AINEWS
The Hidden Cost of Your Daily Coffee Isn't Insomnia. It's Shallow Sleep.
A new review of brain wave studies finds caffeine can flatten the deepest stage of sleep, leaving you rested in body but unrefreshed the next day.
A new review of brain wave studies finds caffeine can flatten the deepest stage of sleep, leaving you rested in body but unrefreshed the next day.
You might be sleeping eight hours a night and still waking up with the kind of fatigue no amount of caffeine seems to fix. A new review of sleep neuroscience suggests the problem is not how long you sleep, but what is happening inside that sleep: caffeine, the researchers argue, can quietly flatten the deepest, most restorative stage of sleep even when it does not keep you awake at all.
The paper, a review of EEG-based caffeine research described by its authors as systematic and mechanistic, published 12 April 2026 in the journal Nutrients by James Chmiel and Prof. Donata Kurpas of Wroclaw Medical University, looks at how the brain behaves under caffeine influence during sleep rather than asking only whether caffeine delays sleep onset. EEG, the brain-wave recording used in sleep clinics, gives researchers a way to see stages of sleep that the sleeper cannot report, and slow-wave activity is the electrical signature of the deepest stage of non-REM sleep, the one most associated with physical restoration, memory consolidation, and next-day cognitive performance.
What the review finds, as summarized in a SciTechDaily write-up of the work, is that the sleeping brain under caffeine influence can shift toward a more wake-like EEG pattern, with reduced slow-wave activity even when total sleep duration looks normal. That matters because a person can fall asleep easily, sleep a full night, and not remember any awakenings while the EEG still shows degraded deep-sleep features. Subjective "I slept fine" reports, the authors note, often do not match the neurophysiology.
The mechanism is not mysterious. Caffeine is a competitive antagonist at adenosine receptors, blocking the brain's own sleepiness signal rather than acting as a stimulant per se. When adenosine signaling is blunted through the night, the slow-oscillation patterns that define restorative sleep are attenuated. The review pulls together human EEG studies showing exactly this pattern: shorter and less powerful slow waves, more high-frequency wake-like intrusions, and a flatter overall architecture of deep sleep.
The effect, however, is not a cliff. The research base Chmiel and Kurpas synthesize is consistent in direction but modest in size for habitual users, and inter-individual variability is large. Response to caffeine is driven by genetics (notably CYP1A2 metabolism), age, chronic stress, sleep debt, hormonal status, and prior caffeine exposure. People who drink coffee daily develop partial tolerance, which is why the dramatic "couldn't sleep at all" stories are not the typical experience and not the one the EEG literature actually documents. The honest version of the finding is: caffeine can shave depth off your deepest sleep in a way you will not feel directly, and the size of that shave depends on who you are.
The framing Kurpas offers, in comments reported by SciTechDaily, is that the relevant variable is often total daily caffeine load plus the time available to metabolize it before bed, not only whether you have a late cup. Caffeine's half-life in most adults sits around five hours, but in slow metabolizers, on a high daily dose, or under sleep deprivation, effective exposure can stretch well past midnight. Two espressos at 2 p.m. plus a black tea at 5 p.m. is not "no afternoon caffeine" in physiological terms.
She also names a vicious cycle the data hint at: daytime caffeine boosts focus and function, nighttime sleep becomes shallower, the next day's fatigue is worse, and the need for stimulation goes up. Demanding cognitive work, athletic training, and habitual reliance on caffeine for performance all sit in this loop, and none of them is solved by "one less cup" in isolation. The relevant question for those readers is not whether they can fall asleep after coffee. It is whether the sleep they are getting is the kind that pays back the next morning.
Caffeine is a biologically active substance whose effects depend on dose, timing, age, lifestyle, baseline sleep quality, stress burden, and individual sensitivity, the authors write. That is not a verdict. It is a better set of inputs for a decision most people are already making on autopilot. The next test, if you suspect this is you, is simple and free: hold caffeine to before noon for two weeks, keep sleep timing constant, and pay attention not to how fast you fall asleep but to how restored you feel on waking. The brain's deepest waves will not tell you they were flattened. That is the entire point of the EEG data.