{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,31]],"date-time":"2026-08-31T06:15:49Z","timestamp":1788156949373,"version":"build-2803163510"},"reference-count":36,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2026,8,31]],"date-time":"2026-08-31T00:00:00Z","timestamp":1788134400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Artif. Intell."],"abstract":"<jats:sec>\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Early diagnosis is critical for effective management of Alzheimer's disease (AD). While prior studies have shown that speech features can be indicative of AD, most existing work consolidates multiple biomarkers, making it difficult to isolate the contribution of individual features.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Objective<\/jats:title>\n                    <jats:p>This study systematically isolates individual speech biomarkers to quantify their distinct contributions to AD classification performance of language models (LMs) and determine whether targeted biomarker selection improves over consolidated feature sets.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>We processed speech transcriptions from DementiaBank to surface discriminatory speech biomarkers\u2014verbal pauses, disfluencies, and unintelligible words. We then fine-tuned and evaluated lightweight LMs (BERT, AlBERT, and DistilBERT) on these biomarker-conditioned transcripts for automatic AD classification.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>Pauses emerged as the most discriminatory speech biomarker (F1 = 0.8326), significantly outperforming the No-biomarker baseline and other biomarkers. Combining all biomarkers degraded performance relative to pauses alone on average across models, though the effect was model-dependent: BERT's All-biomarkers condition exceeded its own No-biomarker baseline, suggesting that feature combination benefits higher-capacity models. BERT yielded the best performance (F1 = 0.8154) across conditions.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusions<\/jats:title>\n                    <jats:p>Selective use of speech biomarkers such as pauses can meaningfully improve AD detection with lightweight LMs, suggesting that targeted biomarker selection may offer a more interpretable and clinically actionable path than broad feature consolidation.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.3389\/frai.2026.1875702","type":"journal-article","created":{"date-parts":[[2026,8,31]],"date-time":"2026-08-31T05:40:45Z","timestamp":1788154845000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Utilising speech-derived biomarkers to detect Alzheimer's disease with BERT-based language models: a machine learning study"],"prefix":"10.3389","volume":"9","author":[{"given":"Zara","family":"Khanna","sequence":"first","affiliation":[{"name":"The Institute for Digital Medicine (WisDM), Yong Loo Lin School of Medicine, National University of Singapore","place":["Singapore, Singapore"]},{"name":"The N.1 Institute for Health, National University of Singapore","place":["Singapore, Singapore"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Dean","family":"Ho","sequence":"additional","affiliation":[{"name":"The Institute for Digital Medicine (WisDM), Yong Loo Lin School of Medicine, National University of Singapore","place":["Singapore, Singapore"]},{"name":"The N.1 Institute for Health, National University of Singapore","place":["Singapore, Singapore"]},{"name":"Department of Biomedical Engineering, College of Design and Engineering, National University of Singapore","place":["Singapore, Singapore"]},{"name":"Department of Pharmacology, Yong Loo Lin School of Medicine, National University of Singapore","place":["Singapore, Singapore"]},{"name":"The Bia-Echo Asia Centre for Reproductive Longevity and Equality, Yong Loo Lin School of Medicine, National University of Singapore","place":["Singapore, Singapore"]},{"name":"Singapore's Health District @ Queenstown, Yong Loo Lin School of Medicine, National University of Singapore","place":["Singapore, Singapore"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Alexandria","family":"Remus","sequence":"additional","affiliation":[{"name":"The