While some autoimmune disorders remain extremely rare, others largely predominate the epidemiology of human autoimmunity. Notably, these include psoriasis, diabetes, vitiligo, thyroiditis, rheumatoid arthritis and multiple sclerosis. Thus, despite the quasi-infinite number of "self" antigens that could theoretically trigger autoimmune responses, only a limited set of antigens, referred here as superautoantigens, induce pathogenic adaptive responses. Several lines of evidence reviewed in this paper indicate that, irrespective of the targeted organ (e.g. thyroid, pancreas, joints, brain or skin), a significant proportion of superautoantigens are highly expressed in the synaptic compartment of the central nervous system (CNS). Such an observation applies notably for GAD65, AchR, ribonucleoproteins, heat shock proteins, collagen IV, laminin, tyrosine hydroxylase and the acetylcholinesterase domain of thyroglobulin. It is also argued that cognitive alterations have been described in a number of autoimmune disorders, including psoriasis, rheumatoid arthritis, lupus, Crohn's disease and autoimmune thyroiditis. Finally, the present paper points out that a great majority of the "incidental" autoimmune conditions notably triggered by neoplasms, vaccinations or microbial infections are targeting the synaptic or myelin compartments. On this basis, the concept of an immunological homunculus, proposed by Irun Cohen more than 25 years ago, is extended here in a model where physiological autoimmunity against brain superautoantigens confers both: i) a crucial evolutionary-determined advantage via cognition-promoting autoimmunity; and ii) a major evolutionary-determined vulnerability, leading to the emergence of autoimmune disorders in Homo sapiens. Moreover, in this theoretical framework, the so called co-development/co-evolution model, both the development (at the scale of an individual) and evolution (at the scale of species) of the antibody and T-cell repertoires are coupled to those of the neural repertoires (i.e. the distinct neuronal populations and synaptic circuits supporting cognitive and sensorimotor functions). Clinical implications and future experimental insights are also presented and discussed.