Alzheimer's disease (AD) is 10-fold more frequent than Parkinson's disease (PD), which in turn is 10-fold more frequent than amyotrophic lateral sclerosis (ALS). The differences between these neurodegenerative diseases have been ascribed to a selective vulnerability of specific neuronal sub-types that then determine each disorder. However, there are non-neuronal cells that are ubiquitously and possibly primarily involved in all of them, and they share regulatory mechanisms through similar interneurons and, typically inhibitory, neurotransmitters. There is recognized clinical and neuropathological overlap between AD, PD and ALS, the best example being Guamanian Lytico-Bodig, but increasingly recognized in larger populations, e.g. carriers of C9orf72 hexanucleotide expansions. From early embryogenesis to adulthood, genetic and experience-dependent functional neural networks develop primarily in relation to the neocortex. From an evolutionary standpoint, cognition, memory, executive function, linguistics and fine motor function are most prominent in humans. It is concluded that neural networks, rather than specific neuronal sub-types defined regionally or by individual transmitters, underlie the marked differences between neurodegenerative disorders in terms of susceptibility and clinical features. This requires the continued development of strategies to study brain function in health and disease as the 'system', greater than the sum of its parts.