The current COVID-19 pandemic led to a considerable reduction in in-person social contacts all over the world. In most individuals, reduced social contacts lead to the perception of social isolation causing feelings of loneliness, which are experienced as stressful. Experiencing social distress due to actual or perceived social isolation has been associated with negative health outcomes such as depression, (cerebro-) vascular disease and mortality. Concrete mechanisms behind this association are still a matter of debate. A group of researchers around Hugo Critchley with special contributions of Sarah Garfinkel and Lisa Quadt proposes a framework for the underlying brain-body interactions including elements from models of social homeostasis and interoceptive predictive processing that provides important insights and testable pathways. While in a previous publication, we reviewed literature on the observed association between social isolation and stroke and coronary heart disease, we now extent this review by presenting a comprehensive model to explain underlying pathomechanisms from the perspective of social neuroscience. Further, we discuss how neurodivergent people, e.g. autistic individuals or persons with attention deficit disorders, might differ in these pathomechanisms and why they are especially vulnerable to social isolation. Finally, we discuss clinical implications for the prevention and therapy of (cerebro-) vascular diseases.