Systematic comparison of Commercial seizure detection Software: Update equals Upgrade?

Area of Science:

• Neurology
• Medical Technology
• Epilepsy Research

Background:

• Commercial seizure detection software is crucial for epilepsy management.
• Previous versions of encevis and Persyst have been evaluated.
• Software updates necessitate re-evaluation of performance.

Purpose of the Study:

• To compare the performance of updated encevis 2.0 and Persyst 14 seizure detection software.
• To assess detection rates and false positive rates in a large video-EEG dataset.

Main Methods:

• A follow-up study evaluated two updated seizure detection software packages (encevis 2.0, Persyst 14).
• Data comprised 80 patients, 6826 hours of EEG, and 912 seizures.
• Different sensitivity settings were utilized for analysis.

Main Results:

• Encevis 2.0 achieved higher detection rates (80.7%-89.1%) compared to Persyst 14 (67.3%-71.1%).
• Persyst 14 exhibited significantly lower false positive rates (2-4 per day) than encevis 2.0 (12-22 per day).
• Mean seizure count was 11.4 per patient, with a 0.7 tonic-clonic seizure rate.

Conclusions:

• Encevis 2.0 offers superior sensitivity, while Persyst 14 provides better specificity.
• Updated software showed divergent performance changes: encevis improved sensitivity, Persyst improved false positive rates.
• Automatic seizure detection software is clinically viable but requires expert interpretation and awareness of software limitations.