Neurological Biomarkers
1 Brain Imaging Markers
1.1 Structural MRI
Documented abnormalities:
- White matter hyperintensities (variable)
- Regional gray matter volume changes
- Brainstem abnormalities in some studies
Certainty: 0.40. A multimodal MRI study from the Griffith University lab reported altered T1w/T2w myelin-signal, diffusion-tensor microstructure, and posterior-cingulate neurochemicals that separated long COVID, COVID-recovered, and never-infected controls, with the myelin-signal correlating with physical and cognitive function (Thapaliya et al. 2025). Combined with independent-lab evidence that brain microstructural alteration is detectable in ME/CFS and post-COVID cohorts (a large-n cingulum DTI finding at 85% diagnostic accuracy (Wu et al. 2026); advanced diffusion-NII showing edema and cellular-infiltration signatures in ME/CFS (Yu2026diffusion?)-neuroinflammation; a post-COVID DWI review finding consistent longitudinal-fasciculus involvement (Jahanshahi et al. 2026)), there is a plausible case that a multimodal MRI panel (structural myelin proxy + diffusion microstructure + MRS neurochemistry) could serve as an objective research-stage biomarker for post-infectious brain involvement. (Raw certainty: 0.47, long-COVID/recovered population → discounted to 0.40.)
Evidence quality: The individual modalities each rest on multiple labs (≥6 organic labs for microstructural alteration, with direction heterogeneity), but no single multimodal protocol has been independently replicated, and the correlational design (physical/cognitive function) does not establish causality. Severity applicability unknown — cohorts were not severity-stratified.
Caveats: Standard clinical MRI cannot substitute for the specialised diffusion-NII or MRS post-processing used in research; the T1w/T2w substrate (remyelination vs gliosis vs iron) is unresolved (Section Elevated T1w/T2w Myelin Signal: Remyelination, Gliosis, or Iron?); direction of diffusion changes varies by tract and cohort; null and severity-gated findings (no overall FA difference in unselected cohorts (Arendt et al. 2026); no FA/MD/RD group effect in ME/CFS (Yu et al. 2025)) constrain diagnostic specificity. The reported 85% diagnostic accuracy of the cingulum DTI classifier (Wu et al. 2026) is a single-lab, single-cohort, in-sample figure with no independent validation set — a classic overfitting hazard — so it should be read as preliminary, not as an established diagnostic accuracy. Several primary findings also rest on a single research group (the Griffith NCNED cluster) in a small, selected sample (Section Elevated T1w/T2w Myelin Signal: Remyelination, Gliosis, or Iron?), so no individual modality’s specificity is yet established.
Falsifiable prediction: If multimodal MRI is a valid post-infectious biomarker, then a multi-site harmonised protocol will reproduce a stable regional signature (myelin-signal and diffusion patterns in SLF/midbrain/caudate + posterior-cingulate neurochemical profile) that discriminates long-COVID/ME/CFS from healthy controls and recovered controls with better accuracy than any single modality, and that signature will track clinical severity longitudinally. The speculation is falsified if independent multi-site replication fails to reproduce a consistent multimodal signature or if the signal tracks non-specific factors (age, deconditioning, mood) as strongly as it tracks post-infectious status.
Consequence: If validated, a multimodal MRI panel would give researchers and clinicians an objective, non-invasive way to detect and track brain involvement in long COVID and ME/CFS — useful for diagnosis and for measuring whether a treatment changes the brain — though it is currently a research tool, not a clinical test.
