{"id":"W4410347660","doi":"10.1101/2025.05.13.653828","title":"Large Language Models Can Extract Metadata for Annotation of Human Neuroimaging Publications","year":2025,"lang":"en","type":"preprint","venue":"bioRxiv (Cold Spring Harbor Laboratory)","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"National Institute on Drug Abuse","keywords":"Metadata; Annotation; Neuroimaging; Computer science; Natural language processing; Information retrieval; Data science; Artificial intelligence; World Wide Web; Psychology; Neuroscience","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0004934752,0.0002940735,0.0003668593,0.0001970436,0.0001380083,0.00009829734,0.0006037666,0.0004204475,0.000004621933],"category_scores_gemma":[0.0004797777,0.000313357,0.0001670633,0.0002184629,0.0001149007,0.0000136822,0.0004579079,0.000274269,4.955102e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000315279,"about_ca_system_score_gemma":0.0004426355,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004963443,"about_ca_topic_score_gemma":0.00001834513,"domain_scores_codex":[0.9982122,0.00008425915,0.0004368726,0.0007382859,0.0001720083,0.0003564415],"domain_scores_gemma":[0.9979241,0.00004149121,0.0003502685,0.001152331,0.0004252908,0.000106487],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.00001805784,0.0001584067,0.0006387463,0.0004088631,0.0001762443,0.000001798632,0.00002403305,0.0000515155,0.9947839,0.001971235,0.00173623,0.00003098638],"study_design_scores_gemma":[0.001305888,0.0001923798,0.01138081,0.0003319471,0.0003721784,2.551757e-8,0.00005899657,0.003337684,0.9414659,0.00003762922,0.04054306,0.0009735767],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8329947,0.003979382,0.154254,0.0008472595,0.0006078925,0.001026134,0.006014369,0.0002240755,0.00005213757],"genre_scores_gemma":[0.9871339,0.0001098673,0.01196477,0.0001957017,0.0001979723,0.000233657,0.00004078169,0.00004149452,0.00008190969],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1541391,"threshold_uncertainty_score":0.9999319,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03010373277821184,"score_gpt":0.2940016720894056,"score_spread":0.2638979393111938,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}