About this Research Topic
Many people with epilepsy continue to experience seizures, despite neuromodulation and medication therapy advancements. Moreover, not all patients who undergo resective surgery achieve long-term seizure freedom, and many have eventual recurrence of seizures. According to people with epilepsy, the most limiting aspect of living with epilepsy is the unpredictability of seizure occurrence. Reliable seizure forecasting has become possible thanks to advances in vast data collections of seizure rates via seizure diaries, wearable devices, and computational algorithms that identify cyclical seizure patterns. Seizure forecasting holds promise to help people living with epilepsy managing their seizures by taking a fast-acting medication, increasing neuromodulation therapy, or modifying their activities.
Recent developments allow ultralong-term data collection and developing forecasting algorithms using data from invasive and non-invasive devices. Although invasive devices may not be acceptable for some patients, and few available devices currently have the capability to sample and telemeter data needed for seizure forecasting, it raises this question if the classifiers trained on physiological signals (Photoplethysmography (PPG), Electrodermal Activity (EDA), Accelerometry (ACC), scalp Electroencephalography (EEG)) recorded with non-invasive devices can outperform iEEG classifiers. Other questions remain, including what physiological signals/features in time-series data are most important in seizure forecasting?; Are physiological data adequate, or are additional data, such as mood, sleep pattern or even demographic information needed?; and finally, is the acceptability of the forecasting algorithms’ accuracy, false alarm rate and sensitivity well defined?
We are inviting the submission of Original Research, Systematic-Review, Review, Methods, Hypothesis and Theory, and Perspective articles on the following sub-topics:
-Applications of invasive and non-invasive devices in ultralong term monitoring & automated analysis. How to manage and share mobile health data?
- Importance of the physiological signals (EEG, ECG, PPG, EDA, ...) and signal quality assessment.
- Exogenous vs Endogenous parameters. Do mood, stress, sleep quality and also demographic characteristics affect seizure risk?
-Possibility of developing generalized seizure forecasting algorithm vs subject specific algorithms.
- Considering patients and caregivers’ perspectives and preferences, how to evaluate performance of a forecasting algorithm?
-Closed-loop systems for responsive brain stimulations and neuromodulation.
Recent developments allow ultralong-term data collection and developing forecasting algorithms using data from invasive and non-invasive devices. Although invasive devices may not be acceptable for some patients, and few available devices currently have the capability to sample and telemeter data needed for seizure forecasting, it raises this question if the classifiers trained on physiological signals (Photoplethysmography (PPG), Electrodermal Activity (EDA), Accelerometry (ACC), scalp Electroencephalography (EEG)) recorded with non-invasive devices can outperform iEEG classifiers. Other questions remain, including what physiological signals/features in time-series data are most important in seizure forecasting?; Are physiological data adequate, or are additional data, such as mood, sleep pattern or even demographic information needed?; and finally, is the acceptability of the forecasting algorithms’ accuracy, false alarm rate and sensitivity well defined?
We are inviting the submission of Original Research, Systematic-Review, Review, Methods, Hypothesis and Theory, and Perspective articles on the following sub-topics:
-Applications of invasive and non-invasive devices in ultralong term monitoring & automated analysis. How to manage and share mobile health data?
- Importance of the physiological signals (EEG, ECG, PPG, EDA, ...) and signal quality assessment.
- Exogenous vs Endogenous parameters. Do mood, stress, sleep quality and also demographic characteristics affect seizure risk?
-Possibility of developing generalized seizure forecasting algorithm vs subject specific algorithms.
- Considering patients and caregivers’ perspectives and preferences, how to evaluate performance of a forecasting algorithm?
-Closed-loop systems for responsive brain stimulations and neuromodulation.
Keywords: Seizure forecasting, neurostimulation, signal processing, machine learning, long-term monitoring, evaluation metrics, seizure risk estimation
Important Note: All contributions to this Research Topic must be within the scope of the section and journal to which they are submitted, as defined in their mission statements. Frontiers reserves the right to guide an out-of-scope manuscript to a more suitable section or journal at any stage of peer review.