1.2 Functional MRI
NIH study and others show:
- Altered activation patterns during tasks
- TPJ dysfunction during effort tasks
- Connectivity changes
- Potential for task-based biomarkers
1.3 PET and SPECT
Metabolic and perfusion imaging:
- Regional hypometabolism
- Reduced cerebral blood flow
- Neuroinflammation markers (TSPO binding)
2 CSF Findings
Beyond the NIH catecholamine findings:
- Elevated inflammatory markers in some studies
- Altered protein profiles
- Potential autoantibodies
- Oligoclonal bands in subset Cerebrospinal Fluid Homocysteine Regland et al. reported elevated CSF homocysteine in all 12 FM/CFS patients examined, with significant correlation to fatiguability, while serum levels were normal (Regland et al. 1997). This CSF–serum discordance, if confirmed in larger ME/CFS-specific cohorts, would make CSF homocysteine a candidate CNS-compartment biomarker for a subset of patients with prominent neurocognitive and fatigue symptoms. The systematic review by Maksoud et al. (2023) of 101 ME/CFS biomarker studies did not identify homocysteine among biomarker candidates, reflecting the absence of larger validation studies rather than a confirmed negative finding (Maksoud et al. 2023). A future biomarker validation study could examine: serum vs CSF homocysteine, MTHFR genotype, serum and CSF B12, and clinical fatigue severity scores simultaneously across ME/CFS, long-COVID, and healthy control populations. Serum Homocysteine and Methylmalonic Acid (MMA) Serum homocysteine and methylmalonic acid (MMA) complete the B12/folate/methylation picture begun by the CSF finding above. Elevated serum homocysteine indicates impaired methylation, which may arise from B12 deficiency, folate deficiency, or MTHFR gene variants. MMA is the functional marker of cellular B12 deficiency: serum B12 concentrations can appear normal while tissue-level deficiency exists, but elevated MMA unmasks this functional shortfall. Both markers are accessible via standard laboratory testing and provide direct mechanistic context for the CSF homocysteine finding reported by Regland et al.
CSF Proteomics in Post-Viral Neuroinflammation Temporal CSF proteomics from herpes simplex encephalitis (HSE) patients provides a prospective model for identifying biomarkers of post-viral neuroinflammation relevant to ME/CFS Analysis of 890 CSF proteins in 48 HSE patients at three time points (days 0–9, 13–28, and 68+) revealed dynamic pathway activation: acute phase response, antimicrobial pattern recognition, and glycolysis/gluconeogenesis were strongly activated early and resolved by two weeks. Six proteins were significantly reduced in patients who developed anti-NMDAR autoantibodies: procathepsin H, heparin cofactor 2, complement factor I, protein AMBP, apolipoprotein A1, and polymeric immunoglobulin receptor. Apolipoprotein A1 (ApoA1) is of particular interest as it has been independently associated with NMDAR encephalitis outside the HSE context ApoA1 is the major protein component of high-density lipoprotein (HDL) and has known anti-inflammatory and neuroprotective functions; its reduction in CSF may indicate impaired neuroprotection or altered lipid metabolism in the CNS compartment. Neurofilament light chain (NFL), a marker of axonal damage, correlated with both impaired neurocognitive recovery and subsequent NMDAR autoantibody development in HSE patients (p=0.006) Whether NFL or ApoA1 are similarly altered in ME/CFS CSF remains untested — this represents a high-priority gap given the documented neuroinflammation in ME/CFS (see Section:cns).
3 Autonomic Function Tests
Quantifiable autonomic biomarkers:
- Tilt table testing: POTS, NMH, OH patterns
- Heart rate variability: Multiple parameters
- Sudomotor function: QSART abnormalities
- Pupillometry: Altered light reflexes
3.1 Renin and Aldosterone: RAAS Dysregulation
Renin-angiotensin-aldosterone system (RAAS) dysregulation is documented in POTS and ME/CFS, typically presenting as low renin with low-to-normal aldosterone despite evidence of reduced circulating volume — a paradoxical failure of the normal compensatory RAAS response This pattern contributes to orthostatic intolerance and is clinically actionable: fludrocortisone (a mineralocorticoid) targets the aldosterone pathway to augment sodium retention and expand plasma volume. Paired plasma renin activity and serum aldosterone, interpreted together with orthostatic testing results, constitute useful biomarkers in the autonomic evaluation of ME/CFS patients with prominent orthostatic intolerance.
4 Cognitive Testing Patterns
Neuropsychological profiles:
- Processing speed reduction (most consistent)
- Attention and working memory deficits
- Variable memory findings
- Pattern different from depression or anxiety