{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"institution":[{"name":"medRxiv"}],"indexed":{"date-parts":[[2026,6,9]],"date-time":"2026-06-09T17:04:44Z","timestamp":1781024684067,"version":"3.54.1"},"posted":{"date-parts":[[2024,4,15]]},"group-title":"Neurology","reference-count":27,"publisher":"openRxiv","license":[{"start":{"date-parts":[[2024,4,15]],"date-time":"2024-04-15T00:00:00Z","timestamp":1713139200000},"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":[[2025,2,3]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                <jats:sec>\n                  <jats:title>Background<\/jats:title>\n                  <jats:p>For patients with drug-resistant focal epilepsy (DRE), surgical resection of the epileptogenic zone (EZ) is an effective treatment to control seizures. Accurate localization of the EZ is crucial and is typically achieved through comprehensive presurgical approaches such as seizure semiology interpretation, electroencephalography (EEG), magnetic resonance imaging (MRI), and intracranial EEG (iEEG). However, interpreting seizure semiology poses challenges because it relies heavily on expert knowledge and is often based on inconsistent and incoherent descriptions, leading to variability and potential limitations in presurgical evaluation. To overcome these challenges, advanced technologies like large language models (LLMs)\u2014with ChatGPT being a notable example\u2014offer valuable tools for analyzing complex textual information, making them well-suited to interpret detailed seizure semiology descriptions and assist in accurately localizing the EZ.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Objective<\/jats:title>\n                  <jats:p>This study evaluates the clinical value of ChatGPT in interpreting seizure semiology to localize EZs in presurgical assessments for patients with focal epilepsy and compares its performance with epileptologists.<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Methods<\/jats:title>\n                  <jats:p>Two data cohorts were compiled: a publicly sourced cohort consisting of 852 semiology-EZ pairs from 193 peer-reviewed journal publications and a private cohort of 184 semiology-EZ pairs collected from Far Eastern Memorial Hospital (FEMH) in Taiwan. ChatGPT was evaluated to predict the most likely EZ locations using two prompt methods: zero-shot prompting (ZSP) and few-shot prompting (FSP). To compare ChatGPT\u2019s performance, eight epileptologists were recruited to participate in an online survey to interpret 100 randomly selected semiology records. The responses from ChatGPT and the epileptologists were compared using three metrics: regional sensitivity (RSens), weighted sensitivity (WSens), and net positive inference rate (NPIR).<\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Results<\/jats:title>\n                  <jats:p>\n                    In the publicly sourced cohort, ChatGPT demonstrated high RSens reliability, achieving 80-90% for the frontal and temporal lobes, 20-40% for the parietal lobe, occipital lobe, and insular cortex, and only 3% for the cingulate cortex. The WSens, which accounts for biased data distribution, consistently exceeded 67%, while the mean NPIR remained around 0. These evaluation results based on the private FEMH cohort are consistent with those from the publicly sourced cohort. A group\n                    <jats:italic>t<\/jats:italic>\n                    -test with 1000 bootstrap samples revealed that ChatGPT-4 significantly outperformed epileptologists in RSens for commonly represented EZs, such as the frontal and temporal lobes (p &lt; 0.001). Additionally, ChatGPT-4 demonstrated superior overall performance in WSens (p &lt; 0.001). However, no significant differences were observed between ChatGPT and the epileptologists in NPIR, highlighting comparable performance in this metric.\n                  <\/jats:p>\n                <\/jats:sec>\n                <jats:sec>\n                  <jats:title>Conclusions<\/jats:title>\n                  <jats:p>ChatGPT demonstrated clinical value as a tool to assist the decision-making in the epilepsy preoperative workup. With ongoing advancements in LLMs, it is anticipated that the reliability and accuracy of LLMs will continue to improve in the future.