Obstructive sleep apnoea (OSA) is characterized by repeated collapse of the upper airway during sleep, producing bouts of intermittent hypoxia (IH). Apnoeas are usually followed by arousal and hyperventilation, producing hypocapnia. During sleep, hypocapnia reduces upper airway dilator muscle activity, rendering the airway susceptible to successive collapse. IH can induce a form of neuroplasticity, called long term facilitation (LTF), within ventilatory motor neurons such as the phrenic and hypoglossal nerves, producing a sustained increase in efferent activity, and also increase carotid body hypoxic sensitivity. As hypoglossal LTF increases upper airway dilator muscle activity, the Mateika laboratory have proposed that IH may be used as a therapeutic treatment for OSA. However, hypocapnia suppresses both phrenic and hypoglossal LTF. Therefore, hypoglossal LTF is unable to maintain airway patency during post-obstructive hypocapnia. Rather than being protective, Mateika and colleagues have shown that IH-induced ventilatory neuroplasticity worsens OSA (Yokhana et al. 2012; Syed et al. 2013). Yokhana et al. (2012) found that although IH induced both ventilatory LTF and increased hypoxic sensitivity, OSA worsened, as evidenced by increased apnoea-hypopnea index (AHI), apnoea duration and ratio of apnoeas to hypopneas, changes which are indicative of increased instability in ventilatory control. The stability of the ventilatory chemoreflex negative feedback system is quantified with the engineering principle of loop gain (LG), calculated as the ratio of the ventilatory response to the disturbance that elicited it. The higher the LG, the greater the ventilatory response and the greater the ensuing hypocapnia and upper airway muscle hypotonia. Thus, increased LG exacerbates OSA. The overall LG of the system is the product of controller and plant gains, being the gain of the ventilatory response to blood gases and the effectiveness of the lungs to alter blood gases, respectively. The report by Alex et al. (2019) published in the current issue of The Journal of Physiology is further analysis of the data published by Yokhana et al. (2012) confirming that OSA severity increases because IH increases LG. Whether increased hypoxic sensitivity or ventilatory LTF contributed to increase LG was not determined, although the authors suggest controller gain increased due to increased hypoxic sensitivity. However, ventilatory LTF may have contributed, as previous work from this laboratory has reported that even when increased hypoxic sensitivity is not expressed, ventilatory LTF increases AHI (Syed et al. 2013). How ventilatory LTF increases AHI is complex. Controller gain is often assumed to be synonomous with chemoreceptor sensitivity, but this is not entirely correct. Increased minute ventilation characteristic of ventilatory LTF reduces , shifting eupnoea left along the metabolic hyperbola. With no change in the apnoeic threshold, the CO2 reserve (difference in CO2 between eupnoea and apnoea) decreases (Chowdhuri et al. 2010). From an engineering perspective, this increases the slope between eupnoea and apnoea and increases controller gain. This has been reported as an increase in central chemoreceptor sensitivity to CO2 (Chowdhuri et al. 2010). However, phrenic LTF is not due to neuroplasticity at either the central or peripheral chemoreceptors. Consequently, phrenic LTF does not alter chemoreceptor sensitivity to CO2. IH can increase the sensitivity of central chemoreceptors to CO2. However, this causes chemoreceptors to be active at lower , thereby reducing the apnoeic threshold (Molkov et al. 2011). With no change in eupnoea, the slope between apnoea and eupnoea decreases (Katayama et al. 2007). Therefore, below eupnoea, increased central chemoreceptor sensitivity to CO2 can reduce controller gain. IH does not appear to alter the apnoeic threshold in humans (Chowdhuri et al. 2010). Consequently, the finding of increased LG by Alex et al. (2019) is likely to be due to both forms of neuroplasticity; by reducing ventilatory LTF reduces the CO2 reserve and increases controller gain below eupnoea, while increased hypoxic sensitivity increases controller gain above eupnoea. As IH-induced LTF is being investigated as a treatment to regain limb function following spinal cord injury, the authors’ advice to test patients for OSA and administer continuous positive airway pressure (CPAP) to prevent deleterious effects of increased LG are wise. Interestingly, despite their results supporting the hypothesis that IH pathogenically increases LG and worsens OSA, Alex and colleagues maintain that IH may have therapeutic potential in OSA, based on their prior data showing that hypoglossal LTF allowed CPAP pressures to be reduced in OSA patients. However, neuroplasticity decays over the course of hours to days, such that IH would have to be administered daily to maintain LTF. Chronic IH induces systemic oxidative stress and inflammation, which can contribute to a multitude of life-threatening conditions such as cardiovascular disease, metabolic disorder and neurocognitive deficits. Therefore, the mild benefits of reduced CPAP pressures are potentially outweighed by the many pathogenic consequences of chronic IH. Alex et al. (2019) also found that arousal threshold increased and suggest this may be a protective neuroplastic mechanism. Previous studies support the postulate that increased sleep pressure/fatigue increases arousal threshold, and even without sleep disruption, IH causes severe neurocognitive deficits including debilitating fatigue. Although Yokhana et al. (2012) found that arousal frequency and total sleep time did not change following IH, it is possible that increased arousal threshold is a symptom of neurocognitive deficit and fatigue, rather than an adaptive form of neuroplasticity. Further work will be required to determine if IH can increase the arousal threshold, without inducing neurocognitive deficits and fatigue. Although the Mateika laboratory maintain advocacy of IH as a potential therapeutic treatment for OSA, the results of this study are in agreement with the bulk of the literature published by the Mateika laboratory, which supports the hypothesis that IH-induced neuroplasticity increases LG and OSA severity, and that the effects of experimentally administered IH induce the same pathogenic consequences as IH experienced during untreated OSA. These results are critical in advancing our understanding of the mechanisms of IH in OSA pathophysiology. No competing interests declared. Sole author. None.