Log in
Enquire now
‌

Towards trustworthy seizure onset detection using workflow notes

OverviewStructured DataIssuesContributors

Contents

Paper abstractTimelineTable: Further ResourcesReferences
Is a
‌
Academic paper
1

Academic Paper attributes

arXiv ID
2306.087281
arXiv Classification
Computer science
Computer science
1
Publication URL
arxiv.org/pdf/2306.0...28.pdf1
Publisher
ArXiv
ArXiv
1
DOI
doi.org/10.48550/ar...06.087281
Paid/Free
Free1
Academic Discipline
Artificial Intelligence (AI)
Artificial Intelligence (AI)
1
Computer science
Computer science
1
Machine learning
Machine learning
1
Electrical engineering
Electrical engineering
1
Signal processing
Signal processing
1
Submission Date
June 14, 2023
2
Author Names
Daniel Rubin1
Siyi Tang1
Mohamed Taha1
Christopher Ré1
Khaled Saab1
Christopher Lee-Messer1
Paper abstract

A major barrier to deploying healthcare AI models is their trustworthiness. One form of trustworthiness is a model's robustness across different subgroups: while existing models may exhibit expert-level performance on aggregate metrics, they often rely on non-causal features, leading to errors in hidden subgroups. To take a step closer towards trustworthy seizure onset detection from EEG, we propose to leverage annotations that are produced by healthcare personnel in routine clinical workflows -- which we refer to as workflow notes -- that include multiple event descriptions beyond seizures. Using workflow notes, we first show that by scaling training data to an unprecedented level of 68,920 EEG hours, seizure onset detection performance significantly improves (+12.3 AUROC points) compared to relying on smaller training sets with expensive manual gold-standard labels. Second, we reveal that our binary seizure onset detection model underperforms on clinically relevant subgroups (e.g., up to a margin of 6.5 AUROC points between pediatrics and adults), while having significantly higher false positives on EEG clips showing non-epileptiform abnormalities compared to any EEG clip (+19 FPR points). To improve model robustness to hidden subgroups, we train a multilabel model that classifies 26 attributes other than seizures, such as spikes, slowing, and movement artifacts. We find that our multilabel model significantly improves overall seizure onset detection performance (+5.9 AUROC points) while greatly improving performance among subgroups (up to +8.3 AUROC points), and decreases false positives on non-epileptiform abnormalities by 8 FPR points. Finally, we propose a clinical utility metric based on false positives per 24 EEG hours and find that our multilabel model improves this clinical utility metric by a factor of 2x across different clinical settings.

Timeline

No Timeline data yet.

Further Resources

Title
Author
Link
Type
Date
No Further Resources data yet.

References

Find more entities like Towards trustworthy seizure onset detection using workflow notes

Use the Golden Query Tool to find similar entities by any field in the Knowledge Graph, including industry, location, and more.
Open Query Tool
Access by API
Golden Query Tool
Golden logo

Company

  • Home
  • Press & Media
  • Blog
  • Careers
  • WE'RE HIRING

Products

  • Knowledge Graph
  • Query Tool
  • Data Requests
  • Knowledge Storage
  • API
  • Pricing
  • Enterprise
  • ChatGPT Plugin

Legal

  • Terms of Service
  • Enterprise Terms of Service
  • Privacy Policy

Help

  • Help center
  • API Documentation
  • Contact Us
By using this site, you agree to our Terms of Service.