Institute for Digital Medicine (WisDM), Yong Loo Lin School of Medicine, National University of Singapore","place":["Singapore, Singapore"]},{"name":"The N.1 Institute for Health, National University of Singapore","place":["Singapore, Singapore"]},{"name":"Department of Biomedical Engineering, College of Design and Engineering, National University of Singapore","place":["Singapore, Singapore"]},{"name":"Heat Resilience & Performance Centre, Yong Loo Lin School of Medicine, National University of Singapore","place":["Singapore, Singapore"]},{"name":"Human Potential Translational Research Programme, Yong Loo Lin School of Medicine, National University of Singapore","place":["Singapore, Singapore"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Marlena","family":"Raczkowska","sequence":"additional","affiliation":[{"name":"The Institute for Digital Medicine (WisDM), Yong Loo Lin School of Medicine, National University of Singapore","place":["Singapore, Singapore"]},{"name":"The N.1 Institute for Health, National University of Singapore","place":["Singapore, Singapore"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Railey","family":"Montalan","sequence":"additional","affiliation":[{"name":"Office of the Deputy President (Research and Technology), National University of Singapore","place":["Singapore, Singapore"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Pramit","family":"Saha","sequence":"additional","affiliation":[{"name":"Department of Engineering Science, University of Oxford","place":["Oxford, United Kingdom"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1965","published-online":{"date-parts":[[2026,8,31]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"3727","DOI":"10.1093\/brain\/awt269","article-title":"Connected speech as a marker of disease progression in autopsy-proven Alzheimer's disease","volume":"136","author":"Ahmed","year":"2013","journal-title":"Brain"},{"key":"B2","doi-asserted-by":"publisher","first-page":"e14530","DOI":"10.1002\/alz.14530","article-title":"Natural language processing-based classification of early Alzheimer's disease from connected speech","volume":"21","author":"Balabin","year":"2025","journal-title":"Alzheimers Dement"},{"key":"B3","doi-asserted-by":"publisher","first-page":"e70051","DOI":"10.1111\/coin.70051","article-title":"Speech2Dementia: a novel deep learning framework integrating enhanced CNN and large language models for automatic detection of Alzheimer's dementia","volume":"41","year":"2025","journal-title":"Expert Syst"},{"key":"B4","doi-asserted-by":"publisher","first-page":"158","DOI":"10.1016\/j.schres.2020.04.032","article-title":"Language as a biomarker for psychosis: a natural language processing approach","volume":"226","author":"Corcoran","year":"2020","journal-title":"Schizophr. Res"},{"key":"B5","first-page":"4171","article-title":"\u201cBERT: pre-training of deep bidirectional transformers for language understanding,\u201d","volume-title":"Proceedings of the 2019 conference of the North American chapter of the association for computational linguistics: human language technologies, Vol. 1","author":"Devlin","year":"2019"},{"key":"B6","doi-asserted-by":"publisher","first-page":"325","DOI":"10.1007\/s10462-024-10961-6","article-title":"Speech based detection of Alzheimer's disease: a survey of AI techniques, datasets and challenges","volume":"57","author":"Ding","year":"2024","journal-title":"Artif. Intell. Rev"},{"key":"B7","article-title":"The Llama 3 herd of models","author":"Dubey","year":"2024","journal-title":"arXiv"},{"key":"B8","doi-asserted-by":"publisher","first-page":"47","DOI":"10.3390\/nu15010047","article-title":"Nutrition strategies promoting healthy aging: from improvement of cardiovascular and brain health to prevention of age-associated diseases","volume":"15","author":"Fekete","year":"2022","journal-title":"Nutrients"},{"key":"B9","doi-asserted-by":"publisher","first-page":"407","DOI":"10.3233\/JAD-150520","article-title":"Linguistic features identify Alzheimer's disease in narrative