is not derivable from knowledge of the brain components in relative isolation. We thus need a framework that respects the prospective system-dependent differences in interactions. In this manuscript, we distil theoretical approaches to emergence into a practical framework for assessing system behaviour focussed on "novelty" as a necessary condition for emergence and identify a range of ways in which novelty can arise. We consider how different approaches for investigating system behaviour can be applied in different contexts and when these might fail. We provide examples of different forms of novelty in neuroscience contexts and then apply the framework to understanding the specific case of epilepsy. We end by outlining a set of guiding principles for designing experiments and models for understanding emergent behaviour in neural systems.