Automatic medullary mechanisms ensure continuity of breathing and adaptation to metabolic fluctuations, right from birth and throughout life. However, breathing is unusual among autonomic functions in that it also responds to motor commands arising from complex cortico-subcortical networks comprising motor and premotor areas (particularly the supplementary motor area). These networks are engaged during voluntary respiratory manoeuvres (such as voluntary apnoea, voluntary sniffing, or voluntary hyperventilation)(1). They are also engaged when the respiratory system is used for non-respiratory purposes such as speech (2), and induced respiratory neuroplasticity experiments using repetitive transcranial magnetic stimulation have suggested that these networks exert a tonic excitatory influence on breathing during wakefulness (3). Inspiratory loading experiments in healthy volunteers have demonstrated that respiratory-related cortical networks are activated when the respiratory system is faced with a mechanical constraint (functional magnetic resonance imaging and electroencephalographic data), leading to the hypothesis of the cortical origin of the “paradoxical” hyperventilation associated with inspiratory loading during wakefulness in humans (4). In disease, respiratory-related cortical activity has been described in patients with deficient respiratory automaticity (congenital central alveolar hypoventilation), severe obstructive sleep apnoea syndrome (most probably related to inspiratory load due to upper airway abnormalities), and respiratory muscle weakness (amyotrophic lateral sclerosis), suggesting that these networks provide corticomedullary cooperation when automatic breathing is compromised or faced with a load-capacity imbalance. Of note, activation of respiratory-related cortical networks in response to experimental inspiratory or expiratory loading is concomitant with respiratory discomfort. In patients with diaphragm dysfunction, alleviating dyspnoea by mechanical ventilatory assistance silences respiratory-related cortical activity (5). These observations have led to the hypothesis that respiratory-related electroencephalographic activity could constitute a surrogate for self-reported dyspnoea in patients unable to directly communicate with their caregivers, and could form the basis for a patient-ventilator interface to improve the clinical management of mechanical ventilation (6).