TL;DR: A 2026 Molecular Psychiatry study found ethanolamine in cerebrospinal fluid (CSF) was lower during active major depression, increased after electroconvulsive therapy, replicated across 4 sites, and tracked with symptom severity.
Key Findings
- Active depression: 12.32 vs 14.07 µM: Active MDD patients had lower CSF ethanolamine than healthy controls (p = 0.00047) in a 380-person cohort. Remitted depression did not show the same gap — this looked tied to current illness.
- ECT raised ethanolamine 10.51 → 11.85 µM: In 13 patients, CSF ethanolamine increased after electroconvulsive therapy (p = 0.0071, d = 0.90). Small sample but the right direction.
- Replicated across 4 sites: Independent multicenter sample again showed lower CSF ethanolamine in MDD vs. controls (12.22 vs 14.56 µM; p = 0.0037; d = -0.91). Not a one-site result.
- Severity correlation: ρ = -0.29: Lower ethanolamine tracked higher HAM-D scores (p = 0.00015), with the clearest separation in moderate-to-severe depression regardless of medication.
- Dopamine chemistry was the closest neighbor: Ethanolamine correlated more strongly with HVA (dopamine metabolite, r = 0.40) than with the serotonin metabolite 5-HIAA (r = 0.26) — reframing depression chemistry away from a serotonin-only story.
- Rats: inflammation lowered it; supplementation reduced immobility: Forced-swim immobility dropped after oral ethanolamine, but open-field and sucrose-preference did not consistently improve. Lead, not supplement.
Source: Molecular Psychiatry (2026) | Ogawa et al.
Ethanolamine is a small molecule involved in membrane lipid biology. It is not a familiar depression target like serotonin, dopamine, inflammation, or stress hormones — and that obscurity is part of why the result deserves attention.
This study tied it to depressive state, treatment response, and depression-like behavior in animals across human cohorts and replication samples that most psychiatric biomarker work cannot match.
CSF Ethanolamine Was Lower in Active Depression Than Remission
The 380-person CSF cohort was the methodological backbone. CSF surrounds the brain and spinal cord, so it gives a closer view of brain-adjacent chemistry than blood. The dataset included healthy controls, active major depression, remitted major depression, bipolar disorder, and schizophrenia — the right structure for testing whether a marker is specific to current illness.
People with active MDD had lower CSF ethanolamine than healthy controls (12.32 vs 14.07 µM; p = 0.00047).
People with remitted depression did not. That state-specific pattern is more useful than a generic lifetime-diagnosis marker, because most proposed psychiatric biomarkers can only separate diagnostic groups in the model. Ethanolamine looked more tied to whether depression was currently active — the kind of signal a clinical state biomarker actually needs.
The schizophrenia and bipolar comparisons helped test whether ethanolamine was just lower across severe psychiatric illness. The schizophrenia comparison weakened after medication adjustment; active depression remained the clearest result.
CSF Ethanolamine Was Lowest in More Severe Depression
Within depression cases, lower ethanolamine correlated with worse HAM-D scores (ρ = -0.29; p = 0.00015).
The clearest separation appeared in moderate-to-severe depression, where ethanolamine remained lower regardless of medication status.
Specific symptom domains moved with it (activity, sleep, psychic anxiety, somatic anxiety) — tying the CSF finding to illness burden, not just diagnostic category.
CSF Ethanolamine Correlated Strongly With Dopamine Chemistry
The CSF analysis compared ethanolamine with monoamine metabolites — chemical readouts related to neurotransmitter systems.
The pattern in medication-free participants was unexpected:
- HVA (dopamine metabolite): r = 0.40, p = 6.7E-8 — the strongest correlation.
- 5-HIAA (serotonin metabolite): r = 0.26, p = 0.00064 — present but weaker.
- MHPG (noradrenaline): not significant.
This does not prove ethanolamine causes dopamine-related depression symptoms.
