John Murphy, CEO, The COVID-19 Long haul Foundation
A mechanistic synthesis of post-acute neurological sequelae of SARS-CoV-2 infection
Abstract
Neurological manifestations constitute one of the most consequential and biologically heterogeneous components of post-acute sequelae of SARS-CoV-2 infection (PASC), commonly termed Long COVID. Cognitive dysfunction, fatigue, headache, sleep disturbance, dysautonomia, anosmia, neuropathic pain, sensory abnormalities, movement disorders, vestibular symptoms and neuropsychiatric disturbances may persist for months or years after the acute infection. The central biological question is no longer whether SARS-CoV-2 can produce persistent neurological disease, but how a transient systemic viral infection can generate prolonged dysfunction in neural, vascular, immune, autonomic and peripheral nervous systems.
Evidence accumulated through 2024–26 indicates that neurological Long COVID is unlikely to represent a single neuropathological entity. Instead, it comprises overlapping biological phenotypes involving persistent systemic inflammation, blood–brain-barrier (BBB) dysfunction, endothelial injury, altered cerebral perfusion, neuroimmune activation, autonomic dysregulation, small-fiber neuropathy, metabolic disturbance and, in a subset of patients, persistent viral or antigenic stimulation. Dynamic-contrast MRI studies have demonstrated persistent BBB disruption in patients with cognitive Long COVID, accompanied by abnormalities of coagulation and adaptive immunity.¹ Subsequent MRI work has reported subtle BBB permeability abnormalities more than two years after infection.² Multimodal imaging has additionally identified alterations in white-matter microstructure, cerebral blood flow, brain metabolites and regional brain volume.³–⁶
The neurological phenotype is also biologically heterogeneous. Some patients exhibit predominantly cognitive dysfunction; others have autonomic or small-fiber neuropathic disease, while still others develop headache, anosmia, vestibular symptoms, sleep disturbance or multisystem neurological illness. Recent histological studies provide evidence of autonomic small-fiber injury even in patients whose conventional sensory nerve-fiber density remains preserved.⁷˒⁸
Genomic evidence indicates that susceptibility to Long COVID is partly heritable. A 2025 genome-wide association study involving up to 6,450 Long-COVID cases and more than 1 million controls identified a significant association near FOXP4, although this locus appears to have a stronger relationship to pulmonary and systemic Long-COVID susceptibility than to a specifically neurological phenotype.⁹ The neurological phenotype is therefore more appropriately understood as the result of interactions among host genetic susceptibility, viral characteristics, immune response, vascular biology, autonomic regulation and tissue-specific vulnerability.
The emerging model is consequently one of postinfectious neurovascular–immune dysregulation rather than persistent primary viral encephalitis. Direct CNS infection may occur in selected circumstances, but available evidence increasingly supports indirect mechanisms—particularly BBB dysfunction, endothelial activation, persistent immune signaling, altered neurovascular coupling and peripheral neuroimmune injury—as major contributors to chronic neurological disease.
Introduction
The neurological consequences of SARS-CoV-2 infection were recognized during the earliest months of the pandemic.
Initially, attention focused on acute neurological complications:
- encephalitis;
- stroke;
- seizures;
- anosmia;
- encephalopathy;
- Guillain–Barré syndrome;
- cerebrovascular thrombosis.
The subsequent recognition of persistent neurological symptoms fundamentally changed the clinical landscape.
Patients who had survived the acute infection began reporting:
- inability to concentrate;
- impaired short-term memory;
- slowed thinking;
- word-finding difficulty;
- headache;
- dizziness;
- tinnitus;
- altered smell and taste;
- neuropathic pain;
- numbness;
- weakness;
- sleep disruption;
- orthostatic intolerance;
- tachycardia;
- exercise intolerance;
- post-exertional deterioration.
The term brain fog became the common clinical description.
Yet brain fog is not a diagnosis.
It is a phenotype.
This distinction is essential because the neurological manifestations of Long COVID are increasingly understood to represent several interacting biological processes rather than a single pathological lesion.