speech","volume":"49","author":"Fraser","year":"2015","journal-title":"J Alzheimers Dis"},{"key":"B10","doi-asserted-by":"publisher","first-page":"5841","DOI":"10.1109\/ICASSP.2019.8682690","article-title":"\u201cAutomatic diagnosis of Alzheimer's disease using neural network language models,\u201d","author":"Fritsch","year":"2019","journal-title":"ICASSP 2019\u20132019 IEEE international conference on acoustics, speech and signal processing (ICASSP)"},{"key":"B11","doi-asserted-by":"publisher","first-page":"642517","DOI":"10.3389\/fcomp.2021.642517","article-title":"Crossing the \u201cCookie Theft\u201d corpus chasm: applying what BERT learns from outside data to the ADReSS challenge dementia detection task","volume":"3","author":"Guo","year":"2021","journal-title":"Front. Comput. Sci"},{"key":"B12","article-title":"Gemma 3 technical report","author":"Kamath","year":"2025","journal-title":"arXiv"},{"key":"B13","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1186\/s13636-015-0052-y","article-title":"Evaluation of linguistic and prosodic features for detection of Alzheimer's disease in Turkish conversational speech","volume":"2015","author":"Khodabakhsh","year":"2015","journal-title":"EURASIP J. Audio Speech Music Process"},{"key":"B14","article-title":"\u201cALBERT: a lite BERT for self-supervised learning of language representations,\u201d","volume-title":"International conference on learning representations (ICLR 2020)","author":"Lan","year":"2020"},{"key":"B15","doi-asserted-by":"publisher","first-page":"426","DOI":"10.1044\/2022_AJSLP-22-00281","article-title":"DementiaBank: theoretical rationale, protocol, and illustrative analyses","volume":"32","author":"Lanzi","year":"2023","journal-title":"Am. J. Speech Lang. Pathol"},{"key":"B16","doi-asserted-by":"publisher","first-page":"431","DOI":"10.1146\/annurev.pu.13.050192.002243","article-title":"Cognitive impairment: dementia and Alzheimer's disease","volume":"13","author":"Larson","year":"1992","journal-title":"Annu. Rev. Public Health"},{"key":"B17","doi-asserted-by":"publisher","first-page":"24720","DOI":"10.1038\/s41598-025-08362-y","article-title":"Deep ensemble learning with transformer models for enhanced Alzheimer's disease detection","volume":"15","author":"Latif","year":"2025","journal-title":"Sci. Rep"},{"key":"B18","first-page":"360","article-title":"TRESTLE: toolkit for reproducible execution of speech, text and language experiments","volume":"2023","author":"Li","year":"2023","journal-title":"AMIA Summits Transl. Sci. Proc"},{"key":"B19","doi-asserted-by":"publisher","first-page":"601","DOI":"10.7326\/0003-4819-159-9-201311050-00730","article-title":"Screening for cognitive impairment in older adults: a systematic review for the US Preventive Services Task Force","volume":"159","author":"Lin","year":"2013","journal-title":"Ann. Intern. Med"},{"key":"B20","doi-asserted-by":"publisher","first-page":"477","DOI":"10.3390\/brainsci13030477","article-title":"Efficient pause extraction and encode strategy for Alzheimer's disease detection using only acoustic features from spontaneous speech","volume":"13","author":"Liu","year":"2023","journal-title":"Brain Sci"},{"key":"B21","doi-asserted-by":"publisher","first-page":"44","DOI":"10.1007\/s12559-013-9229-9","article-title":"On automatic diagnosis of Alzheimer's disease based on spontaneous speech analysis and emotional temperature","volume":"7","author":"Lopez-de Ipi na","year":"2015","journal-title":"Cognit. Comput"},{"key":"B22","first-page":"19","article-title":"\u201cA method for analysis of patient speech in dialogue for dementia detection,\u201d","volume-title":"Proceedings of the LREC 2018 workshop, resources and processing of linguistic, para-linguistic and extra-linguistic data from people with various forms of cognitive\/psychiatric impairments (RaPID-2)","author":"Luz","year":"2018"},{"key":"B23","doi-asserted-by":"publisher","first-page":"373","DOI":"10.3233\/JAD-160507","article-title":"Toward the automation of diagnostic conversation analysis in patients with memory complaints","volume":"58","author":"Mirheidari","year":"2017","journal-title":"J. Alzheimers