A stressful stimulus will evoke a number of coincident neuroendocrine, autonomic and behavioural responses which serve both to minimize the potentially damaging effects of the threat and to restore homeostasis. Understanding how this integrated response is generated requires the elucidation of the neural pathways which transmit the stressful stimulus to the effectors systems, and the identification of transmitters that play pivotal roles in these pathways. This paper will present a conceptual framework that integration of the stress response depends on a hierarchy of interacting response networks, and that particular transmitters have a pre-eminent role in mediating or modulating network activation. Importantly these transmitters offer potential targets for pharmacological modulation of stress physiology. Numerous studies employing immediate-early gene expression or functional magnetic resonance imaging have provided evidence of the neural networks which are activated by different stress modalities and which serve to coordinate the different elements of the stress response. For example, as expected from the activation of the hypothalamo-pituitary-adrenal axis, c-fos mRNA or its protein product Fos can be detected in the hypothalamic paraventricular nucleus in response to both psychological (exteroceptive) and physical (introceptive) stress stimuli. However, these two modalities are differentiated by their activation of limbic and brainstem nuclei, notably areas of the amygdala and bed nuclei of the stria terminalis display marked modality-dependent activation. Nevertheless, while different stress modalities may utilize distinct neural pathways which converge on common outputs, each may be generalised to comprise two elements. Firstly, an afferent limb comprising either a cognitive or sensory system that responds to introceptive or exteroceptive stimuli which under certain conditions or above a particular threshold can take on a ‘stressful’ quality. Secondly, an efferent limb comprising a network which distributes this ‘stress signal’ to the various effector systems to generate an appropriate physiological or behavioural response. Diversity of stress responses and integration within this network is proposed to be achieved through a hierarchy of overlapping pathways, with some having the capacity for direct, rapid activation in response to an immediate threat to homeostasis, while higher order pathways provide a distributed signal to several output systems. It serves little purpose to identify the neurochemical phenotype of all the neurones which contribute either to the afferent or efferent limbs of these stress-activated networks. However, certain transmitters have been shown to have a pre-eminent role in regulating stress response, either acting to coordinate activation of diverse elements of the response (notable corticotropin-releasing hormone), or to attenuate activation of the response network and, thereby, have properties of endogenous anti-stress agents (e.g. oxytocin or GABA). It is suggested that the organisation of the response-activating network enables these key transmitters to fulfil important roles in integrating the stress response or in modulating its magnitude. Both classes of transmitter are particularly important as they may play important roles in the aetiology of stress-related disorders and, consequently have become the focus for pharmacological intervention. Furthermore, the involvement of transmitter receptor subtypes opens the possibility for the development of selective ligands for modulating the diverse responses to stress. In this respect an on-going challenge is to translate the knowledge of transmitter involvement into clinically effective therapies.