The contemporary neurological model can therefore be expressed as:
SARS-CoV-2 infection
↓
systemic and tissue-specific immune response
↓
vascular, endothelial and autonomic disturbance
↓
BBB dysfunction ± peripheral nerve injury ± neuroimmune activation
↓
altered cerebral perfusion and neural network function
↓
cognitive, sensory, autonomic and neurological symptoms.
1. The Neurological Phenotype
Neurological Long COVID encompasses at least seven partially overlapping domains.
Cognitive
- impaired attention;
- executive dysfunction;
- working-memory impairment;
- slowed processing speed;
- word-retrieval difficulty;
- impaired learning.
Autonomic
- postural orthostatic tachycardia syndrome;
- orthostatic intolerance;
- inappropriate sinus tachycardia;
- abnormal sweating;
- gastrointestinal dysmotility;
- thermoregulatory abnormalities.
Peripheral neuropathic
- burning pain;
- paresthesias;
- numbness;
- altered temperature sensation;
- dysesthesia;
- small-fiber neuropathy.
Cerebrovascular/neurovascular
- headache;
- dizziness;
- cerebral hypoperfusion;
- impaired neurovascular coupling.
Sensory
- anosmia;
- dysgeusia;
- tinnitus;
- visual disturbances;
- vestibular symptoms.
Sleep and neuropsychiatric
- insomnia;
- hypersomnia;
- fragmented sleep;
- anxiety;
- depression;
- altered arousal.
Motor
- weakness;
- impaired coordination;
- tremor;
- exercise intolerance.
The phenotypes frequently coexist.
A patient with small-fiber neuropathy may simultaneously have POTS, fatigue and cognitive dysfunction. Thus, neurological Long COVID is better represented as a network of interacting phenotypes than as discrete diseases.
2. Etiology: From Viral Infection to Chronic Neurological Dysfunction
The initiating event is SARS-CoV-2 infection.
The unresolved question is why neurological symptoms persist after the acute infection has apparently resolved.
Several non-mutually exclusive mechanisms have emerged.
2.1 Persistent viral antigen
SARS-CoV-2 RNA or protein has been detected in tissues after acute infection in some studies.
The critical distinction is between:
persistent replication
and
persistent antigenic material.
The latter could continue stimulating immune pathways without requiring high-level productive viral replication.
This mechanism could produce a chronic state of:
antigen → innate immune activation → endothelial dysfunction → neuroimmune signaling.
Evidence for persistent viral material does not establish that it is responsible for every neurological phenotype.
2.2 Persistent systemic inflammation
Patients with neurological Long COVID may exhibit persistent abnormalities in immune signaling.
A large-scale mechanistic phenotyping study found persistent myeloid inflammation and complement activation across several Long-COVID clinical phenotypes, including cognitive and fatigue groups.¹⁰
This is significant because complement and myeloid pathways can affect:
- endothelial permeability;
- platelet activation;
- leukocyte trafficking;
- synaptic homeostasis;
- microglial activation.
The brain therefore need not be directly infected for neurological dysfunction to occur.
3. The Blood–Brain Barrier as a Central Pathological Interface
The BBB is one of the most compelling biological bridges between systemic COVID-19 and chronic neurological disease.
The barrier consists principally of:
- endothelial cells;
- tight junctions;
- pericytes;
- basement membrane;
- astrocytic endfeet.
Its physiological function is to regulate movement of molecules and immune cells between blood and neural tissue.
SARS-CoV-2-associated systemic inflammation can disrupt this equilibrium.
The resulting sequence may be:
endothelial activation
↓
tight-junction alteration
↓
increased permeability
↓
entry of inflammatory mediators
↓
microglial/astrocytic activation
↓
altered synaptic and neuronal function.
4. Direct Evidence for BBB Dysfunction
Greene and colleagues provided some of the strongest evidence for BBB involvement in cognitive Long COVID.
Using dynamic contrast-enhanced MRI, they identified increased BBB permeability in patients with persistent brain fog. Peripheral blood mononuclear cells exhibited increased adhesion to human brain endothelial cells, while serum from affected patients induced inflammatory responses in brain endothelial cells in vitro.¹
The importance of this study is conceptual.
