Patients frequently describe difficulty concentrating, losing words mid-sentence, forgetting what they went upstairs for, or finding that work that used to feel straightforward now requires considerably more effort. They often attribute this to getting older, to stress, or to their phones. What they have not considered is that they may be experiencing the measurable neuropsychological consequences of interrupted sleep architecture — and that these consequences are not simply the tiredness that anyone feels after a bad night.
OSA-related cognitive impairment is distinct from ordinary fatigue. It is structural, measurable on objective tests, and in some domains appears to reflect genuine neurological change rather than simply functional underperformance driven by tiredness. Understanding what the research shows, domain by domain, helps make sense of why the clinical picture looks the way it does.
How Many People with OSA Have Measurable Cognitive Impairment?
A 2025 meta-analysis covering 23 studies and more than 33,000 individuals found that 37 per cent of adults with OSA — more than 1 in 3 — have measurable cognitive impairment on standardised neuropsychological testing. In those with severe OSA (defined as more than 30 apnoea events per hour of sleep), that rises to 44 per cent — nearly 1 in 2. For context: the population prevalence of mild cognitive impairment in adults over 60 is estimated at around 15 to 20 per cent. OSA patients are experiencing it at twice the background rate, and many of them are younger.
Women with OSA show higher rates than men — 59 per cent versus 47 per cent — a finding that is probably related to both the diagnostic delay women experience (meaning more time with untreated OSA before testing) and genuine sex differences in neurobiological vulnerability to intermittent hypoxia.
Which Cognitive Domains Are Affected — and How Badly?
When researchers pool results from many studies, they express how far apart two groups are using a measure called an effect size, or standardised mean difference (SMD). An SMD of 0.2 is considered small, 0.5 is medium, and 0.8 is large. These measures are not intuitive on their own, so I will translate each one into practical terms.
Sustained attention — the ability to remain focused on a task over time — shows an SMD of about -0.5, which is a medium-to-large impairment. In practical terms, this means OSA patients perform substantially worse on tasks requiring maintained vigilance than comparable healthy adults. This is the cognitive domain most responsible for the driving risk associated with OSA, and it is also the domain most relevant to sustained work performance, particularly in jobs requiring monitoring, quality checking, or continuous operational awareness.
Working memory — the ability to hold and manipulate information in mind over seconds to minutes — shows a medium-sized impairment (SMD around -0.4). In everyday terms, this is what determines whether you can keep track of a phone number while dialling it, follow the thread of a complex conversation, or hold the previous steps of a task in mind while completing the next one. When this capacity is reduced, people experience it as mental fogginess, losing their train of thought, or needing to re-read the same paragraph multiple times.
Processing speed — how quickly the brain can take in and respond to information — is also significantly impaired (SMD around -0.45). A slowing of processing speed is often the earliest and most easily noticed cognitive change in adults with OSA, and it is frequently misinterpreted as a side effect of ageing. But when a 45-year-old notices that things feel slower cognitively than they did at 35, and they have untreated OSA, the OSA is a more parsimonious explanation than two decades of biological ageing for a loss that should not be this pronounced yet.
Executive function — planning, decision-making, cognitive flexibility, and error correction — shows a small-to-medium impairment (SMD around -0.3 to -0.4). This is the domain that determines how well someone manages complex projects, adapts to unexpected problems, or inhibits impulsive responses under pressure. In workplaces where these skills are critical, even a small reduction in executive function can have meaningful consequences for professional performance.
OSA patients perform significantly worse on objective tests of attention, memory, processing speed, and executive function. These are not simply tired people performing below their best. They are people whose brains are measurably operating below their structural potential.
Why Does OSA Impair Cognition?
Three mechanisms operate in parallel. First, the direct effects of intermittent nocturnal hypoxia on brain tissue — repeated oxygen drops cause oxidative stress and mitochondrial dysfunction in neurons, particularly in the hippocampus (the region most critical to memory formation) and the prefrontal cortex (the region most critical to attention and executive function). Neuroimaging studies show measurable reductions in grey matter density in these regions in severe OSA patients compared with healthy controls.
Second, sleep fragmentation prevents the memory consolidation that depends on slow-wave and REM sleep. Memory consolidation — the process by which experiences from the day are transferred from temporary to long-term storage — occurs largely during sleep. When sleep architecture is disrupted, consolidation fails, and memories that should have been stored durably are instead weakly encoded or lost. This is why OSA patients frequently report good recall of events from years ago but poor recall of conversations, commitments, or information encountered in the past few days.
Third, the cardiovascular consequences of OSA — endothelial dysfunction, arterial stiffness, small vessel disease — affect cerebral blood flow and white matter integrity over time. White matter lesions, visible on MRI, are more common in severe OSA patients and are themselves associated with processing speed impairment and increased dementia risk. This is a structural brain change that accumulates with years of untreated disease.
