What If We Fixed Alpha-Intrusion? A Thought Experiment in ME/CFS
Nobody has run this experiment yet. But the question is worth asking anyway.
Alpha-delta sleep — the intrusion of waking-frequency alpha waves into deep sleep — is one of the most documented objective abnormalities in ME/CFS and fibromyalgia (Moldofsky et al. 1975, Psychosomatic Medicine (Moldofsky et al. 1975); Roizenblatt et al. 2001, Arthritis & Rheumatism (Roizenblatt et al. 2001)). It prevents the brain from entering the slow-wave oscillations that drive overnight restoration. A previous article in this series described the thalamic calcium channel dysfunction hypothesised to cause it (not yet directly tested in ME/CFS patients).
Suppose a drug, a brain stimulation protocol, or some future intervention could cleanly eliminate that intrusion and restore pure delta oscillations during slow-wave sleep. What would change? What wouldn’t? The answer turns out to reveal something important about why ME/CFS resists single-target interventions — and it isn’t a particularly encouraging picture.
1 What would improve
The first and most direct benefit would be partial restoration of glymphatic clearance. Delta oscillations are the primary driver of glymphatic flow — the brain’s waste-clearance system that flushes metabolic debris during deep sleep (Xie et al. 2013, Science (Xie et al. 2013)). The spaces between brain cells expand during delta sleep, allowing cerebrospinal fluid to flow through and carry out accumulated waste. With alpha intrusion eliminated and sustained delta epochs restored, this clearance system would regain its driving force. The result: less accumulation of inflammatory mediators, misfolded proteins, and metabolic waste in brain tissue. The morning “hangover” feeling — that toxic, poisoned sensation patients describe on waking — would likely improve. Morning headaches driven by impaired overnight clearance would likely decrease.
Memory consolidation would also improve. Sleep spindles (12–14 Hz thalamic bursts) couple with hippocampal sharp-wave ripples during the up-states of slow oscillations, transferring memories from short-term hippocampal storage to long-term cortical storage (Diekelmann & Born 2010, Nature Reviews Neuroscience (Diekelmann and Born 2010); Staresina et al. 2015, Nature Neuroscience (Staresina et al. 2015)) — but this coupling requires intact slow-wave oscillations as the scaffold. With delta waves restored, one component of ME/CFS cognitive dysfunction, the difficulty consolidating new information, should improve.
The neuroinflammatory vicious cycle would partially break. Impaired glymphatic clearance leaves inflammatory mediators in the brain, which activate the NLRP3 inflammasome in microglia, producing cytokines that further disrupt sleep architecture. Better delta sleep and improved clearance would remove one of the inputs driving this loop — poor sleep → waste accumulation → neuroinflammation → worse sleep — though it would not eliminate the loop entirely.
Finally, pain thresholds might improve. During normal slow-wave sleep, delta oscillations actively suppress nociceptive pathway activity — the brain’s pain circuits are downregulated via thalamic gating, and this is part of what makes sleep restorative. Alpha intrusion leaves those pain pathways partially active throughout the night. Restoring delta might restore this overnight suppression, improving morning pain and potentially reducing central sensitisation over time.
2 What would not improve
Here is where the thought experiment earns its keep.
Even with perfect delta oscillations, the aquaporin-4 water channels that form the molecular gateway for glymphatic exchange remain suppressed by elevated nocturnal norepinephrine — a consequence of autonomic dysfunction, not sleep architecture (Xie et al. 2013 (Xie et al. 2013)). You can restore the driving force while the plumbing stays partially blocked. Glymphatic clearance would improve but not normalise.
The kynurenine pathway keeps running regardless. The tryptophan-to-kynurenine shunt — driven by chronic immune activation via IDO1/IDO2 enzymes — produces quinolinic acid, an NMDA receptor agonist that is directly excitotoxic to neurons. ME/CFS patients show significantly elevated quinolinic acid and reduced neuroprotective KA/QA ratios (Groven et al. 2021, Psychoneuroendocrinology (Groven et al. 2021)). Sleep oscillations have no bearing on this pathway. The brain fog driven by quinolinic acid-mediated excitotoxicity persists at whatever level it was at before.