It connects:
systemic immune abnormalities
with
brain vascular dysfunction
and
cognitive symptoms.
A 2025 MRI study involving 97 individuals with Long COVID and neuropsychiatric symptoms likewise reported increased BBB permeability compared with recovered controls, although cerebral blood flow and global brain volume did not differ significantly.²
Thus, BBB abnormalities may persist even when conventional structural MRI appears relatively normal.
5. White-Matter Pathology
The white matter is particularly vulnerable to vascular and inflammatory perturbations.
Diffusion MRI provides indirect measures of white-matter microstructure.
A multimodal longitudinal study of patients with post-COVID cognitive impairment found abnormalities involving:
- BBB permeability;
- white-matter integrity;
- brain metabolites.
Some white-matter abnormalities improved during follow-up, whereas others persisted.³
More recent diffusion MRI work in individuals with neurological PASC persisting for approximately 2.7 years demonstrated global and tract-specific white-matter abnormalities, including prominent changes involving the fornix and minor forceps.⁴
These findings are intriguing because the affected pathways participate in:
- memory;
- executive processing;
- limbic integration.
However, diffusion abnormalities cannot by themselves distinguish:
- inflammation;
- demyelination;
- axonal injury;
- altered extracellular water.
Consequently, they should be interpreted as evidence of altered tissue microstructure rather than proof of a particular histopathological lesion.
6. Neuroinflammation
Neuroinflammation is one of the dominant mechanistic hypotheses.
Potential cellular participants include:
- microglia;
- astrocytes;
- endothelial cells;
- perivascular macrophages;
- infiltrating monocytes.
Microglia are the resident immune cells of the CNS.
Persistent activation can alter:
- synaptic pruning;
- neurotransmitter signaling;
- neuronal metabolism;
- network connectivity.
Astrocytes regulate:
- extracellular potassium;
- glutamate;
- metabolic support;
- BBB integrity;
- synaptic function.
Abnormal astrocytic signaling could therefore produce substantial neurological dysfunction without causing widespread neuronal death.
This distinction may explain why patients can have severe cognitive symptoms despite relatively modest conventional MRI abnormalities.
7. A Critical Distinction: Dysfunction versus Neuronal Destruction
One of the most important observations in neurological Long COVID is that severe symptoms do not necessarily correspond to widespread neuronal destruction.
Neurofilament light chain (NfL), a biomarker of axonal injury, has not consistently been elevated in Long-COVID cohorts.
A 2026 controlled study reported no significant increase in NfL or GFAP among its Long-COVID participants, arguing against ongoing widespread neuronal or astroglial destruction in that cohort.¹¹
This does not refute neuroinflammation.
Instead, it supports a more subtle model:
Neural dysfunction may be generated by altered vascular, immune, metabolic and synaptic physiology without extensive neuronal death.
That distinction has profound implications for reversibility.
8. Cerebral Hypoperfusion and Neurovascular Dysfunction
The brain consumes approximately one-fifth of resting cardiac output despite representing only a small fraction of total body mass.
Cognition depends upon tightly regulated cerebral blood flow.
Consequently:
small changes in perfusion
can produce
large changes in cognitive performance.
Neurovascular dysfunction may arise from:
- endothelial injury;
- impaired nitric-oxide signaling;
- autonomic dysfunction;
- abnormal vascular tone;
- altered carbon-dioxide responsiveness.
Recent imaging work in patients with neuropsychological Long COVID has identified associations between cognitive and psychiatric measures and regional cerebral blood-flow abnormalities.¹²
These findings provide a physiological pathway linking:
vascular dysfunction → cerebral hypoperfusion → impaired cognition.
9. Autonomic Nervous-System Dysfunction
The autonomic phenotype is one of the strongest bridges between peripheral and central neurological Long COVID.
Patients may exhibit:
- excessive heart-rate elevation on standing;
- reduced blood-pressure stability;
- impaired vasoconstriction;
- venous pooling;
- abnormal sweating;
- gastrointestinal dysmotility.
POTS is particularly important.
On standing:
gravity → venous pooling → reduced venous return → reduced stroke volume.
The normal response is:
sympathetic activation → vasoconstriction + increased heart rate.