Dis"},{"key":"B24","first-page":"2825","article-title":"Scikit-learn: machine learning in Python","volume":"12","author":"Pedregosa","year":"2011","journal-title":"J. Mach. Learn. Res"},{"key":"B25","doi-asserted-by":"publisher","first-page":"1784","DOI":"10.1093\/jamia\/ocaa174","article-title":"A systematic literature review of automatic Alzheimer's disease detection from speech and language","volume":"27","author":"Petti","year":"2020","journal-title":"J. Am. Med. Inf. Assoc"},{"key":"B26","doi-asserted-by":"publisher","first-page":"276","DOI":"10.1044\/2021_PERSP-21-00174","article-title":"Speech as a biomarker: opportunities, interpretability, and challenges","volume":"7","author":"Ramanarayanan","year":"2022","journal-title":"Perspect. ASHA Spec. Interest Groups"},{"key":"B27","doi-asserted-by":"publisher","first-page":"866","DOI":"10.1016\/j.neuropsychologia.2010.12.024","article-title":"Affective prosody: what do comprehension errors tell us about hemispheric lateralization of emotions, sex and aging effects, and the role of cognitive appraisal","volume":"49","author":"Ross","year":"2011","journal-title":"Neuropsychologia"},{"key":"B28","article-title":"\u201cDistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter,\u201d","volume-title":"5th workshop on energy efficient machine learning and cognitive computing-NeurIPS","author":"Sanh","year":"2019"},{"key":"B29","doi-asserted-by":"publisher","first-page":"463","DOI":"10.1097\/COH.0b013e32833ed177","article-title":"What are biomarkers?","volume":"5","author":"Strimbu","year":"2010","journal-title":"Curr. Opin. HIV AIDS"},{"key":"B30","article-title":"\u201cAttention is all you need,\u201d","volume-title":"Advances in neural information processing systems, Vol. 30","author":"Vaswani","year":"2017"},{"key":"B31","doi-asserted-by":"publisher","first-page":"3117","DOI":"10.21437\/Interspeech.2017-112","article-title":"\u201cManual and automatic transcriptions in dementia detection from speech,\u201d","author":"Weiner","year":"2017","journal-title":"Proceedings of interspeech"},{"key":"B32","doi-asserted-by":"publisher","first-page":"38","DOI":"10.18653\/v1\/2020.emnlp-demos.6","article-title":"\u201cTransformers: state-of-the-art natural language processing,\u201d","author":"Wolf","year":"2020","journal-title":"Proceedings of the 2020 conference on empirical methods in natural language processing: system demonstrations"},{"key":"B33","doi-asserted-by":"publisher","first-page":"2162","DOI":"10.21437\/Interspeech.2020-2516","article-title":"\u201cDisfluencies and fine-tuning pre-trained language models for detection of Alzheimer's disease,\u201d","author":"Yuan","year":"","journal-title":"Proceedings of interspeech"},{"key":"B34","doi-asserted-by":"publisher","first-page":"624488","DOI":"10.3389\/fcomp.2020.624488","article-title":"Pauses for detection of Alzheimer's disease","volume":"2","author":"Yuan","year":"","journal-title":"Front. Comput. Sci"},{"key":"B35","doi-asserted-by":"publisher","first-page":"9851","DOI":"10.1007\/s13042-024-02443-6","article-title":"A comprehensive survey on pretrained foundation models: a history from BERT to ChatGPT","volume":"16","author":"Zhou","year":"2025","journal-title":"Int. J. Mach. Learn. Cybern"},{"key":"B36","doi-asserted-by":"publisher","first-page":"9","DOI":"10.1038\/s44387-025-00011-z","article-title":"Large language models for disease diagnosis: a scoping review","volume":"1","author":"Zhou","year":"2025","journal-title":"npj Artif. Intell"}],"container-title":["Frontiers in Artificial Intelligence"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2026.1875702\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,31]],"date-time":"2026-08-31T05:40:48Z","timestamp":1788154848000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2026.1875702\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,8,31]]},"references-count":36,"alternative-id":["10.3389\/frai.2026.1875702"],"URL":"https:\/\/doi.org\/10.3389\/frai.2026.1875702","relation":{},"ISSN":["2624-8212"],"issn-type":[{"value":"2624-8212","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,8,31]]},"article-number":"1875702"}}