Mast cell-mediated cognitive impairment is similarly untouched. Histamine released from activated mast cells crosses into the brain and acts on H3 receptors to suppress acetylcholine release, directly impairing attention, working memory, and executive function. Sleep oscillations are not part of that circuit.
Cerebral hypoperfusion continues. ME/CFS patients show reduced cerebral blood flow — 26% reductions documented during orthostatic challenge versus 7% in controls (van Campen et al. 2020, Clinical Neurophysiology Practice (Campen et al. 2020)) — driven by autonomic dysfunction, endothelial dysfunction, and reduced blood volume. The brain’s oxygen supply is limited regardless of how well it sleeps. Cognitive tasks that demand increased blood flow would still fail at the supply level.
The mitochondrial picture does not change either. The energy deficit that drives ME/CFS — impaired oxidative phosphorylation, electron transport chain dysfunction, elevated mitochondrial ROS — is a peripheral and systemic problem that sleep architecture alone cannot resolve. Better sleep might modestly improve overnight mitochondrial repair, but the fundamental metabolic defect persists.
And the autonomic dysfunction that caused much of the sleep problem in the first place remains. The sympathetic-parasympathetic imbalance, orthostatic intolerance, inappropriate tachycardia, and vascular dysregulation that characterise ME/CFS are upstream of sleep architecture, not downstream of it. The “tired but wired” state at bedtime, the nocturnal awakenings, the night sweats — all of these would likely persist even with improved delta oscillations during the sleep that does occur.
3 The partial improvement problem
Put the predictions together and you get: significant but incomplete improvement in brain fog, moderate improvement in pain, modest improvement in fatigue, and essentially no change in autonomic symptoms, PEM, or exercise intolerance.
That asymmetry illustrates something fundamental about ME/CFS that most treatment discussions fail to acknowledge. The disease has multiple independent pathological mechanisms running simultaneously, and they do not share a single upstream cause. They share a common trigger — in many cases, a viral infection — and a common sustaining environment of immune dysregulation and metabolic failure, but once established, the individual mechanisms run on partially independent tracks.
Brain fog is the clearest example. It has at least three independent mechanistic drivers: glymphatic clearance failure (sleep-dependent, potentially fixable by restoring delta); kynurenine pathway excitotoxicity (immune-dependent, unaffected by sleep); and mast cell histamine suppression of acetylcholine via H3 receptors (mast cell-dependent, also unaffected by sleep). Fixing the first of those helps. The patient may feel substantially better — but not well. This is exactly the partial-improvement pattern clinicians actually observe: each effective intervention moves something in the right direction without resolving the disease.
4 What not to do — and why standard advice backfires
Most sleep advice for ME/CFS patients is borrowed from the general insomnia literature, where it works. In ME/CFS, where sleep dysfunction has a neurological and autonomic basis alongside any behavioural component, some standard recommendations need modification — not because behavioural sleep medicine has nothing to offer, but because specific components interact badly with the underlying physiology.
Sleep restriction therapy is the clearest case. CBT-I protocols typically reduce time in bed to build “sleep pressure” — mild sleep deprivation that consolidates fragmented sleep. In ME/CFS, that logic fails at the first step: the energy deficit means sleep deprivation does not build homeostatic sleep pressure normally, because the adenosine mechanism is already dysfunctional (Porkka-Heiskanen et al. 1997, Science (Porkka-Heiskanen et al. 1997); covered in the earlier article in this series). Instead, sleep restriction depletes an already bankrupt energy reserve, can trigger post-exertional malaise, and worsens the autonomic dysfunction driving the sleep problem. The NICE guideline for ME/CFS (2021, updated (National Institute for Health and Care Excellence 2021)) explicitly recommends personalised sleep management rather than sleep restriction. If a sleep clinician proposes CBT-I, the sleep restriction component should be discussed explicitly — patients should ensure their clinician understands that ME/CFS-specific physiology makes sleep deprivation counterproductive. Other CBT-I components are a different matter: stimulus control can address conditioned arousal at bedtime (a real phenomenon that layers on top of the organic sleep dysfunction), cognitive restructuring may help with the anxiety that accumulates around sleep, and sleep scheduling can support circadian alignment. The problem is specifically sleep restriction, not behavioural sleep medicine as a whole.
