IFN-beta → BDNF → MDD

• Alterations in serum melatonin, IL-6 and BDNF levels and their clinical correlations in first-episode major depressive disorder

  B→C
  Ann Gen Psychiatry· 2026· PMID 41906159· doi
• Potential value of serum brain-derived neurotrophic factor, vascular endothelial growth factor, and S100B for identifying major depressive disorder in knee osteoarthritis patients

  B→C
  Front Psychiatry· 2022· PMID 36386998· doi
• The Effect of IFN-β Treatment on Plasma Levels of BDNF and IL-6 in Relapsing-Remitting Multiple Sclerosis Patients

  A→B
  Neuroimmunomodulation· 2021· PMID 34182566· doi
• Reduced serum BDNF levels are associated with the increased risk for developing MDD: a case–control study with or without antidepressant therapy

  B→C
  BMC Res Notes· 2020· PMID 32085720· doi

Each row is one structured triple extracted by deepseek-v4-flash (prompt v1_2026_05_07, source: configured). Confidence is the LLM's self-rating.

• IFN-betaupregulatesBDNFA→B
  "in relapsing-remitting multiple sclerosis patients"polarity: negativeconfidence 0.90· extraction run #25· claim #144128
• BDNFassociated_withmajor depressive disorderB→C
  "as early risk assessment marker"confidence 0.70· extraction run #18· claim #75381
• BDNFpredictsMDDB→C
  "ROC analysis in drug-naïve patients with AUC 0.821"confidence 0.80· extraction run #18· claim #75380
• BDNFpredictsMDDB→C
  "in knee osteoarthritis patients over 3-year follow-up"confidence 0.85· extraction run #24· claim #136588
• BDNFassociated_withMDDB→C
  "first-episode MDD patients"confidence 0.95· extraction run #24· claim #140115

No claims match this filter.

edge_support_log_ab

log(ab + 1) = 0.693

edge_support_log_bc

log(bc + 1) = 1.386

novelty_factor

1.000 (no direct drug→outcome evidence — full novelty)

direct_a_to_c_count

0

Embeddings are used at the L1 filtering stage to prune off-topic abstracts before claim extraction. They do not directly score the hypothesis.

model

bge-m3

dim

1024

max_length

8192 tokens

similarity_scores

No per-hypothesis cosine scores stored.

method

swanson_abc_log_evidence_with_novelty_penalty

formula

score = min((log(ab+1) + log(bc+1)) * novelty / 10, 1.0); novelty = 1.0 if direct==0 else 0.7 / (direct+1)

raw score

0.2079

outreach score

0.2079

quality tier

mid