Ensemble Learning Using Individual Neonatal Data for Seizure Detection.

OBJECTIVE: Sharing medical data between institutions is difficult in practice due to data protection laws and official procedures within institutions. Therefore, most existing algorithms are trained on relatively small electroencephalogram (EEG) data sets which is likely to be detrimental to prediction accuracy. In this work, we simulate a case when the data can not be shared by splitting the publicly available data set into disjoint sets representing data in individual institutions. METHODS AND PROCEDURES: We propose to train a (local) detector in each institution and aggregate their individual predictions into one final prediction. Four aggregation schemes are compared, namely, the majority vote, the mean, the weighted mean and the Dawid-Skene method. The method was validated on an independent data set using only a subset of EEG channels. RESULTS: The ensemble reaches accuracy comparable to a single detector trained on all the data when sufficient amount of data is available in each institution. CONCLUSION: The weighted mean aggregation scheme showed best performance, it was only marginally outperformed by the Dawid-Skene method when local detectors approach performance of a single detector trained on all available data. CLINICAL IMPACT: Ensemble learning allows training of reliable algorithms for neonatal EEG analysis without a need to share the potentially sensitive EEG data between institutions.

Past, Present and Future of Home video-electroencephalographic telemetry: A review of the development of in-home video-electroencephalographic recordings.

Video-electroencephalographic (EEG) monitoring is an essential tool in epileptology, conventionally carried out in a hospital epilepsy monitoring unit. Due to high costs and long waiting times for hospital admission, coupled with technological advances, several centers have developed and implemented video-EEG monitoring in the patient's home (home video-EEG telemetry [HVET]). Here, we review the history and current status of three general approaches to HVET: (1) supervised HVET, which entails setting up video-EEG in the patient's home with daily visiting technologist support; (2) mobile HVET (also termed ambulatory video-EEG), which entails attaching electrodes in a health care facility, supplying the patient and carers with the hardware and instructions, and then asking the patient and carer to set up recording at home without technologist support; and (3) cloud-based HVET, which adds to either of the previous models continuous streaming of video-EEG from the home to the health care provider, with the option to review data in near real time, troubleshoot hardware remotely, and interact remotely with the patient. Our experience shows that HVET can be highly cost-effective and is well received by patients. We note limitations related to long-term electrode attachment and correct camera placing while the patient is unsupervised at home, and concerns related to regulations regarding data privacy for cloud services. We believe that HVET opens significant new opportunities for research, especially in the field of understanding the many influences in seizure occurrence. We speculate that in the future HVET may merge into innovative new multisensor approaches to continuously monitoring people with epilepsy.