{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"institution":[{"name":"medRxiv"}],"indexed":{"date-parts":[[2026,1,16]],"date-time":"2026-01-16T13:54:47Z","timestamp":1768571687948,"version":"3.49.0"},"posted":{"date-parts":[[2025,10,2]]},"group-title":"Neurology","reference-count":42,"publisher":"openRxiv","license":[{"start":{"date-parts":[[2025,10,2]],"date-time":"2025-10-02T00:00:00Z","timestamp":1759363200000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by-nc-nd\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"accepted":{"date-parts":[[2025,11,7]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                <jats:p>Dementia encompasses diverse clinical syndromes where diseases of the brain can manifest as impaired cognitive abilities, such as in Alzheimer\u2019s disease (AD) and behavioral-variant frontotemporal dementia (bvFTD). The diversity of symptom presentations often results in challenges in diagnosis. Crucial clinical information remains in unstructured narrative notes within electronic health records (EHRs). We leverage large language models (LLMs) for symptom phenotyping from notes in UCSF Information Commons, focusing on patients with expert dementia syndrome diagnosed from a multidisciplinary team of specialists from the UCSF Memory and Aging Center. We developed a pipeline to extract findings in a validated structured output, clustered into symptom groups, and then classified patients into syndromes with traditional machine learning paradigms. From over 9,000 cross-referenced patients and over 350,000 specialty-related notes, matched cohorts of bvFTD (122 patients) and AD (170) syndromes were identified. From notes, 12,637 distinct symptom phrases were extracted, with clustering analysis revealing 51 symptom groups. A logistic regression model separated AD and bvFTD with an AUC of 0.83. Disinhibition and obsessive-compulsive behaviors favored bvFTD, while anxiety and visuospatial abnormalities favored AD. This novel approach, combining LLM-based structured information extraction with traditional interpretable prediction paradigms, demonstrates a promising approach for enhanced symptom characterization in dementia. Our findings suggest potential future applications in improving diagnostic accuracy, developing prediction models, and optimizing treatment strategies in dementia care.<\/jats:p>","DOI":"10.1101\/2025.10.01.25336815","type":"posted-content","created":{"date-parts":[[2025,10,2]],"date-time":"2025-10-02T20:45:13Z","timestamp":1759437913000},"source":"Crossref","is-referenced-by-count":0,"title":["Characterizing Dementia Phenotypes from Unstructured EHR Notes with Generative AI and Interpretable Machine Learning"],"prefix":"10.64898","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-4745-0714","authenticated-orcid":false,"given":"Alice S.","family":"Tang","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4653-3772","authenticated-orcid":false,"given":"Billy Z.D.","family":"Zeng","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3611-0848","authenticated-orcid":false,"given":"Katherine P.","family":"Rankin","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Bruce","family":"Miller","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Maria Luisa","family":"Gorno-Tempini","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"William W.","family":"Seeley","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Howard J.","family":"Rosen","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Gil D.","family":"Rabinovici","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7393-5120","authenticated-orcid":false,"given":"Tomiko T.","family":"Oskotsky","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-7246-6083","authenticated-orcid":false,"given":"Marina","family":"Sirota","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8512-0113","authenticated-orcid":false,"given":"Pedro","family":"Pinheiro-Chagas","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"54368","reference":[{"key":"2025111006050569000_2025.10.01.25336815v2.1","unstructured":"Emmady, P. D. , Schoo, C. & Tadi, P. Major Neurocognitive Disorder (Dementia). in StatPearls (StatPearls Publishing, Treasure Island (FL), 2025)."},{"key":"2025111006050569000_2025.10.01.25336815v2.2","doi-asserted-by":"publisher","DOI":"10.1001\/jama.2019.4782"},{"key":"2025111006050569000_2025.10.01.25336815v2.3","first-page":"891","article-title":"Racial disparities and temporal trends in dementia misdiagnosis risk in the United States","volume":"5","year":"2019","journal-title":"Alzheimers Dement. N. Y. N"},{"key":"2025111006050569000_2025.10.01.25336815v2.4","doi-asserted-by":"crossref","unstructured":"Qian, W. , Schweizer, T. , Munoz, D. & Fischer, C. E. O3\u25a104\u25a106: Misdiagnosis of Alzheimer\u2019s Disease: Inconsistencies Between Clinical Diagnosis and Neuropathological Confirmation. Alzheimers Dement. 