If vasoconstriction is inadequate, the heart compensates excessively.
The result is:
tachycardia + cerebral hypoperfusion + dizziness + fatigue + cognitive dysfunction.
This mechanism can therefore generate symptoms commonly attributed exclusively to the brain.
10. Small-Fiber Neuropathy
Small-fiber neuropathy provides another important mechanistic pathway.
Small sensory and autonomic fibres include:
- C fibres;
- A-delta fibres.
They regulate:
- pain;
- temperature;
- sweating;
- vascular tone;
- visceral function.
Conventional nerve-conduction studies primarily evaluate large myelinated fibres.
Consequently, a patient can have substantial small-fiber pathology despite normal routine electrophysiology.
A 2025 cohort study of patients with neuropathic pain after COVID-19 found abnormal skin-biopsy results in 14 of 17 evaluable participants (82%); autonomic symptoms occurred in 70% and were associated with abnormal biopsy findings. At six months, 10 of 17 participants reported improvement in pain and/or dysautonomia.⁷
Even more recent histological work has demonstrated reductions in autonomic nerve-fiber densities in painful Long COVID, including patients with preserved intraepidermal nerve-fiber density.⁸
This is an important observation.
It suggests:
autonomic small-fiber injury may occur independently of conventional somatic small-fiber loss.
11. Peripheral Nerve Injury: Potential Mechanisms
Potential mechanisms include:
Immune-mediated injury
Autoantibodies or inflammatory cells could damage peripheral nerves.
Microvascular ischemia
Small nerves depend upon an extensive microvascular network.
Endothelial dysfunction could therefore impair nerve perfusion.
Metabolic injury
Persistent metabolic abnormalities could impair axonal maintenance.
Neuroimmune interaction
Peripheral nerve terminals communicate extensively with immune cells.
The result may be a self-reinforcing loop:
immune activation → nerve dysfunction → autonomic dysregulation → vascular dysfunction → further immune activation.
12. Neurocognitive Long COVID
Brain fog is clinically heterogeneous.
Objective deficits have been reported in:
- attention;
- executive function;
- processing speed;
- working memory;
- verbal fluency.
However, not every patient with subjective brain fog demonstrates abnormalities on conventional neuropsychological testing.
This discrepancy is not necessarily evidence against biological disease.
Several mechanisms may produce fluctuating cognitive dysfunction:
- cerebral hypoperfusion;
- autonomic instability;
- sleep disruption;
- neuroinflammation;
- BBB dysfunction;
- metabolic dysfunction;
- sensory overload.
Cognition is therefore a systems-level output.
13. Glymphatic Dysfunction
The glymphatic system is involved in clearance of metabolites and interstitial solutes from the brain.
It is strongly influenced by:
- sleep;
- vascular pulsatility;
- astrocytic function;
- perivascular fluid movement.
A proof-of-concept MRI study reported asymmetrical glymphatic abnormalities in patients with neurocognitive PASC and examined their relationship to BBB permeability.⁵
If confirmed, this could provide an additional link between:
sleep disruption + vascular dysfunction + neuroinflammation + cognitive symptoms.
The hypothesis remains preliminary.
14. Glutamatergic Dysfunction and Excitotoxicity
Glutamate is the principal excitatory neurotransmitter of the CNS.
Excessive extracellular glutamate can produce excitotoxic injury.
A multimodal MRI/MRS study of patients with post-COVID cognitive impairment reported abnormalities involving:
- BBB permeability;
- myoinositol;
- glutamate/glutamine;
- white-matter microstructure.
Some abnormalities changed during 12-month follow-up.³
This raises the possibility that BBB dysfunction and altered glutamatergic metabolism contribute to cognitive impairment.
Again, however, these findings require replication in larger cohorts.
15. Genomics
The genetic architecture of neurological Long COVID remains incompletely characterized.
The largest GWAS to date identified a significant association near FOXP4 among Long-COVID cases.⁹
The study included up to:
6,450 Long-COVID cases
and
1,093,995 controls
across 24 studies and 16 countries.
