The coupling of non-neural internal states essential to an agent’s survival with artificial nervous systems can increase adaptivity in terms of (1) exploitation of sensorimotor possibilities, (2) regulation of internal and behavioural activity and (3) behavioural emergence via complex network dynamics that enable the agent to contend with a challenging and unpredictable world. This paper provides a review of recent research on the relevance of non-neural internal states to adaptive behaviour in the field of adaptive robotics. The paper derives a methodological approach that promises to further extend our understanding of how non-neural internal states can increase adaptivity in robots as relevant to the proposed core benefits extracted.