Using sampled visual EEG review in combination with automated detection software at the EMU

Area of Science:

• Clinical Neurophysiology
• Epileptology
• Medical Technology

Background:

• Complete visual review of prolonged video-electroencephalography (EEG) recordings in Epilepsy Monitoring Units (EMUs) is time-intensive.
• A shortage of trained personnel can further complicate the review process.
• Developing efficient diagnostic methods is crucial for timely epilepsy management.

Purpose of the Study:

• To evaluate the non-inferiority of a sampled review combined with EEG analysis software (Persyst 13) compared to complete visual review for electroclinical diagnosis.
• To assess the diagnostic accuracy of a novel, time-efficient approach for prolonged video-EEG interpretation.

Main Methods:

• Fifty adult prolonged video-EEG recordings were prospectively analyzed.
• A sampled review included specific wake and sleep periods, plus all marked clinical events.
• This sampled approach was combined with automated detection software and compared to a complete visual review.

Main Results:

• Electroclinical diagnoses derived from the sampled review with automated detection were non-inferior to those from complete visual review.
• The automated detection software successfully identified all records containing epileptiform abnormalities and epileptic seizures.
• This indicates the sampled approach maintains diagnostic integrity.

Conclusions:

• Sampled visual review combined with Persyst 13 automated detection is a non-inferior alternative to complete visual review for electroclinical diagnosis in EMUs.
• This method offers a promising, more efficient approach for interpreting prolonged video-EEG recordings.
• The findings support the adoption of this technique to address personnel and time constraints in EMU settings.