Pitfalls in EEG Analysis in Patients With Nonconvulsive Status Epilepticus: A Preliminary Study

Area of Science:

• Neurology
• Medical Technology
• Epilepsy Research

Background:

• Electroencephalography (EEG) interpretation for nonconvulsive status epilepticus (NCSE) lacks reliability.
• Both visual (human) and automated (computer) analyses present challenges in accurate NCSE diagnosis.
• Identifying and rectifying common pitfalls in EEG analysis is crucial for improving diagnostic accuracy.

Purpose of the Study:

• To identify specific pitfalls in EEG interpretation for nonconvulsive status epilepticus (NCSE).
• To investigate factors contributing to unreliable visual EEG interpretations in suspected NCSE cases.
• To propose strategies for avoiding misinterpretations in automated and visual EEG analysis.

Main Methods:

• Analysis of EEG recordings from patients with confirmed or suspected NCSE.
• Visual interpretation of ictal discharges by two independent raters to assess interrater agreement.
• Automated analysis using in-house algorithms and expert review of misinterpreted EEG patterns.

Main Results:

• Short ictal discharges (gradual onset, <3 seconds) and abnormal background activity were frequently misinterpreted.
• Misinterpretation of continuous interictal discharges and movement artifacts also contributed to diagnostic errors.
• Increased EEG duration (e.g., 2 minutes) improved interrater agreement (kappa > 0.1).

Conclusions:

• A longer duration criterion (beyond 10 seconds) for NCSE EEG analysis is recommended.
• Utilizing historical EEG data and individualized algorithms can enhance diagnostic accuracy.
• Context-dependent alarm thresholds in automated systems may help mitigate interpretation pitfalls.