Waking Local Sleep: A Convergent Electrophysiological Mechanism for Brain Fog
The mechanisms described in this chapter — adenosine accumulation (Section Adenosine Accumulation and Pathological Sleep Pressure), cytokine-induced somnolence (Section Inflammatory Cytokine-Induced Somnolence and Fatigue), microglial activation (Section Microglia Activation and Neuroinflammatory Fatigue), catecholamine deficiency (Section Why ME/CFS Patients Cannot Tolerate Stimulation: An Integrated Model), and failed inhibitory GABAergic brakes (this section, above) — all act upon a shared electrophysiological substrate: the ability of cortical circuits to maintain desynchronised, information-processing activity during wakefulness. When this substrate fails, individual cortical columns transition into sleep-like slow-wave oscillations (\(\leq\) 4 Hz delta) while the rest of the brain remains behaviourally awake — a phenomenon termed “local sleep” (Andrillon and Oudiette 2023). This section proposes that waking local sleep represents a convergent mechanism through which the diverse upstream pathologies of ME/CFS produce the subjective experience of cognitive dysfunction.
In neurotypical individuals, waking slow waves are a normal, use-dependent phenomenon: they increase with cognitive fatigue, predict attention lapses and mind-wandering, and are suppressed by noradrenergic/dopaminergic tone (Pinggal et al. 2022). Adults with ADHD show significantly elevated waking slow-wave density over parieto-temporal electrodes compared to controls, and mediation analysis confirms that slow-wave density explains a significant portion of ADHD-related attentional difficulties (Pinggal et al. 2026). Pediatric ADHD subtypes show dissociable waking slow-wave topography — prefrontal in combined type, more posterior in inattentive type — suggesting that the specific cortical regions entering local sleep determine the cognitive domain affected (Gong et al. 2026). Post-COVID patients with persistent attention deficits exhibit global slowing and increased intra-individual RT variability — the behavioral signature predicted by elevated waking slow-wave activity (Ortelli et al. 2022).
The threshold for local sleep entry is governed by three interacting factors, all of which are disrupted in ME/CFS:
Catecholamine gatekeeping. Noradrenergic projections from the locus coeruleus maintain cortical desynchronisation. When noradrenergic tone is pharmacologically reduced, waking slow-wave density increases; when enhanced (methylphenidate), it decreases (Pinggal et al. 2022). ME/CFS involves documented central noradrenergic deficiency (Walitt et al. 2024).
Inflammatory priming. Systemic LPS administration produces region-specific waking slow-wave intrusions in animals without requiring the usual cognitive-load precondition (Leemburg et al. 2025). ME/CFS involves chronic systemic and neuroinflammatory signalling (Section Inflammatory Cytokine-Induced Somnolence and Fatigue, Section Microglia Activation and Neuroinflammatory Fatigue).
Energy substrate limitation. Maintaining waking ion gradients (Na+/K+-ATPase, KCC2 chloride extrusion) is metabolically expensive. When ATP availability drops — as in ME/CFS mitochondrial dysfunction (Chapter Energy Metabolism and Mitochondrial Function) — the probability that any given cortical column will exhaust its local energy budget and tip into slow-wave oscillation increases (Van Dongen 2025). Intracellular chloride accumulation during sustained waking activity, when not adequately cleared by energy-dependent KCC2 transporters, directly increases local sleep pressure (Alfonsa et al. 2023).
The convergence of these three disrupted gatekeeping factors — impaired noradrenergic suppression of slow waves, lowered slow-wave threshold from neuroinflammation, and depleted energy substrate for ion gradient maintenance — predicts that ME/CFS patients will experience elevated waking slow-wave density during cognitive tasks, proportional to cognitive-fatigue severity. The behavioral consequences of these electrophysiological intrusions — attention lapses, slowed processing, response variability, and the subjective experience of “brain fog” — represent a convergent endpoint of the diverse upstream pathologies characterised in this chapter.