It places ethanolamine closer to motivation, reward, and psychomotor function than to a generic monoamine explanation. Depression chemistry has been quietly moving past serotonin-only framing for years, and this result fits that direction.
ECT Raised Ethanolamine in a Small Treatment Sample
Treatment evidence came from 13 patients receiving ECT. After treatment, average CSF ethanolamine increased from 10.51 to 11.85 µM (p = 0.0071; d = 0.90).
The 13-person sample cannot establish a definitive treatment-response biomarker. Even so, a CSF molecule that moves with treatment is more informative than one that only differs between groups at a single time point.
Ethanolamine change correlated most clearly with improvement on the HAM-D somatic-anxiety subscale. The data do not show ethanolamine explains ECT’s antidepressant effect — but they show CSF ethanolamine can track treatment.
Independent Cohort and Proteomics Supported the Ethanolamine Signal
Biomarker findings often weaken when they leave the original lab. The independent 4-site replication held: MDD averaged 12.22 µM vs.
14.56 µM in healthy controls (Cohen’s d = -0.91). That alone gives the result more credibility than most novel CSF biomarkers ever earn.
The proteomic subset added biological coherence. In 191 participants, ethanolamine correlated with 40 proteins after correction for multiple comparisons — the strongest hit was CHL1, a neural cell adhesion molecule involved in nervous system development and synaptic plasticity.
The broader protein set was enriched for axon-guidance biology, which concerns how neurons extend, connect, and maintain signaling paths. That fits ethanolamine’s membrane-biology role better than a vague stress-marker reading. The pathway is coherent, not just one isolated metabolite hit.
Animal Models of Depression Linked Ethanolamine to Monoamine Changes
The rat work asked two questions: does inflammation lower ethanolamine, and does adding it back change behavior?
Repeated lipopolysaccharide reduced CSF ethanolamine in a dose-dependent pattern, with parallel depression- and anxiety-like behavior. The model does not reproduce human depression, but it moved ethanolamine in the same direction as the human CSF data.
A 4-week oral ethanolamine experiment then tested whether supplementation could change behavior. The clearest change was reduced forced-swim immobility.
Open-field and sucrose-preference measures did not show the same broad improvement. That uneven behavioral profile supports ethanolamine as a possible lead for future drug development — not as an over-the-counter mood treatment.
CSF Ethanolamine Still Needs Larger Depression Validation
CSF is not easy to collect in routine depression care. A biomarker requiring lumbar puncture will not become a normal outpatient screen unless its clinical value is unusually strong.
The next translational step is whether ethanolamine, related metabolites, or linked proteins can be measured reliably in blood, or combined into a less invasive marker panel.
Larger treatment studies would also need to show ethanolamine changes consistently with symptom improvement across different therapies, not just ECT.
For now, this is a strong early biomarker argument.
Active depression was associated with lower CSF ethanolamine. ECT raised it in a small sample. Replication held across sites. Animal experiments suggested ethanolamine may be more than measurement noise.
That is more layers of evidence than most psychiatric biomarker candidates assemble — and the dopamine connection plus axon-guidance pathway gives the field something specific to test next.
Citation: DOI: 10.1038/s41380-026-03559-7; Ogawa et al; Ethanolamine as a potential biomarker and therapeutic target for depressive disorder; Molecular Psychiatry; 2026.
Study Design: Human CSF cohort + 4-site replication, ECT treatment sample, CSF proteomics, rat inflammation and supplementation experiments.
Sample Size: 380 main CSF cohort, 13 ECT patients, 191 proteomics subset, multicenter replication cohort.
Key Statistic: Active MDD vs. controls: 12.32 vs 14.07 µM (p = 0.00047); ECT pre/post: 10.51 → 11.85 µM (p = 0.0071, d = 0.90); replication d = -0.91.
Caveat: CSF collection limits clinical use; ECT sample small; cohorts Japanese; animal data does not support self-treatment.
