{"id":"W4416402424","doi":"10.3389/fdgth.2025.1659366","title":"CharMark: character-level Markov modeling for interpretable linguistic biomarkers of cognitive decline","year":2025,"lang":"en","type":"article","venue":"Frontiers in Digital Health","topic":"Neurobiology of Language and Bilingualism","field":"Neuroscience","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"University of Pittsburgh; National Institute on Aging; National Science Foundation","keywords":"Cognitive decline; Discriminative model; Dementia; Cognition; Interpretation (philosophy); Hidden Markov model; Character (mathematics); Logistic regression; Computational linguistics","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003697379,0.0001821748,0.0004384184,0.0003553447,0.00008678258,0.00003923273,0.0002919399,0.00008147535,0.000003886109],"category_scores_gemma":[0.003349683,0.0001758023,0.0001069156,0.0003450412,0.0001533765,0.000170736,0.0001369143,0.0001743023,0.000001492668],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00007666352,"about_ca_system_score_gemma":0.0002762026,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003707136,"about_ca_topic_score_gemma":0.000009234628,"domain_scores_codex":[0.9982669,0.00006162578,0.0005960111,0.0004968167,0.0001097643,0.0004688914],"domain_scores_gemma":[0.9989356,0.0005368639,0.0001811273,0.0001799135,0.0000846597,0.00008180486],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0121104,0.003007645,0.05712027,0.004561309,0.0003211462,0.0004452811,0.01071878,0.0001529378,0.01461839,0.001482242,0.01277162,0.88269],"study_design_scores_gemma":[0.02995194,0.004712586,0.004814169,0.01450377,0.0002466805,0.0003026043,0.01009166,0.7021239,0.1313739,0.0858999,0.01199781,0.003981067],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8771414,0.0008877879,0.1116303,0.0007748987,0.003220181,0.00133032,0.001350389,0.00007996873,0.003584786],"genre_scores_gemma":[0.9943677,0.0001056851,0.001725885,0.003099501,0.0000582509,0.00003375324,0.00007054293,0.00001818113,0.0005204618],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8787089,"threshold_uncertainty_score":0.7169012,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03546149579726589,"score_gpt":0.3344673724246631,"score_spread":0.2990058766273972,"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."}}