Perturbation-induced responses improved seizure forecasting in epileptic rats

Area of Science:

• Neuroscience
• Epilepsy Research
• Biomarker Discovery

Background:

• Epilepsy seizure unpredictability poses significant challenges for patient management.
• Current seizure risk forecasting relies on passive intracranial electroencephalography (EEG) signals, with no available clinical devices.
• Dynamical theory suggests brain response to stimulation may offer a more robust seizure risk measure.

Purpose of the Study:

• To test the hypothesis that brain response to electrical stimulation is a superior seizure risk biomarker.
• To compare the efficacy of perturbation-evoked responses versus passive EEG features for seizure prediction.

Main Methods:

• Periodic electrical stimulation and continuous EEG recordings were conducted in rats with induced epilepsy over 8 weeks.
• A preictal detection system was developed using perturbation-evoked responses as a predictive biomarker.
• A comparative system utilizing only passive EEG features was also developed.

Main Results:

• The perturbation-evoked response system achieved excellent accuracy (AUC > 0.95) in distinguishing preictal from interictal states.
• In contrast, the system using only passive EEG features could not predict the preictal state better than chance.
• This highlights the superior predictive power of active brain stimulation responses.

Conclusions:

• Perturbation-evoked responses are a promising biomarker for seizure prediction.
• This approach could significantly enhance the clinical efficacy of seizure forecasting.
• The findings advocate for the clinical application of perturbation-based seizure prediction.