The lead variant, rs9367106, was associated with increased Long-COVID risk and replicated in independent cohorts.⁹
The association was particularly interesting because FOXP4 expression is relevant to:
- lung biology;
- immune cells;
- epithelial regeneration;
- hypothalamic tissue.
However, it would be scientifically incorrect to describe FOXP4 as a neurological Long-COVID gene.
The current evidence suggests that it is a general Long-COVID susceptibility locus, potentially acting through pulmonary, immune and tissue-repair mechanisms.
16. Genetic Susceptibility Is Not Genetic Determinism
The neurological phenotype probably reflects an interaction:
host genotype
×
viral genotype
×
acute disease severity
×
immune response
×
vascular susceptibility
×
environmental factors.
Thus, two genetically similar individuals can have different clinical trajectories.
Conversely, individuals with very different genetic backgrounds may develop similar neurological phenotypes through convergent biological mechanisms.
17. Epigenomics and Immune Memory
An important unresolved question is whether SARS-CoV-2 leaves a durable molecular imprint on immune cells.
Potential mechanisms include alterations in:
- DNA methylation;
- chromatin accessibility;
- transcription-factor activity;
- mitochondrial signaling;
- immune-cell differentiation.
Such changes could explain how a transient infection produces persistent immune dysregulation.
The hypothesis is compelling but currently less established than BBB dysfunction, autonomic abnormalities and small-fiber pathology.
18. Viral Reactivation
Latent-virus reactivation is another possible component.
Epstein–Barr virus and other herpesviruses can reactivate during immune perturbation.
A possible sequence is:
SARS-CoV-2 infection
↓
immune dysregulation
↓
loss of latent-virus immune control
↓
viral reactivation
↓
secondary inflammatory signaling
↓
neurological symptoms.
This mechanism could be especially relevant to patients whose neurological symptoms emerge after an apparently complete recovery from the acute infection.
It should not, however, be interpreted as proof that EBV reactivation causes neurological Long COVID in all patients.
19. The Gut–Brain Axis
The intestine contains:
- extensive immune tissue;
- enteric neurons;
- endocrine cells;
- microbial ecosystems.
SARS-CoV-2 infection can alter the intestinal environment.
Potential consequences include:
- dysbiosis;
- altered microbial metabolites;
- epithelial dysfunction;
- intestinal immune activation.
The vagus nerve provides a major communication pathway between the gut and brain.
Consequently:
intestinal inflammation → vagal signaling → brainstem/autonomic pathways → CNS effects
represents another plausible route to neurological symptoms.
20. Mitochondrial and Metabolic Contributions
The nervous system has extraordinarily high metabolic requirements.
Neurons depend heavily upon oxidative phosphorylation.
If systemic illness produces:
- mitochondrial dysfunction;
- impaired fatty-acid utilization;
- altered glucose metabolism;
- abnormal redox balance;
the brain may become functionally vulnerable.
This could contribute to:
- cognitive fatigue;
- impaired attention;
- exercise intolerance;
- post-exertional worsening.
Metabolic dysfunction may therefore be an amplifier rather than a primary cause.
21. Why Brain Fog Fluctuates
One of the characteristic features of neurological Long COVID is fluctuation.
Symptoms may worsen with:
- physical exertion;
- prolonged cognitive activity;
- sleep deprivation;
- orthostasis;
- infection;
- stress;
- temperature changes.
This fluctuation is difficult to reconcile with a fixed structural brain lesion.
It is more compatible with a dynamic physiological disorder.
The likely model involves interaction among:
autonomic state
cerebral perfusion
immune activation
metabolic reserve
sleep
neural network stability.
22. Clinical Course
Neurological Long COVID has no single trajectory.
Several patterns are apparent.
Persistent from acute illness
Symptoms never completely resolve.
Delayed neurological onset
The patient initially recovers and develops cognitive or autonomic symptoms later.
Relapsing–remitting
Symptoms fluctuate over weeks or months.
Progressive multisystem phenotype
Neurological symptoms become accompanied by autonomic, gastrointestinal or vascular disease.
Partial recovery
One domain improves while another persists.
Chronic persistence
Symptoms remain for years.