The “no naps” rule is wrong in ME/CFS, though uncontrolled napping is its own problem. Standard insomnia advice prohibits daytime naps to preserve nighttime sleep drive. Naps in ME/CFS serve a different purpose: horizontal rest reduces cardiac preload demands (important with orthostatic intolerance), decreases skeletal muscle energy expenditure, and may improve cerebral perfusion even when the patient does not actually fall asleep. The catch is that ME/CFS naps are also not restorative in the way a healthy person’s nap would be, because the adenosine-clearance mechanism that makes naps refreshing is itself impaired. The practical guidance is controlled naps — early afternoon (13:00–15:00), 15–20 minutes maximum, alarm set. Late naps risk shifting the already-fragile circadian phase. And if afternoon sleepiness follows unusual exertion from the previous 24–72 hours, it may be the prodrome of PEM rather than ordinary tiredness; pacing assessment is more appropriate than sleep.
Nap length is where real damage accumulates. Past 30–40 minutes, the brain enters slow-wave sleep. In a healthy person, waking from that state produces sleep inertia that clears in 15–30 minutes; in ME/CFS the same transition takes 2–4 hours, because the thalamic coordination required is itself energy-limited. A 90-minute nap that enters and exits a full sleep cycle can leave the patient functionally incapacitated for the rest of the day. Very long naps — 2–3 hours or more — compound this further by stealing slow-wave sleep from the night. The total daily slow-wave budget is finite, and the daytime portion is alpha-contaminated anyway, so the patient ends up with two blocks of poor-quality sleep instead of one.
The exception is important: during a crash or severe PEM episode, the body may demand extended sleep of 12–16 hours. That is not a nap — it is a physiological emergency response, and fighting it is counterproductive. The relevant distinction is between habitual long daytime sleep (a pattern worth discussing with a clinician, as it may signal undertreated sleep architecture dysfunction or unrecognised chronic overexertion) and crash-driven hypersomnia (the body enforcing rest to limit further damage).
Zolpidem deserves specific mention because it may worsen the problem it was prescribed for. As described in the previous article, in mouse models, zolpidem suppresses the norepinephrine oscillations that drive glymphatic clearance during sleep (Hauglund et al. 2025, Cell (Hauglund et al. 2025); human validation pending). If this translates to humans, the patient sleeps longer but the brain’s waste-clearance system does not activate properly. The morning “toxic” feeling may persist or worsen despite the increased total sleep time. Important: patients currently prescribed zolpidem or any Z-drug should not self-discontinue. Abrupt withdrawal carries risks including rebound insomnia and, in longer-duration users, seizure. Any change should be discussed with the prescriber.
Two other habits deserve brief mention. Forcing oneself to stay awake all day to “save sleep for tonight” costs more metabolically than it gains: the energy of keeping a depleted system upright and active for hours typically exceeds the circadian cost of a brief early-afternoon rest, and patients often arrive at bedtime in worse metabolic shape with paradoxically worse sleep quality. Alcohol is similarly counterproductive — it promotes initial sedation but fragments the second half of the night, suppresses REM sleep, worsens autonomic instability, and impairs glymphatic function. In a system where every minute of quality slow-wave sleep matters, its net effect is strongly negative.
5 Current pharmacological approaches to alpha-delta sleep
Several existing medications partially target alpha intrusion, though none eliminates it cleanly. The dose ranges below reflect published clinical use — they are not prescribing guidance. All require physician evaluation and monitoring, particularly given the cardiac, autonomic, and drug-interaction risks common in ME/CFS. Do not start, stop, or adjust any medication based on this article.