Can CPAP Reverse Cognitive Impairment?
The honest answer is: partly, and not reliably across all domains. A meta-analysis of 14 randomised controlled trials found that CPAP treatment produced statistically robust improvements in attention and processing speed in patients with severe OSA — the changes were large enough to be unlikely due to chance. However, the same analysis found that improvements in memory and executive function were less consistent and, in some studies, modest.
The pattern in the literature is that the cognitive deficits most reliably reversed by CPAP are those driven primarily by sleep fragmentation and daytime sleepiness — sustained attention and processing speed are in this category. The deficits most resistant to reversal appear to be those driven by structural neurological change from cumulative hypoxia — particularly aspects of working memory and some executive functions. This is not a universal finding, and some studies report broader recovery, but the weight of evidence suggests that earlier treatment produces better cognitive outcomes than treating the condition after many years.
A 2020 meta-analysis found that CPAP improved processing speed by an SMD of 0.43 — a medium-sized improvement that represents a genuine and clinically perceptible change in mental efficiency. Sustained attention improved by an SMD of 0.36. These are meaningful numbers: they suggest that many of the cognitive difficulties experienced by people with undiagnosed OSA are substantially reversible once the airway is treated, particularly when diagnosis and treatment happen before structural brain changes are advanced.
The implication is clear: earlier diagnosis and treatment matters not only for cardiovascular and metabolic outcomes but for the long-term functional integrity of the brain.
Daytime Sleepiness and Its Consequences
Excessive daytime sleepiness is one of the primary cognitive complaints of OSA patients, and it is worth distinguishing it from the deeper cognitive impairments described above because it responds particularly well to treatment and has its own significant consequences. The Epworth Sleepiness Scale (ESS) is the standard clinical questionnaire used to measure subjective sleepiness; a score above 10 out of 24 is considered excessive.
CPAP reduces ESS scores by an average of 2.5 to 3 points in meta-analyses of randomised trials — a change large enough to move most patients from the clinically significant range into the normal range. This improvement in alertness has direct implications for driving safety, workplace performance, and quality of life.
In patients with severe OSA, the degree of daytime sleepiness can reach levels that would be clinically indistinguishable from narcolepsy — a defined neurological sleep disorder — on objective testing. Yet many of these patients have no formal sleep disorder diagnosis, have never been tested, and are managing their lives around cognitive limitations they have come to regard as their normal baseline. For them, a successful treatment outcome does not mean returning to how they felt some years ago — it means discovering, sometimes with considerable surprise, how they were actually supposed to feel.
References
[1] Su K et al. Prevalence of cognitive impairment among adults with obstructive sleep apnea: a systematic review and meta-analysis. Sleep and Breathing. 2025. 23 studies, 33,226 individuals; pooled prevalence 36.9% in OSA; 44.5% in severe OSA; women 59% vs men 47%.
[2] Olaithe M et al. Cognitive deficits in obstructive sleep apnea: insights from a meta-review and comparison with deficits observed in COPD, insomnia, and sleep deprivation. Sleep Medicine Reviews. 2018. Meta-review of 7 meta-analyses; SMD -0.50 attention, -0.43 memory, -0.48 psychomotor speed, -0.31 executive function in OSA patients.
[3] Wang MY et al. Cognitive effects of treating obstructive sleep apnea: a meta-analysis of randomized controlled trials. Journal of Alzheimer's Disease. 2020. 14 RCTs, 1,926 participants; CPAP significantly improved attention (SMD 0.36) and processing speed (SMD 0.43); improvements in memory and executive function less consistent.
[4] Pase MP et al. Sleep architecture, obstructive sleep apnea, and cognitive function in adults. JAMA Network Open. 2023. Sleep and Dementia Consortium, 5,946 adults from 5 US population cohorts; even mild OSA significantly associated with worse global cognition over 5-year follow-up.
[5] Gagnon K et al. Neuropsychological, neurophysiological, and sleep characteristics of obstructive sleep apnea subtypes. Journal of Sleep Research. 2022. Neuroimaging evidence of reduced grey matter density in hippocampus and prefrontal cortex in severe OSA patients.
[6] Vaessen TJA et al. Cognitive complaints in obstructive sleep apnea: a systematic review. Sleep Medicine Reviews. 2015. Systematic review; strongest and most consistent deficits in sustained attention and psychomotor vigilance; memory deficits prominent in long-term untreated OSA.
Contact Professor Vik Veer
If you would like to discuss a sleep assessment or arrange a consultation, please use the details below.
Private secretary: 0207 458 4584