A 2026 longitudinal cohort reported domain-specific trajectories: cognitive impairment decreased substantially over 24 months, autonomic dysfunction showed a biphasic course, whereas peripheral neuropathy remained relatively persistent.¹³
Although that study requires confirmation in larger prospective cohorts, the finding is conceptually important:
neurological Long COVID does not behave as one disease with one recovery curve.
23. Reversibility
The possibility of recovery is biologically important.
If symptoms were caused primarily by irreversible neuronal destruction, recovery would be difficult to explain.
Instead, some studies show improvement in:
- cognitive performance;
- white-matter abnormalities;
- autonomic symptoms;
- neuropathic symptoms.
This is compatible with a disease characterized partly by functional neural dysregulation and potentially reversible neurovascular injury.
It does not exclude permanent injury in a subset of patients.
24. A Unified Pathophysiological Model
The emerging model can be expressed as follows:
Stage 1 — Viral infection
SARS-CoV-2 infects respiratory and extrapulmonary tissues.
↓
Stage 2 — Systemic immune activation
Interferon, cytokine, complement and myeloid pathways become activated.
↓
Stage 3 — Vascular injury
Endothelial dysfunction and altered coagulation develop.
↓
Stage 4 — BBB dysfunction
Neurovascular-unit integrity becomes impaired.
↓
Stage 5 — Neuroimmune activation
Microglia, astrocytes and perivascular immune cells respond.
↓
Stage 6 — Autonomic and peripheral nerve injury
Small autonomic fibres and regulatory pathways become dysfunctional.
↓
Stage 7 — Physiological network dysfunction
Cerebral perfusion, metabolic efficiency and neural signaling become abnormal.
↓
Stage 8 — Clinical phenotype
brain fog + fatigue + dysautonomia + neuropathy + headache + sensory disturbance + sleep dysfunction.
This model accommodates both central and peripheral neurological disease.
25. Why the Brain May Be a “Downstream Organ”
An important conceptual shift is required.
The neurological manifestations of Long COVID may not originate entirely within the CNS.
The brain may be a downstream target of systemic disease.
For example:
systemic endothelial dysfunction
→ cerebral microvascular dysfunction
→ impaired cognition.
Similarly:
peripheral autonomic dysfunction
→ abnormal blood-pressure regulation
→ cerebral hypoperfusion
→ brain fog.
And:
systemic immune activation
→ BBB dysfunction
→ neuroimmune activation
→ cognitive symptoms.
Thus, neurological Long COVID is partly a disease of the neurovascular unit, not merely of neurons.
26. Clinical Implications
The neurological evaluation of a patient with Long COVID should therefore be phenotype-driven.
Depending upon symptoms, appropriate evaluation may include:
- detailed neurological examination;
- formal neuropsychological testing;
- autonomic testing;
- orthostatic vital signs;
- tilt-table testing;
- nerve-conduction studies;
- skin biopsy for suspected small-fiber neuropathy;
- quantitative sensory testing;
- sleep evaluation;
- MRI where clinically indicated;
- laboratory assessment for alternative neurological causes.
Routine brain MRI may be normal.
That does not exclude neurological Long COVID.
27. Therapeutic Implications
The mechanistic heterogeneity argues strongly against a universal treatment.
Potential treatment targets include:
Persistent viral antigen
Antiviral therapy.
Autoimmune mechanisms
Immunomodulation.
Autonomic dysfunction
Volume expansion, compression, rehabilitation and phenotype-specific autonomic therapy.
Small-fiber neuropathy
Neuropathic pain treatment and, in selected cases, investigation of immune-mediated mechanisms.
Neuroinflammation
Targeted anti-inflammatory or immunomodulatory strategies.
Sleep dysfunction
Treatment of sleep disorders.
Metabolic dysfunction
Carefully individualized rehabilitation and metabolic interventions.
However, most mechanistic therapies remain investigational.
The evidence should not be extrapolated from biological plausibility to clinical efficacy.
28. The Problem of Post-Exertional Malaise
One of the most important clinical constraints is PEM.
Conventional rehabilitation assumes:
exercise → conditioning → increased capacity.
In patients with PEM, excessive exertion may instead produce:
activity → delayed physiological deterioration → prolonged symptom exacerbation.
