سامانه همایش های دانشگاه آزاد اسلامی واحد نجف آباد

صفحه اصلی / اولین همایش بین المللی تحول خدمات سلامت: تلفیق مراقبت، درمان، بازتوانی و امیدبخشی

Artificial Intelligence for Etiological and Biomarker Discovery in Epilepsy of Unknown Etiology

نویسندگان :

Sarina Rezaei¹ Samira Shariati²

1- دانشگاه آزاد نجف آباد - دانشکده پزشکی 2- Clinical Research Development Center, Najafabad Branch, Islamic Azad University, Najafabad, Ira

کلمات کلیدی :

Epilepsy of unknown etiology،Artificial intelligence،Machine learning،Genetic epilepsy،Biomarkers،Multi-omics،Precision medicine

چکیده :

Abstract: Background: Epilepsy is one of the most frequent neurological disorders; however, the cause of epilepsy still remains undetermined in a large number of patients. Genetic variations, autoimmunity, neuroinflammation, metabolic dysfunction, structural anomalies, and molecular changes can be responsible for epilepsy with unknown causes but may go unnoticed using traditional methods of diagnosis. The application of artificial intelligence (AI) opens new possibilities for recognizing complicated patterns and analyzing various aspects for identifying the possible causes of the disorder. Methods: The present study is a narrative review done by searching the PubMed/MEDLINE, Scopus, Web of Science, Embase, and Google Scholar databases. Literature from 2007 to 2026 was selected without language limitations. Search keywords include idiopathic epilepsy, artificial intelligence, machine learning, genetic epilepsy, multi-omics, and precision medicine. Finding: Artificial intelligence and machine learning tools can aid in the detection and interpretation of genetic variants, polygenic risk, clinical phenotypes, EEG phenotypes, and imaging biomarkers related to epilepsy. Genomics, transcriptomics, proteomics, and metabolomics data integration can be useful for detecting molecular biomarkers and pathways. Further, blood and cerebrospinal fluid biomarkers, extracellular vesicles, neural autoantibodies, MRI, EEG, and clinical phenotypes can be combined using multimodal analysis for biologically distinct subgroups and treatment etiology. However, some limitations such as small dataset sizes, non-diverse datasets, overfitting, lack of external validation, and sampling biases still persist. Conclusion: Artificial intelligence-driven multimodal and multi-omics approaches have significant promise for exploring idiopathic forms of epilepsy through the analysis of different clinical, genetic, imaging, electrophysiological, and molecular variables. This may help uncover novel etiologies and biomarkers for epilepsy, and promote stratification of patients to tailor their diagnosis and treatment. Nevertheless, further validation in larger cohorts is necessary prior to their clinical application. Keywords: Epilepsy of unknown etiology; Artificial intelligence; Machine learning; Genetic epilepsy; Multi-omics; Biomarkers; Precision medicine