<\/jats:p>\n                <\/jats:sec>","DOI":"10.1101\/2024.04.13.24305773","type":"posted-content","created":{"date-parts":[[2024,4,15]],"date-time":"2024-04-15T18:25:22Z","timestamp":1713205522000},"source":"Crossref","is-referenced-by-count":5,"title":["The Clinical Value of ChatGPT for Epilepsy Presurgical Decision Making: Systematic Evaluation on Seizure Semiology Interpretation"],"prefix":"10.64898","author":[{"given":"Yaxi","family":"Luo","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Meng","family":"Jiao","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Neel","family":"Fotedar","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jun-En","family":"Ding","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ioannis","family":"Karakis","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Vikram R.","family":"Rao","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Melissa","family":"Asmar","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xiaochen","family":"Xian","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Orwa","family":"Aboud","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yuxin","family":"Wen","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jack J.","family":"Lin","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Fang-Ming","family":"Hung","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hai","family":"Sun","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Felix","family":"Rosenow","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5225-8199","authenticated-orcid":false,"given":"Feng","family":"Liu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"54368","reference":[{"key":"2025020511551005000_2024.04.13.24305773v4.1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ncl.2016.06.015"},{"key":"2025020511551005000_2024.04.13.24305773v4.2","doi-asserted-by":"publisher","DOI":"10.1001\/jama.2016.18625"},{"key":"2025020511551005000_2024.04.13.24305773v4.3","doi-asserted-by":"publisher","DOI":"10.1001\/jamaneurol.2017.3949"},{"key":"2025020511551005000_2024.04.13.24305773v4.4","doi-asserted-by":"crossref","unstructured":"Arno Klein and Jason Tourville . 101 labeled brain images and a consistent human cortical labeling protocol. Frontiers in neuroscience, 6:33392, 2012.","DOI":"10.3389\/fnins.2012.00171"},{"issue":"1","key":"2025020511551005000_2024.04.13.24305773v4.5","doi-asserted-by":"crossref","first-page":"22","DOI":"10.1017\/S0317167100007526","article-title":"Seizure semiology: value in identifying seizure origin","volume":"35","year":"2008","journal-title":"Canadian Journal of Neurological Sciences"},{"issue":"4","key":"2025020511551005000_2024.04.13.24305773v4.6","doi-asserted-by":"crossref","first-page":"243","DOI":"10.3988\/jcn.2012.8.4.243","article-title":"Seizure semiology: its value and limitations in localizing the epileptogenic zone","volume":"8","year":"2012","journal-title":"Journal of Clinical Neurology"},{"issue":"1","key":"2025020511551005000_2024.04.13.24305773v4.7","doi-asserted-by":"crossref","first-page":"12","DOI":"10.5698\/1535-7597.18.1.12","article-title":"The epileptogenic zone: concept and definition","volume":"18","year":"2018","journal-title":"Epilepsy currents"},{"key":"2025020511551005000_2024.04.13.24305773v4.8","doi-asserted-by":"publisher","DOI":"10.3109\/9780203091708-107"},{"key":"2025020511551005000_2024.04.13.24305773v4.9","doi-asserted-by":"publisher","DOI":"10.1046\/j.1528-1157.2001.35100.x"},{"key":"2025020511551005000_2024.04.13.24305773v4.10","first-page":"514","article-title":"Surgical treatment of the epilepsies: presurgical evaluation","volume":"38","year":"1992","journal-title":"Clin Neurosurg"},{"key":"2025020511551005000_2024.04.13.24305773v4.11","doi-asserted-by":"publisher","DOI":"10.1016\/j.eplepsyres.2015.05.012"},{"issue":"5","key":"2025020511551005000_2024.04.13.24305773v4.12","doi-asserted-by":"crossref","first-page":"1122","DOI":"10.1109\/JAS.2023.123618","article-title":"A brief overview of ChatGPT: The history, status quo and potential future development","volume":"10","year":"2023","journal-title":"IEEE\/CAA Journal of Automatica