Cross-disease context. The ADHD parallel is instructive: ADHD is a neurodevelopmental disorder, while ME/CFS is acquired — but both produce a final common pathway of elevated waking slow-wave activity and attentional dysfunction, suggesting the local-sleep mechanism is domain-general rather than diagnosis-specific. The use-dependent nature of local sleep (Type 2, cortically driven (Deboer 2026)) also provides a mechanistic foundation for post-exertional cognitive malaise: cognitive effort depletes local cortical energy stores, increasing slow-wave pressure that is inadequately suppressed by the deficient noradrenergic system, producing a delayed cognitive crash proportional to prior cognitive load.
Certainty: 0.35. The phenomenon of “local sleep” — sleep-like slow waves occurring in localized cortical patches during wakefulness — provides a mechanistic bridge between the neurochemical and energy-limitation mechanisms described in this chapter and the subjective experience of cognitive dysfunction (“brain fog”). The upstream drivers converge into a tractable electrophysiological endpoint: individual cortical columns entering slow-wave oscillations (1–4 Hz delta) while the rest of the brain maintains waking activity.
Mechanism. In neurotypical individuals, sleep-like slow waves appear during demanding or repetitive waking tasks as a use-dependent consequence of prior cortical activity (Andrillon and Oudiette 2023). These intrusions are gated by catecholamine tone: noradrenaline and dopamine suppress waking slow waves, while serotonin promotes them (Pinggal et al. 2022). Intracellular chloride accumulation during wake at synapses with low KCC2 expression increases the probability that local patches of cortex tip into slow-wave oscillations when metabolic or attentional demands exceed the available energy budget (Alfonsa et al. 2023). Systemic inflammation can trigger region-specific waking slow-wave intrusions even without the usual cognitive-load precondition (Leemburg et al. 2025).
In ME/CFS, all three gatekeeping factors are disrupted simultaneously:
Catecholamine deficiency: The NIH deep phenotyping study documented significant deficits in central noradrenergic signalling (reduced CSF DHPG, blunted LC response) (Walitt et al. 2024). Without adequate noradrenergic tone, the neural circuits that normally maintain waking cortical desynchronisation lose their primary arousal support — cortical columns that would normally be held in a depolarised, information-processing state drift toward the depolarised-downstate alternation characteristic of slow-wave sleep.
Neuroinflammation and altered chloride homeostasis: Chronic microglial activation (Section Microglia Activation and Neuroinflammatory Fatigue) elevates extracellular cytokines that shift KCC2/NKCC1 transporter expression ratios, increasing intracellular chloride in cortical neurons (Alfonsa et al. 2023). The resultant shift in GABA reversal potential means that what should be inhibitory GABAergic input instead produces paradoxical depolarisation or fails to prevent the transition to slow-wave oscillations. This is mechanistically convergent with the KCC2 downregulation documented in neuroinflammatory states Coull et al. (2003, Nature).
Energy limitation: The ATP synthesis deficit (Chapter Energy Metabolism and Mitochondrial Function) and reduced cerebral blood flow (Chapter Neurological and Neurocognitive Dysfunction) mean that the ion pumps maintaining waking membrane potentials (Na+/K+-ATPase, KCC2) operate on a depleted energy substrate. Cortical columns that in a healthy brain would sustain waking activity during a demanding task instead exhaust their local ATP budget more rapidly, triggering local slow-wave transitions at cognitive loads that would be trivially sustained by a non-ME/CFS brain (Van Dongen 2025).
Direct inflammation-to-local-sleep pathway: The LPS-challenge model demonstrates that peripheral inflammation can produce region-specific waking slow-wave intrusions without requiring cognitive load (Leemburg et al. 2025). In ME/CFS — where systemic inflammation markers are elevated at baseline and surge after exertion (Section Inflammatory Cytokine-Induced Somnolence and Fatigue) — this pathway may operate tonically and episodically, producing both background cognitive impairment and post-exertional cognitive exacerbation.
