{
  "abstract": "Normative age models are trained to predict chronological age in a nominally healthy cohort. Applied to patients, they deviate, and the gap between predicted and chronological age is read as disease risk. Here, we attribute the disease-related effect size of the age gap directly to individual training samples, rather than using a prediction-level loss as the attribution target. For Cohen's $d$, the resulting closed-form influence functional, validated against leave-one-out retraining, ranks training samples by their effect on held-out case-control separation. Across four diseases and two biomarker modalities in UK Biobank, removing the 10% most influential training samples raises held-out disease-related effect size in every seed. It more than doubles the metabolomic-age effect for type-2 diabetes and raises the brain-age effect for multiple sclerosis by roughly a third. Random removal leaves effect size flat even at 50% removal, confirming the gain comes from which samples are removed, not how many. Flagged subjects carry subclinical cardiometabolic burden that diagnosis-based exclusion misses, on markers the model never sees. For type-2 diabetes, where the method gains most, the marker recovered is HbA1c, the standard measure of blood sugar control. We release pyinfluence, our influence-function package, for reproducibility and reuse.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": "2609.07729",
  "author_total": 2,
  "authors": [
    {
      "name": "Jakob Snel"
    },
    {
      "name": "Marc-Andre Schulz"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/00rx7g0b182a6.bib",
  "cited_by": 0,
  "doi": "10.48550/arxiv.2609.07729",
  "fulltext_url": null,
  "id": "00rx7g0b182a6",
  "issns": [],
  "landing_url": "https://arxiv.org/abs/2609.07729",
  "language": null,
  "licence": "LicenseRef-arXiv-nonexclusive-1.0",
  "licence_label": "arXiv licence",
  "licence_source": "article",
  "licence_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
  "metadata_licence": "CC0-1.0",
  "native_subjects": [
    [
      "arxiv",
      "cs.LG"
    ],
    [
      "arxiv",
      "q-bio.QM"
    ],
    [
      "openalex",
      "1702|Machine Learning in Healthcare"
    ]
  ],
  "openalex": "W7212045362",
  "pdf_url": "https://arxiv.org/pdf/2609.07729",
  "pmcid": null,
  "pmid": null,
  "published": "2026-09-18",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/00rx7g0b182a6.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:23:08Z",
      "native_id": "2609.07729",
      "source": "arxiv",
      "url": "https://arxiv.org/abs/2609.07729"
    }
  ],
  "subjects": [
    {
      "field": "biology",
      "origin": "arxiv:q-bio.QM",
      "subfield": "bioinformatics"
    }
  ],
  "title": "Attributing Cohen's d: Training Data Attribution for Disease-Related Effects in Normative Age Biomarkers",
  "title_text": "Attributing Cohen's d: Training Data Attribution for Disease-Related Effects in Normative Age Biomarkers",
  "type": "preprint",
  "updated_at": "2026-10-30T12:23:08Z",
  "url": "https://research.draconitic.com/paper/00rx7g0b182a6",
  "venue": "arXiv",
  "versions": []
}