A Nonlinear Method to Identify Seizure Dynamic Trajectory Based on Variance of Recurrence Rate in Human Epilepsy...

Area of Science:

• Neuroscience
• Medical Technology
• Epilepsy Research

Background:

• Surgery is a key treatment for drug-resistant epilepsy, but success is limited, especially without visible lesions.
• Interpreting electroencephalogram (EEG) data to pinpoint seizure origin and spread is a significant challenge.
• Accurately tracing seizure signal trajectories in the complex brain remains a critical hurdle.

Purpose of the Study:

• To develop and validate a novel method for precisely identifying seizure-involved brain regions using EEG data.
• To enhance the accuracy of localizing the epileptogenic zone for improved surgical planning in epilepsy.
• To provide a tool for real-time clinical monitoring during epilepsy treatment.

Main Methods:

• Analyzed EEG data from 17 epilepsy patients, using clinical interpretations as the gold standard.
• Employed quantification analysis of recurrence plots, focusing on recurrence rate variance, to identify seizure-involved regions.
• Conducted stage-wise analysis across EEG electrodes to highlight simultaneously involved brain areas.

Main Results:

• The recurrence plot method achieved a macro-averaged F-score of 95.54 in distinguishing involved from non-involved brain regions.
• Demonstrated high accuracy (up to 86.96%), sensitivity (up to 82.79%), and specificity (up to 86.96%) in identifying seizure-involved areas.
• Showcased robust performance across anterior, posterior, and temporal regions, with other regions yielding 66.0% to 89.13% accuracy.

Conclusions:

• The developed approach effectively pinpoints brain regions involved in seizures at any stage.
• This method shows promise for clinical monitoring and surgical planning in epilepsy management.
• The technique's simplicity and strong performance suggest potential for real-time application in epilepsy treatment.