Sinica"},{"key":"2025020511551005000_2024.04.13.24305773v4.13","doi-asserted-by":"publisher","DOI":"10.1038\/s41591-023-02448-8"},{"issue":"2","key":"2025020511551005000_2024.04.13.24305773v4.14","first-page":"AIoa2300068","article-title":"Almanac\u2014retrieval-augmented language models for clinical medicine","volume":"1","year":"2024","journal-title":"NEJM AI"},{"issue":"11","key":"2025020511551005000_2024.04.13.24305773v4.15","doi-asserted-by":"crossref","first-page":"e2343689","DOI":"10.1001\/jamanetworkopen.2023.43689","article-title":"Leveraging large language models for decision support in personalized oncology","volume":"6","year":"2023","journal-title":"JAMA Network Open"},{"issue":"2","key":"2025020511551005000_2024.04.13.24305773v4.16","doi-asserted-by":"crossref","first-page":"e0000198","DOI":"10.1371\/journal.pdig.0000198","article-title":"Performance of ChatGPT on USMLE: Potential for AI-assisted medical education using large language models","volume":"2","year":"2023","journal-title":"PLOS Digital Health"},{"key":"2025020511551005000_2024.04.13.24305773v4.17","doi-asserted-by":"publisher","DOI":"10.1148\/radiol.231040"},{"key":"2025020511551005000_2024.04.13.24305773v4.18","doi-asserted-by":"crossref","unstructured":"Lars Mehnen , Stefanie Gruarin , Mina Vasileva , and Bernhard Knapp . ChatGPT as a medical doctor? A diagnostic accuracy study on common and rare diseases. medRxiv, pages 2023\u201304, 2023.","DOI":"10.1101\/2023.04.20.23288859"},{"key":"2025020511551005000_2024.04.13.24305773v4.19","unstructured":"Kathi Canese and Sarah Weis . PubMed: The bibliographic database. The NCBI handbook, 2(1), 2013."},{"issue":"3","key":"2025020511551005000_2024.04.13.24305773v4.20","doi-asserted-by":"crossref","first-page":"fcac130","DOI":"10.1093\/braincomms\/fcac130","article-title":"Probabilistic landscape of seizure semiology localizing values","volume":"4","year":"2022","journal-title":"Brain Communications"},{"key":"2025020511551005000_2024.04.13.24305773v4.21","first-page":"22199","article-title":"Large language models are zero-shot reasoners","volume":"35","year":"2022","journal-title":"Advances in neural information processing systems"},{"key":"2025020511551005000_2024.04.13.24305773v4.22","doi-asserted-by":"crossref","first-page":"3081","DOI":"10.1609\/aaai.v36i3.20215","article-title":"An empirical study of GPT-3 for few-shot knowledge-based vqa","volume":"36","year":"2022","journal-title":"Proceedings of the AAAI Conference on Artificial Intelligence"},{"key":"2025020511551005000_2024.04.13.24305773v4.23","doi-asserted-by":"publisher","DOI":"10.1001\/jamaneurol.2024.0676"},{"key":"2025020511551005000_2024.04.13.24305773v4.24","unstructured":"https:\/\/survey.zohopublic.com\/zs\/NECl0I"},{"key":"2025020511551005000_2024.04.13.24305773v4.25","doi-asserted-by":"publisher","DOI":"10.5121\/ijdkp.2015.5201"},{"key":"2025020511551005000_2024.04.13.24305773v4.26","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1016\/j.seizure.2023.11.013","article-title":"Assessing the performance of ChatGPT\u2019s responses to questions related to epilepsy: A cross-sectional study on natural language processing and medical information retrieval","volume":"114","year":"2024","journal-title":"Seizure: European Journal of Epilepsy"},{"key":"2025020511551005000_2024.04.13.24305773v4.27","doi-asserted-by":"crossref","unstructured":"YuXin Wu , Zaiyu Zhang , Xinyu Dong , Siqi Hong , Yue Hu , Ping Liang , Lusheng Li , Bin Zou , Xuanxuan Wu , Difei Wang , et al. Evaluating the performance of the language model ChatGPT in responding to common questions of people with epilepsy. Epilepsy & Behavior, 151:109645, 2024.","DOI":"10.1016\/j.yebeh.2024.109645"}],"container-title":[],"original-title":[],"link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.1101\/2024.04.13.24305773","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,1,15]],"date-time":"2026-01-15T19:58:19Z","timestamp":1768507099000},"score":1,"resource":{"primary":{"URL":"http:\/\/medrxiv.org\/lookup\/doi\/10.1101\/2024.04.13.24305773"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,4,15]]},"references-count":27,"URL":"https:\/\/doi.org\/10.1101\/2024.04.13.24305773","relation":{"is-preprint-of":[{"id-type":"doi","id":"10.2196\/69173","asserted-by":"subject"}]},"subject":[],"published":{"date-parts":[[2024,4,15]]},"subtype":"preprint"}}