
Frontlines with Frontera: Using Artificial Intelligence to Rethink Seizure Management in the Neuro ICU
Frontlines with Frontera, a podcast hosted by neurointensivist Jennifer Frontera, MD, brings you an exclusive interview with Sahar Zafar, MD, MBBS. [LISTEN TIME: 30 minutes]
Episode 1 of the Frontlines with Frontera®
Frontlines with Frontera is a NeurologyLive podcast hosted by Jennifer Frontera, MD, Professor of Neurology in the Division of Neurocritical Care and Stroke at NYU Langone. Drawing on her extensive experience in clinical care, research, and guideline development, Dr. Frontera leads in-depth conversations focused on complex, high-stakes decision-making in neurocritical care and acute neurology. Each episode features Dr. Frontera alongside expert guests, offering clinicians practical, evidence-informed perspectives on emerging research, evolving standards of care, and unresolved questions at the front lines of neurologic practice.
In this initial episode, host Jennifer Frontera, MD, is joined by Sahar Zafar, MD, associate professor of neurology at Mass General Brigham and vice chair of quality, to discuss the expanding role of artificial intelligence (AI) in neurocritical care. The conversation explores how large-scale EEG data, electronic health records, and advanced causal modeling can be leveraged to address longstanding gaps in seizure management, prognostication, and trial design. In addition, Zafar outlines how AI-driven approaches are being used to annotate neurophysiologic data at scale, quantify seizure burden, and extract clinically meaningful outcomes from unstructured records. Together, they examine the challenges of model validation, generalizability across health systems, and how these tools may ultimately inform more precise, data-driven treatment strategies in the neuro ICU.
EPISODE BREAKDOWN
- 1:30 – Why evidence gaps and practice variation persist in neurocritical seizure management
- 4:50 – How AI and real-world data can address limitations of traditional clinical trials
- 8:10 – Using EEG-based AI models to quantify seizure burden and predict outcomes
- 10:40 – Applying NLP and large language models to extract meaningful data from EHRs
- 21:20 – Challenges of model validation, portability, and multicenter collaboration
- 26:20 – Future applications of AI for seizure prophylaxis, multimodal monitoring, and trial design
Hoping to listen on your favorite podcast app? See below:
Newsletter
Keep your finger on the pulse of neurology—subscribe to NeurologyLive for expert interviews, new data, and breakthrough treatment updates.



