Gabapentin and pregabalin are the most commonly prescribed. They bind alpha-2-delta voltage-gated calcium channel subunits, increasing slow-wave sleep and reducing alpha-frequency activity during NREM (Lo et al. 2002, Epilepsia (Lo, Yang, and Lo 2002)). Gabapentin (100–600 mg at bedtime) is first-line when pain and sleep disturbance co-occur; pregabalin (25–150 mg) has similar efficacy with more predictable pharmacokinetics. Both improve delta power on polysomnography, but neither fully eliminates alpha intrusion — they shift the ratio rather than resolve the underlying oscillatory dysfunction.
Low-dose tricyclics (amitriptyline 5–25 mg, doxepin 3–6 mg) improve subjective sleep quality through histamine H1 blockade and serotonin effects, but their impact on alpha-delta architecture specifically is modest. Trazodone (25–100 mg) is widely used for sleep initiation but does not consistently alter the alpha/delta power ratio.
The dual orexin receptor antagonists suvorexant and lemborexant target the orexin-driven hyperarousal that may sustain alpha intrusion (López-Amador et al. 2025, Exploration (López-Amador 2025)), and are mechanistically interesting given the documented orexin system dysfunction in ME/CFS — though clinical experience in this population remains limited. Z-drugs (zolpidem, eszopiclone) are effective for sleep initiation but do not address the unrefreshing quality of ME/CFS sleep; zolpidem’s suppression of glymphatic-driving norepinephrine oscillations may actively worsen waste accumulation (Hauglund et al. 2025 (Hauglund et al. 2025)).
The pattern across all of these is the same: partial improvement in sleep quality, no specific correction of the thalamic oscillatory dysfunction that produces alpha-delta intrusion. We do not yet have a pharmacological tool that targets the mechanism directly.
6 A more specific target — and a trial that could test it
Ethosuximide is a selective T-type calcium channel blocker used in absence epilepsy (Coulter et al. 1989, Neuroscience Letters (Coulter, Huguenard, and Prince 1989)), and it directly targets the CaV3.1 channels that generate thalamic oscillations. A single-dose crossover study — ethosuximide 250 mg versus placebo, overnight polysomnography measuring delta power and the alpha-delta index — would directly answer whether T-type channel blockade can shift ME/CFS sleep from alpha-contaminated to pure delta.
The direction is genuinely uncertain. If alpha-delta intrusion results from insufficient T-type current for delta generation (the channel already hypoactive due to energy failure), then blocking it further would worsen sleep. If it results from residual T-type current firing at the wrong frequency (the channel fires, but at alpha rather than delta frequency due to altered competing currents), then selective blockade might suppress the aberrant alpha component specifically. Those two possibilities predict opposite outcomes from the same drug. One night, one drug, one PSG — and the thalamic calcium hypothesis for ME/CFS sleep dysfunction is either supported or refuted, and we learn whether pharmacologically fixing alpha intrusion is even possible with currently available tools.
7 Why this matters beyond sleep
The thought experiment’s real value is not in predicting how much a single intervention would help. It is in making visible a structural feature of the disease that often goes unnamed: ME/CFS is multi-mechanism, and those mechanisms are partially independent of each other.
Each article in this sleep series has described a different piece — alpha-wave intrusion, glymphatic failure, adenosine dysregulation, autonomic handoff failure, thermoregulatory instability, circadian desynchronisation, orexin suppression. All of them independently impair sleep quality, and all have their own partially distinct upstream causes. Addressing one genuinely helps, and that matters to the people living with this. But it is not enough.
Some patients already understand this intuitively. They track their sleep with wearables, notice that some nights are clearly “alpha nights” while others are “autonomic nights” or “temperature nights,” and are effectively doing empirical phenotyping of their own disease. The research community should be matching that granularity. Polysomnography with spectral analysis, autonomic testing, and circadian phase measurement should be the standard of care assessment, not the exception.
The field is not waiting for a single breakthrough drug. It is waiting for the recognition that ME/CFS requires the same multi-target, mechanism-matched thinking that became standard in oncology and cardiology decades ago — and has still not been systematically applied here.
Previous article in this series: Tired But Wired — the autonomic betrayal at bedtime.