Fifty years have passed since social anxiety disorder (SAD) was first differentiated from other phobias. In the years since research has largely aligned with the zeitgeist of categorical classificatory frameworks, and has spanned identifying causes, maintenance factors and innovative interventions. Despite significant advances in the field, the capacity to conceptualise SAD as an independent entity is limited given the heterogeneity and dimensionality of diagnostic criteria, high rates of comorbidity, and non-specificity of aetiological mechanisms, maintaining factors and approaches to treatment. The Research Domain Criteria (RDoC) initiative was developed in an effort to overcome the inherent limitations posed by descriptive diagnostic systems – particularly in terms of reliability and validity – and in doing so seeks to facilitate research into underlying pathophysiological and behavioural mechanisms that cut across traditional diagnostic boundaries. The RDoC framework is furnished with a ‘matrix’, which in essence corresponds to a set of research principles that attempt to reconcile neuroscience and psychopathology. This review outlines a rationale for integrating SAD research with the RDoC approach, and offers examples of how future studies may wish to frame hypotheses and design experiments as the field moves towards classifying dimensions of psychopathology through a mechanistic understanding of underlying neurobiological and behavioural processes.