TO THE EDITOR: Obstructive sleep apnea (OSA) results in awake neurocognitive impairment with societal implications (22). However, the relevant pathophysiological mechanism(s) for this impairment are unclear, impacting on disease monitoring and targeted treatment (1, 17). Among a number of candidate factors, hypoxemia and sleep fragmentation are generally recognized as two key underlying mechanisms (1, 4), with hypoxemia having the strongest evidence base (7, 17). However, recent large clinical trials failed to find a strong link between nocturnal hypoxia and neurobehavioral impairment. In the landmark APPLES study, the severity of O 2 desaturation explained only Ͻ2% of the variance of neurocognitive performance improvement in 1,204 OSA patients, and arousal index was not a significant predictor (17). In the recent LATINO study with the largest sample size of 8,059, apnea hypopnea index (AHI) with 3% desaturation was very weakly associated with neurocognitive functions after controlling for confounders ( ranged from 0.003 to 0.005) (18). In a multicenter European study, 1,649 OSA patients with excessive daytime sleepiness (EDS) and 1,233 OSA patients without EDS had only 1% of difference in SpO 2 nadir, no difference in mean SpO 2 , and similar arousal index (37/h vs. 33/h) (20). In human experimental studies, neurocognitive function was measured in 11 healthy subjects at ground level and simulated altitude of 13,000 feet. Two weeks of nocturnal continuous hypoxia at altitude did not induce subjective sleepiness or impair objective vigilance and working memory (24). In addition, the effect of 4-wk nocturnal intermittent hypoxia on attention, working memory, Multiple Sleep Latency Test, and the Rey Auditory Verbal Learning Test was measured in 8 healthy subjects and again no effect was detected (29). Moreover, if hypoxemia is the dominant factor for neurocognitive impairment, then EDS should improve with supplemental O 2 therapy. However, studies have not found that supplemental O 2 improves daytime hypersomnolence in OSA patients despite improving oxygenation (13, 16).Interestingly, although OSA usually comes with repetitive episodes of both hypoxia and hypercapnia, only the effect of hypoxia has been more intensively studied. A simple test entering key words of "Sleep Apnea and Hypoxia" in PubMed returns 2,866 publications; In contrast, there are only 867 publications with the key words of "Sleep Apnea and Hypercapnia." It is therefore not surprising that hypercapnia has not been considered as a major factor in OSA-related neurocognitive impairment in any major review in this field (1,5,12).In this context, we conducted a series of clinical and experimental studies to evaluate the effect of hypoxia and hypercapnia on EEG activation and neurobehavioral performance. First, we demonstrated that 97 hypercapnic sleep-disordered breathing (SDB) patients have a significantly high degree of EEG slowing, which was best predicted by increased wake PCO 2 measured by arterial blood gases. Hypoxia-related parameters were not...