In adults with ADHD, waking slow-wave density is significantly elevated compared to controls and mediates attentional lapses, slowed reaction times, increased response variability, and subjective sleepiness (Pinggal et al. 2026). The behavioral signature — omission errors, intra-individual RT variability, task disengagement (“mind blanking”) — maps precisely onto the cognitive complaints reported by ME/CFS patients as “brain fog.” In post-COVID attention studies, global slowing and increased intra-individual variability — the behavioral correlate of elevated waking slow-wave activity — are characteristic findings (Ortelli et al. 2022). No study has yet applied the waking slow-wave EEG paradigm to ME/CFS patients.
The local-sleep model integrates several seemingly disconnected findings from the ME/CFS sleep literature: the ultra-slow power (0.3–0.79 Hz) deficit during slow-wave sleep in CFS (Neu et al. 2015) — the same frequency band in which waking slow waves occur; the reduced parasympathetic activity specifically during SWS, indicating that even when patients achieve deep sleep, it is not restorative (Fatt et al. 2020); and the paradoxical finding that ME/CFS patients can have normal or increased total SWS duration while still reporting profoundly unrefreshing sleep (Maksoud et al. 2021). If the same oscillatory defect — a failure to maintain desynchronised, information-processing cortical states — occurs during both wake (as local sleep intrusions causing brain fog) and sleep (as alpha intrusion and microarousals preventing restorative slow-wave function), the mechanism becomes parsimonious: a single sleep-wake boundary dysregulation phenotype produces daytime cognitive dysfunction and non-restorative sleep through the same underlying neurochemical and metabolic substrate.
Falsifiable prediction: In ME/CFS patients performing a sustained attention task under EEG, waking slow-wave density (SW/min over parieto-temporal electrodes) will be significantly higher than in matched controls, and will correlate with subjective brain fog severity (MFS or CFQ), omission error rate, and RT variability. Falsified if no group difference in waking SW density, or if SW density does not correlate with cognitive performance.
Consequence: This provides a single testable explanatory mechanism for cognitive dysfunction in ME/CFS that bridges neuroinflammation, catecholamine deficiency, and energy metabolism. If confirmed, waking slow-wave density would become an objective electrophysiological biomarker of brain fog severity, and treatment strategies could target catecholamine support, anti-neuroinflammatory interventions, and energy-substrate optimization with a quantifiable electrophysiological endpoint rather than subjective report alone. The immediate practical consequence is to identify a specific, tractable research protocol that could disambiguate this mechanism from competing cognitive-dysfunction models within 2–3 years.
Limitations. This hypothesis relies entirely on cross-disease extrapolation from ADHD, post-COVID, and animal models; zero ME/CFS waking slow-wave data exist. The behavioral signature overlap (RT slowing, variability, lapses) is circumstantial — these are generic cognitive dysfunction markers present across many neurological conditions, not specific to local sleep. Competing explanations for the same cognitive phenotype include: glymphatic failure with metabolite accumulation, reduced cerebral blood flow during upright posture (orthostatic cognitive hypoperfusion), direct cytokine effects on prefrontal function, kynurenine pathway neurotoxicity, and sleep fragmentation impairing overnight memory consolidation. These are not mutually exclusive — multiple mechanisms may converge on the same behavioral endpoint. Catecholamine deficiency in ME/CFS has been documented at the group level; intra-individual correlation between catecholamine levels and waking slow-wave density has not been measured. BDNF-TrkB signalling, which contributes to local sleep buildup in animal models (ElGrawani et al. 2024), shows mixed findings in ME/CFS (both elevated and reduced reported). The local-sleep framework provides a mechanistic convergence point for existing observations but does not yet explain why some ME/CFS patients have brain fog as their dominant symptom and others do not. The ADHD data were collected in medication-withdrawn patients; whether concurrent psychostimulant treatment modifies the relationship is unknown.
