Seizure prediction and forecasting: a scoping review.

Area of Science:

• Epilepsy research
• Neurological disorder forecasting
• Biomedical signal processing

Background:

• Seizure forecasting aims to predict epileptic seizures.
• Advances in monitoring technologies like EEG and wearables have been explored.
• Clinical application of these forecasting tools remains limited.

Purpose of the Study:

• To review key developments in seizure forecasting.
• To identify limitations hindering clinical translation of seizure forecasting tools.

Main Methods:

• Scoping review of seizure forecasting literature.
• Analysis of various forecasting modalities (EEG, wearables, diaries).
• Evaluation of current validation and performance comparison metrics.

Main Results:

• Progress in seizure forecasting modalities including intracranial/subcutaneous EEG, wearables, and diaries.
• Clinical translation is hampered by lack of external validation and inadequate comparison benchmarks (e.g., chance forecaster).
• Patients desire even imprecise forecasts and may alter behavior, despite potential inaccuracies.

Conclusions:

• Seizure forecasting tools show potential but require further development for clinical use.
• Future research must focus on external validation, the 'napkin test' benchmark, and seizure frequency-dependent metrics.
• Addressing patient behavior in response to forecasts is crucial.