To quantify seizure spread using EEG features and deep learning models, and to investigate the association of seizure spread with patient outcomes and brain connectivity.
Approach:
Patient Inclusion: Seventy-one patients with treatment-resistant epilepsy underwent intracranial EEG (iEEG) monitoring and provided consent for data analysis.
EEG Data Acquisition: Continuous iEEG signals were recorded during the patients' stay in the epilepsy monitoring unit, with seizure onset times defined by unequivocal electrographic onset.
Deep Learning Model Development: Deep learning algorithms were trained to measure seizure spread, utilizing preprocessed EEG data and single features like absolute slope and broadband power.
Key Findings:
Seizure spread patterns may correlate with patient outcomes and structural brain connectivity.
Automated measures for quantifying seizure activity spread are lacking, with poor reliability in human evaluations.
Deep learning models can effectively quantify seizure spread based on EEG data.
Interpretation:
The study highlights the potential of AI in improving the understanding of seizure dynamics and outcomes in refractory epilepsy.
Limitations:
The study's findings are based on a limited sample size of 71 patients.
Reliability of human evaluations in intracranial EEG remains a challenge.
Challenges related to the performance and validation of deep learning models were noted.
Conclusion:
Further research is needed to validate the findings and improve automated seizure spread quantification methods.
by Andrew Y. Revell, Marc Jaskir, Akash R. Pattnaik, William K. S. Ojemann, Erin Conrad, Nishant Sinha, Brittany H. Scheid, Alfredo Lucas, John M. Bernabei, John Beckerle, Joel M. Stein, Sandhitsu R. Das, Brian Litt, Kathryn A. Davis
Federal prosecutors allege that a Florida physician and research staff fabricated clinical trial records that were submitted into database systems used to evaluate investigational drugs.