Evidence note. The core local-sleep phenomenon (slow waves during wake) is well-established in human EEG literature (Andrillon and Oudiette 2023). PDE: 0.82 (review), 0.80 (ADHD primary, n=63). The link to ME/CFS is mechanistic inference only. Severity applicability: unknown — no waking slow-wave data exist in any ME/CFS severity stratum.
Research protocol. A testable prediction emerges directly from the model described below.
Certainty: prediction only — test has not been performed.
If waking local sleep is a convergent mechanism for ME/CFS cognitive dysfunction, a sustained-attention EEG study in ME/CFS patients versus matched controls should yield the following:
Primary prediction: ME/CFS patients will show significantly higher waking slow-wave density (SW/min, delta 1–4 Hz) over parieto-temporal and prefrontal electrodes during a sustained attention task (e.g., gradual-onset continuous performance task or psychomotor vigilance task) compared to healthy controls and fibromyalgia controls.
Dose-response: Slow-wave density will correlate positively with: (a) subjective brain fog severity (Mental Fatigue Scale, Chalder Fatigue Scale cognitive subscale); (b) omission error rate; (c) intra-individual RT coefficient of variation; and negatively with (d) CSF DHPG (noradrenergic tone marker), (e) spectral power in the alpha band during task performance (indicating preserved waking cortical desynchronisation).
PEM effect: Slow-wave density will increase 24–48 hours after a standardised exertional challenge (CPET) above baseline, tracking the temporal course of post-exertional cognitive worsening, with peak density at 24h matching peak subjective brain fog.
Pharmacological modulation: A single dose of a selective noradrenaline reuptake inhibitor (e.g., atomoxetine 40 mg) or a low-dose DORA (e.g., daridorexant 25 mg), administered in a crossover design, will reduce waking slow-wave density and improve cognitive task performance compared to placebo. Methylphenidate (per [Pinggal et al. (2022) in healthy adults) should also reduce SW density; atomoxetine should test whether selective noradrenergic enhancement alone is sufficient.
Negative control: Waking slow-wave density should NOT correlate with sleep onset latency (preserved in ME/CFS) or with measures of motor slowing (finger tapping, simple RT) — confirming that the effect is cortical/cognitive, not global psychomotor.
Falsification: The hypothesis is falsified if ME/CFS patients do NOT show higher waking SW density than controls, OR if SW density does not correlate with cognitive performance, OR if pharmacological noradrenergic enhancement does not reduce SW density.
Consequence: A positive result would establish waking slow-wave density as the first direct electrophysiological biomarker of ME/CFS brain fog — objectively measurable, pharmacologically responsive, and anchored to a specific mechanistic framework rather than subjective report. A negative result would be equally informative: it would eliminate one candidate mechanism for the ME/CFS cognitive phenotype, directing research resources toward alternative explanations (glymphatic failure, orthostatic hypoperfusion, kynurenine neurotoxicity). In either outcome, the experiment has discriminating value. The protocol is technically achievable now: the EEG paradigm is established ((Pinggal et al. 2026)), CSF catecholamine measurement is available through existing lumbar puncture protocols, and CPET-integrated cognitive testing is routinely performed in ME/CFS research settings. Expected timeline: 2–3 years if funded.
Evidence note. The prediction is grounded in: (1) established human EEG paradigm for waking slow-wave detection (Pinggal 2026, n=63; Gong 2026, n=120); (2) pharmacological modulation of waking SW by catecholaminergic agents (Pinggal 2022, n=32 RCT crossover); (3) documented noradrenergic deficiency in ME/CFS (Walitt 2024, n=17 NIH deep phenotyping); (4) post-COVID attention studies showing the predicted behavioral signature (Ortelli 2022, n=117); (5) LPS-induced waking SW in animals (Leemburg 2025). No component of the prediction chain requires novel technology or unvalidated assays. :::