Presented By: DCMB Seminar Series
Gilbert S. Omenn Department of Computational Medicine and Bioinformatics Weekly Seminar
Todd C. Hollon, MD (Assistant Professor in Neurological Surgery, University of Michigan), "Multimodal AI models for neuroimaging
Abstract:
Advancements in multimodal artificial intelligence (AI) are revolutionizing neuroimaging, offering solutions to longstanding clinical and operational challenges. This talk explores how foundation models trained on large-scale, multimodal datasets can optimize diagnostic accuracy, streamline workflows, and address healthcare disparities. By leveraging diverse neuroimaging data, these models integrate visual and linguistic modalities to extract generalizable features, enabling precise diagnostics across a wide spectrum of neurological conditions. Their capabilities extend to real-time applications, such as intraoperative decision-making, where rapid and interpretable predictions enhance surgical outcomes. These AI systems demonstrate superior performance compared to traditional methods, achieving high diagnostic accuracy while maintaining fairness across patient demographics and healthcare settings. Emphasizing scalability and generalizability, these approaches not only improve diagnostic workflows but also reduce bias and expand access to advanced neuroimaging in low-resource environments. Attendees will gain insights into the transformative potential of multimodal AI, from improving patient care to setting new benchmarks in the integration of artificial intelligence into healthcare systems.
Bio:
Assistant Professor of Neurosurgery, Computational Medicine and Bioinformatics, and Computer Science and Engineering. Principal Investigator of the Machine Learning in Neurosurgery Lab.
Advancements in multimodal artificial intelligence (AI) are revolutionizing neuroimaging, offering solutions to longstanding clinical and operational challenges. This talk explores how foundation models trained on large-scale, multimodal datasets can optimize diagnostic accuracy, streamline workflows, and address healthcare disparities. By leveraging diverse neuroimaging data, these models integrate visual and linguistic modalities to extract generalizable features, enabling precise diagnostics across a wide spectrum of neurological conditions. Their capabilities extend to real-time applications, such as intraoperative decision-making, where rapid and interpretable predictions enhance surgical outcomes. These AI systems demonstrate superior performance compared to traditional methods, achieving high diagnostic accuracy while maintaining fairness across patient demographics and healthcare settings. Emphasizing scalability and generalizability, these approaches not only improve diagnostic workflows but also reduce bias and expand access to advanced neuroimaging in low-resource environments. Attendees will gain insights into the transformative potential of multimodal AI, from improving patient care to setting new benchmarks in the integration of artificial intelligence into healthcare systems.
Bio:
Assistant Professor of Neurosurgery, Computational Medicine and Bioinformatics, and Computer Science and Engineering. Principal Investigator of the Machine Learning in Neurosurgery Lab.
Related Links
Explore Similar Events
-
Loading Similar Events...
