{"id":26299,"date":"2026-01-21T10:17:05","date_gmt":"2026-01-21T09:17:05","guid":{"rendered":"https:\/\/sano.science\/?post_type=seminars&p=26299"},"modified":"2026-01-21T10:17:06","modified_gmt":"2026-01-21T09:17:06","slug":"toward-trustworthy-ai-for-acute-ischemic-stroke-triage-with-eeg-from-signal-processing-to-uncertainty-aware-decision-support","status":"publish","type":"seminars","link":"https:\/\/sano.science\/seminars\/toward-trustworthy-ai-for-acute-ischemic-stroke-triage-with-eeg-from-signal-processing-to-uncertainty-aware-decision-support\/","title":{"rendered":"174. Toward Trustworthy AI for Acute Ischemic Stroke Triage with EEG: From Signal Processing to Uncertainty-Aware Decision Support"},"content":{"rendered":"\n

Abstract: <\/strong><\/h2>\n\n\n\n

Electroencephalography (EEG) holds promise for rapid pre-hospital stroke triage but is hindered by signal noise [1]<\/sup>. In this seminar, I will present the evolution of a comprehensive framework for acute ischemic stroke (AIS) triage, developed during my Visiting Scholar internship and ongoing collaboration with Amsterdam UMC, within the context of the ELECTRA-Stroke and AI-Stroke trials.<\/p>\n\n\n\n

I will first discuss the foundational role of signal quality, demonstrating how a dedicatedartefact removal pipeline significantly enhances EEG reliability in real-world clinical data[<\/sup>2<\/sup>]<\/sup>. Bridging signal processing with clinical decision support, I will then introduce anUncertainty-Aware framework that incorporates a \u201cRejection Option\u201d as a safety mechanism. This approach enables automated decision-making in approximately 70% of cases with high reliability, while flagging uncertain predictions for expert review[<\/sup>3,4<\/sup>]<\/sup>.<\/p>\n\n\n\n

Finally, I will outline future research directions toward a Hierarchical Time-Aware Cascade Model[<\/sup>4<\/sup>]<\/sup>, aiming to transform low-density EEG into a granular triage tool capable of identifying Large Vessel Occlusions, thereby supporting time-critical patient stratification for endovascular treatment.<\/p>\n\n\n\n

<\/div>\n\n\n\n

References:<\/strong><\/p>\n\n\n\n

1<\/strong>. Dhillon, P. S., Singh, N., Ospel, J. M., Roozenbeek, B., Goyal, M., & Hill, M. D. (2025). Pre-Hospital Stroke Triage and Research: Challenges and Opportunities. Cerebrovascular Diseases<\/em>, 54<\/em>(2), 282-287.<\/p>\n\n\n\n

2<\/strong>. Arpaia, P., De Luca, M., Di Marino, L., Duran, D., Gargiulo, L., Lanteri, P., … & Visani, E. (2025). A systematic review of techniques for artifact detection and artifact category identification in electroencephalography from wearable devices. Sensors<\/em>, 25<\/em>(18), 5770.<\/p>\n\n\n\n

3<\/strong>. Campagner, A., Biganzoli, E. M., Balsano, C., Cereda, C., & Cabitza, F. (2025). Modeling Unknowns: A Vision for Uncertainty-Aware Machine Learning in Healthcare. International Journal of Medical Informatics<\/em>, 106014.<\/p>\n\n\n\n

4<\/strong>. Islam, M. R. (2026). An Explainable Agentic AI Framework for Uncertainty-Aware and Abstention-Enabled Acute Ischemic Stroke Imaging Decisions. arXiv preprint arXiv:2601.01008<\/em>.<\/p>\n\n\n\n

<\/div>\n\n\n\n

About the author:<\/strong><\/h2>\n\n\n\n

Rosmary Blanco<\/a> holds a BSc in Neurophysiopathology from the University of Udine (Italy) and an MSc in Molecular and Medical Biotechnology from the University of Verona (Italy). She brings clinical experience from hospital neurology, neurosurgery, and neurorehabilitation departments.<\/p>\n\n\n\n

Currently, Rosmary is a PhD student working on Computational Neuroscience and AI, with a focus on acute stroke and early dementia assessment. Her research combines electrophysiological signal processing, network neuroscience, and emerging technologies, with a particular focus on EEG, to improve decision-making and patient risk stratification. She is dedicated to translating her research into clinical practice by developing Clinical Decision Support Systems (CDSS).<\/p>\n\n\n\n

<\/div>\n\n\n\n
\n