12, (2016).","DOI":"10.1016\/j.jalz.2016.06.529"},{"key":"2025111006050569000_2025.10.01.25336815v2.5","doi-asserted-by":"publisher","DOI":"10.1097\/WAD.0b013e3181a6bebc"},{"key":"2025111006050569000_2025.10.01.25336815v2.6","doi-asserted-by":"crossref","first-page":"654","DOI":"10.1016\/j.jalz.2016.09.015","article-title":"Mixed neuropathologies and estimated rates of clinical progression in a large autopsy sample","volume":"13","year":"2017","journal-title":"Alzheimers Dement. J. Alzheimers Assoc"},{"key":"2025111006050569000_2025.10.01.25336815v2.7","doi-asserted-by":"crossref","first-page":"535","DOI":"10.1016\/j.jalz.2018.02.018","article-title":"NIA-AA Research Framework: Toward a biological definition of Alzheimer\u2019s disease","volume":"14","year":"2018","journal-title":"Alzheimers Dement. J. Alzheimers Assoc"},{"key":"2025111006050569000_2025.10.01.25336815v2.8","doi-asserted-by":"publisher","DOI":"10.1097\/WAD.0b013e31815c3445"},{"key":"2025111006050569000_2025.10.01.25336815v2.9","doi-asserted-by":"crossref","first-page":"101771","DOI":"10.1016\/j.artmed.2019.101771","article-title":"Predicting dementia with routine care EMR data","volume":"102","year":"2020","journal-title":"Artif. Intell. Med"},{"key":"2025111006050569000_2025.10.01.25336815v2.10","doi-asserted-by":"crossref","first-page":"379","DOI":"10.1038\/s43587-024-00573-8","article-title":"Leveraging electronic health records and knowledge networks for Alzheimer\u2019s disease prediction and sex-specific biological insights","volume":"4","year":"2024","journal-title":"Nat. Aging"},{"key":"2025111006050569000_2025.10.01.25336815v2.11","doi-asserted-by":"publisher","DOI":"10.1080\/08952841.2020.1735925"},{"key":"2025111006050569000_2025.10.01.25336815v2.12","doi-asserted-by":"crossref","first-page":"1847","DOI":"10.1038\/s41591-024-03074-8","article-title":"Harnessing EHR data for health research","volume":"30","year":"2024","journal-title":"Nat. Med"},{"key":"2025111006050569000_2025.10.01.25336815v2.13","doi-asserted-by":"publisher","DOI":"10.2147\/JMDH.S17773"},{"key":"2025111006050569000_2025.10.01.25336815v2.14","doi-asserted-by":"crossref","first-page":"866","DOI":"10.1002\/alz.12253","article-title":"The search for a convenient procedure to detect one of the earliest signs of Alzheimer\u2019s disease: A systematic review of the prediction of brain amyloid status","volume":"17","year":"2021","journal-title":"Alzheimers Dement"},{"key":"2025111006050569000_2025.10.01.25336815v2.15","doi-asserted-by":"publisher","DOI":"10.4088\/JCP.08m04181"},{"key":"2025111006050569000_2025.10.01.25336815v2.16","doi-asserted-by":"crossref","first-page":"5905","DOI":"10.1002\/alz.13427","article-title":"Artificial intelligence for dementia genetics and omics","volume":"19","year":"2023","journal-title":"Alzheimers Dement"},{"key":"2025111006050569000_2025.10.01.25336815v2.17","doi-asserted-by":"publisher","DOI":"10.1126\/sciadv.abo6764"},{"key":"2025111006050569000_2025.10.01.25336815v2.18","doi-asserted-by":"publisher","DOI":"10.1093\/hmg\/ddp406"},{"key":"2025111006050569000_2025.10.01.25336815v2.19","doi-asserted-by":"publisher","DOI":"10.1002\/alz.12011"},{"key":"2025111006050569000_2025.10.01.25336815v2.20","doi-asserted-by":"publisher","DOI":"10.1016\/j.neuroimage.2015.01.048"},{"key":"2025111006050569000_2025.10.01.25336815v2.21","doi-asserted-by":"publisher","DOI":"10.1016\/j.neuroimage.2017.03.057"},{"key":"2025111006050569000_2025.10.01.25336815v2.22","doi-asserted-by":"publisher","DOI":"10.1007\/s10462-024-10961-6"},{"key":"2025111006050569000_2025.10.01.25336815v2.23","first-page":"1","article-title":"Deep representation learning of electronic health records to unlock patient stratification at scale","volume":"3","year":"2020","journal-title":"Npj Digit. Med"},{"key":"2025111006050569000_2025.10.01.25336815v2.24","doi-asserted-by":"crossref","first-page":"104442","DOI":"10.1016\/j.jbi.2023.104442","article-title":"AD-BERT: Using pre-trained language model to predict the progression from mild cognitive impairment to Alzheimer\u2019s disease","volume":"144","year":"2023","journal-title":"J. Biomed. Inform"},{"key":"2025111006050569000_2025.10.01.25336815v2.25","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-022-28273-0"},{"key":"2025111006050569000_2025.10.01.25336815v2.26","doi-asserted-by":"crossref","first-page":"1438","DOI":"10.1093\/jamia\/ocad054","article-title":"Automatic extraction of social determinants of health from medical notes of chronic lower back pain patients","volume":"30","year":"2023","journal-title":"J. Am. Med. Inform. Assoc. JAMIA"},{"key":"2025111006050569000_2025.10.01.25336815v2.27","doi-asserted-by":"crossref","first-page":"ooad112","DOI":"10.1093\/jamiaopen\/ooad112","article-title":"Topic modeling