Research › QEEG, Methods & Mechanisms
Using Machine Learning to Enhance the EEG Screening Review by Pre-Screening the EEG
STS Instructor Authored by Bill Brubaker, P. David Ims, Robert Turner and Thomas Collura
Share
Study at a glance
- Area
- QEEG, Methods & Mechanisms
- Study type
- Observational study
- Authors
- Collura T, Rosace A, Turner R, Ims D, Brubaker B
- Journal
- Artificial Intelligence and Applications
- Year
- 2026
- Participants
- 100 EEGs
- Protocol
- machine-learning EEG pre-screening (Brain Panel)
Key findings
- Automated pre-screen estimates artifact, overall quality, drowsiness and likelihood of abnormality
- 100 Brain Panel reports were checked against 100 matching physician reports
- Best metric combinations reached 89-95% sensitivity for clinically relevant findings
- Positioned as a support for, not a substitute for, expert EEG reading
The authors built an automated machine-learning tool, the Brain Panel, that screens an EEG before anyone reads it, flagging recording quality problems, artifacts, drowsiness and possible abnormalities. They compared 100 Brain Panel reports with the physician reports for the same EEGs, using two general-purpose AI systems to evaluate them independently. The tool proved sensitive to technical and clinical issues, and combinations of its measures detected clinically relevant findings with 89-95% sensitivity; it is meant to speed up neurologist review, not replace it.