This distinction is particularly important for neurological Long COVID because autonomic dysfunction, metabolic abnormalities and neuroimmune signaling may all contribute.
Clinical rehabilitation therefore requires individualized pacing rather than an assumption that progressive exercise is universally beneficial.
29. What Remains Uncertain
Several major questions remain unanswered.
Is persistent virus necessary?
Probably not in every patient.
Is autoimmunity universal?
Almost certainly not.
Is neuroinflammation universal?
Evidence suggests substantial heterogeneity.
Is BBB dysfunction universal?
No.
Is small-fiber neuropathy universal?
No.
Is cerebral hypoperfusion universal?
No.
Does FOXP4 specifically cause neurological Long COVID?
There is currently insufficient evidence for that conclusion.
Can neurological Long COVID produce permanent neuronal injury?
Probably in some patients, but the magnitude and frequency remain uncertain.
30. The Most Important Research Priority
Future studies should abandon the assumption that all neurological Long-COVID patients have the same disease.
The optimal study would simultaneously measure:
genome
transcriptome
proteome
metabolome
immune repertoire
viral persistence
BBB permeability
cerebral perfusion
autonomic physiology
peripheral nerve structure
cognitive function
longitudinal clinical trajectory.
Only such multimodal studies can establish whether clinically similar patients actually have different biological diseases.
31. Conclusions
Neurological Long COVID represents one of the most complex consequences of SARS-CoV-2 infection.
The available evidence increasingly argues against a simple model of persistent viral encephalitis.
Instead, neurological disease appears to emerge from interactions among:
immune dysregulation
endothelial dysfunction
BBB disruption
neuroinflammation
autonomic dysfunction
small-fiber injury
altered cerebral perfusion
metabolic dysfunction
and, potentially, persistent viral antigen or secondary viral reactivation.
The pathology is consequently neither exclusively central nor exclusively peripheral.
It is neurovascular, neuroimmune and systemic.
The BBB appears to represent a particularly important interface. Evidence from dynamic MRI, molecular profiling and experimental endothelial models demonstrates that patients with cognitive Long COVID can exhibit persistent vascular-barrier abnormalities accompanied by systemic immune and coagulation disturbances.¹˒²
Peripheral nervous-system pathology provides an additional biological substrate. Skin-biopsy studies increasingly demonstrate small-fiber abnormalities, particularly involving autonomic fibres, even when conventional neurological testing remains normal.⁷˒⁸
Neuroimaging further suggests that the disorder can leave measurable physiological and structural signatures, including altered white-matter microstructure, cerebral perfusion, metabolites and regional brain morphology.³–⁶
Yet the clinical course is heterogeneous. Cognitive dysfunction may gradually improve, autonomic disease may fluctuate or follow a biphasic trajectory, while peripheral neuropathy can remain comparatively persistent.¹³
The most defensible contemporary formulation is therefore:
Neurological Long COVID is a heterogeneous postinfectious neurovascular–immune disorder in which systemic immune and vascular abnormalities interact with central and peripheral nervous-system vulnerabilities to produce persistent disturbances of cognition, autonomic regulation, sensory processing and neural function.
This formulation has an important consequence for clinical science.
The question is no longer simply whether SARS-CoV-2 damages the nervous system.
It is:
Which biological pathway is responsible for neurological disease in a particular patient, at what stage of the illness, and which component remains reversible?
Answering that question will require a transition from symptom-based Long-COVID research to mechanism-defined neurological endotypes.
That transition is likely to determine whether the next generation of therapies produces modest symptomatic benefit—or genuine disease modification.
Numbered References / Footnotes
1. Greene C, Connolly R, Brennan D, et al. Blood-brain barrier disruption and sustained systemic inflammation in individuals with long COVID-associated cognitive impairment. Nature Neuroscience. 2024;27:421–432. Dynamic-contrast MRI demonstrated BBB disruption in patients with Long-COVID brain fog, accompanied by coagulation and immune abnormalities.