on clinical social work notes for exploring social determinants of health factors","volume":"7","year":"2024","journal-title":"JAMIA Open"},{"key":"2025111006050569000_2025.10.01.25336815v2.28","doi-asserted-by":"crossref","first-page":"111223","DOI":"10.1016\/j.exger.2020.111223","article-title":"Motor signs in Alzheimer\u2019s disease and vascular dementia: Detection through natural language processing, co-morbid features and relationship to adverse outcomes","volume":"146","year":"2021","journal-title":"Exp. Gerontol"},{"key":"2025111006050569000_2025.10.01.25336815v2.29","doi-asserted-by":"crossref","first-page":"399","DOI":"10.1017\/S1041610222000126","article-title":"Suicidal ideation in dementia: associations with neuropsychiatric symptoms and subtype diagnosis","volume":"34","year":"2022","journal-title":"Int. Psychogeriatr"},{"key":"2025111006050569000_2025.10.01.25336815v2.30","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2502.01789"},{"key":"2025111006050569000_2025.10.01.25336815v2.31","doi-asserted-by":"crossref","first-page":"105401","DOI":"10.1016\/j.ebiom.2024.105401","article-title":"Enhancing early detection of cognitive decline in the elderly: a comparative study utilizing large language models in clinical notes","volume":"109","year":"2024","journal-title":"eBioMedicine"},{"key":"2025111006050569000_2025.10.01.25336815v2.32","doi-asserted-by":"crossref","first-page":"833","DOI":"10.3233\/JAD-190924","article-title":"Alzheimer\u2019s Disease or Behavioral Variant Frontotemporal Dementia? Review of Key Points Toward an Accurate Clinical and Neuropsychological Diagnosis","volume":"73","year":"2020","journal-title":"J. Alzheimers Dis. JAD"},{"key":"2025111006050569000_2025.10.01.25336815v2.33","doi-asserted-by":"crossref","first-page":"1572947","DOI":"10.3389\/fcomm.2025.1572947","article-title":"Energy costs of communicating with AI","volume":"10","year":"2025","journal-title":"Front. Commun"},{"key":"2025111006050569000_2025.10.01.25336815v2.34","doi-asserted-by":"crossref","first-page":"1373341","DOI":"10.3389\/fneur.2024.1373341","article-title":"Linguistic changes in neurodegenerative diseases relate to clinical symptoms","volume":"15","year":"2024","journal-title":"Front. Neurol"},{"key":"2025111006050569000_2025.10.01.25336815v2.35","first-page":"249","volume-title":"in International Review of Neurobiology","volume":"149","year":"2019"},{"key":"2025111006050569000_2025.10.01.25336815v2.36","doi-asserted-by":"crossref","first-page":"51","DOI":"10.1016\/j.jns.2016.11.033","article-title":"Clinical characteristics of parkinsonism in frontotemporal dementia according to subtypes","volume":"372","year":"2017","journal-title":"J. Neurol. Sci"},{"key":"2025111006050569000_2025.10.01.25336815v2.37","doi-asserted-by":"crossref","first-page":"e14628","DOI":"10.1002\/alz.14628","article-title":"Parkinsonism in Alzheimer\u2019s disease without Lewy bodies in association with nigral neuron loss: A data-driven clinicopathologic study","volume":"21","year":"2025","journal-title":"Alzheimers Dement. J. Alzheimers Assoc"},{"key":"2025111006050569000_2025.10.01.25336815v2.38","doi-asserted-by":"crossref","first-page":"2474","DOI":"10.1212\/WNL.96.15_supplement.2474","article-title":"History of Psychiatric Disease Inversely Correlates with Age of Onset in Alzheimer\u2019s Disease. (2474)","volume":"96","year":"2021","journal-title":"Neurology"},{"key":"2025111006050569000_2025.10.01.25336815v2.39","doi-asserted-by":"publisher","DOI":"10.1016\/j.biopsych.2021.11.025"},{"key":"2025111006050569000_2025.10.01.25336815v2.40","doi-asserted-by":"publisher","DOI":"10.1016\/j.jad.2015.09.069"},{"key":"2025111006050569000_2025.10.01.25336815v2.41","first-page":"1447","article-title":"The Impact of AI Scribes on Streamlining Clinical Documentation: A Systematic Review","volume":"13","year":"2025","journal-title":"Healthc. Basel Switz"},{"key":"2025111006050569000_2025.10.01.25336815v2.42","doi-asserted-by":"crossref","first-page":"236","DOI":"10.1186\/s12911-025-03061-0","article-title":"Evaluating the performance of artificial intelligence-based speech recognition for clinical documentation: a systematic review","volume":"25","year":"2025","journal-title":"BMC Med. Inform. Decis. Mak"}],"container-title":[],"original-title":[],"link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.1101\/2025.10.01.25336815","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,1,15]],"date-time":"2026-01-15T20:47:28Z","timestamp":1768510048000},"score":1,"resource":{"primary":{"URL":"http:\/\/medrxiv.org\/lookup\/doi\/10.1101\/2025.10.01.25336815"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,10,2]]},"references-count":42,"URL":"https:\/\/doi.org\/10.1101\/2025.10.01.25336815","relation":{},"subject":[],"published":{"date-parts":[[2025,10,2]]},"subtype":"preprint"}}