2. Rubin LH, Shi W, Azola A, et al. Blood-brain barrier disruption in long COVID and cognitive correlates: a cross-sectional MRI study. Brain, Behavior, and Immunity. 2025;129:989–999. The study found increased BBB permeability in 97 participants with Long COVID compared with 31 recovered controls, with an association between permeability and motor performance.
3. Chaganti S, Poudel P, Cysique LA, et al. Blood brain barrier disruption and glutamatergic excitotoxicity in post-acute sequelae of SARS-CoV-2 infection cognitive impairment. Multimodal longitudinal MRI/MRS study demonstrating BBB, white-matter and metabolic abnormalities.
4. White matter microstructural abnormalities in neurological post-acute sequelae of coronavirus disease: imaging signatures consistent with persistent neuroinflammation. 2026. Diffusion MRI demonstrated global and tract-specific white-matter abnormalities in individuals with neurological PASC persisting for approximately 2.7 years.
5. Asymmetrical glymphatic dysfunction in patients with long COVID-associated neurocognitive impairment—correlation with BBB disruption. 2025. Proof-of-concept MRI study examining glymphatic function and BBB permeability in neurocognitive PASC.
6. Nakase T, Takano Y, Nomura S, et al. Association of symptoms of neuropsychological long COVID with imaging and plasma biomarkers. Journal of the Neurological Sciences. 2026;483:125808. Regional cerebral-blood-flow abnormalities correlated with cognitive and psychiatric measures.
7. Drobinska N, Nehme M, Assal F, et al. Small Fiber Neuropathy in Long COVID: A Cohort Study with Multimodal Assessment and Follow-Up. 2025. In a cohort of patients with post-COVID neuropathic pain and normal large-fiber nerve conduction, 82% had abnormal skin biopsies and 70% reported autonomic symptoms.
8. Falco G, Galosi A, Litewczuk A, et al. Autonomic small fiber involvement in painful long COVID: a histological and clinical study. 2026. Histological evidence demonstrated reduced autonomic small-fiber densities, including in patients with preserved intraepidermal nerve-fiber density.
9. Lammi V, Nakanishi T, Jones SE, et al. Genome-wide association study of long COVID. Nature Genetics. 2025. Meta-analysis involving up to 6,450 Long-COVID cases and 1,093,995 controls identified and replicated an association at the FOXP4 locus.
10. Liew F, Efstathiou C, Fontanella S, et al. Large-scale phenotyping of patients with long COVID post-hospitalization reveals mechanistic subtypes of disease. Nature Immunology. 2024. The investigators identified persistent myeloid inflammation and complement activation across multiple clinical Long-COVID phenotypes, including cognitive and fatigue groups.
11. Long-COVID: assessment of circulating markers suggests no cerebral neuronal damage, neuroinflammation or systemic inflammation—a controlled study. Scientific Reports. 2026. The study found no significant differences in circulating NfL or GFAP between Long-COVID participants and controls, illustrating the heterogeneity of proposed neuroinflammatory biomarkers.
12. Nakase T, Takano Y, Nomura S, et al. Journal of the Neurological Sciences. 2026;483:125808. The investigators found associations between regional cerebral-blood-flow abnormalities and neuropsychological measures in patients with neuropsychological Long COVID.
13. Temporal patterns of neurological deterioration in COVID-19 survivors: a longitudinal cohort analysis of post-acute neurological sequelae. 2026. In a 386-person longitudinal cohort, cognitive impairment decreased over 24 months, autonomic dysfunction demonstrated a biphasic pattern, and peripheral neuropathy remained comparatively persistent.
14. Chakraborty C, Bhattacharya M, Chatterjee S, Lee S-S. Long COVID-associated neurological symptoms and brain fog: understanding the mechanism of neuroinflammation, BBB disruption, diagnostics, and therapeutics. Molecular Biology Reports. 2026;53:401. Contemporary review synthesizing evidence for neuroinflammation, BBB disruption and neurological mechanisms in Long COVID.
15. Talkington M, Kolluru A, Gressett C, et al. Neurological sequelae of long COVID: a comprehensive review of diagnostic imaging, underlying mechanisms, and potential therapeutics. 2025. Review of MRI, PET, EEG and other neurological investigations and proposed mechanisms.