Supervised learning in automatic channel selection for epileptic seizure detection

作者：

Highlights：

• An automatic channel selection for seizure detection is proposed.

• Computational efficiency is improved by 49.4%, while maintaining accuracy close to 96.5%.

• Mean detection delay is improved by 400 ms to 2.77s without degrading specificity.

• Seizure onsets are detected at 91.95% sensitivity and 94.05% specificity.

摘要

•An automatic channel selection for seizure detection is proposed.•Computational efficiency is improved by 49.4%, while maintaining accuracy close to 96.5%.•Mean detection delay is improved by 400 ms to 2.77s without degrading specificity.•Seizure onsets are detected at 91.95% sensitivity and 94.05% specificity.

论文关键词：Seizure detection,iEEG,Random Forest,Automatic channel selection

论文评审过程：Received 1 February 2017, Revised 21 April 2017, Accepted 20 May 2017, Available online 29 May 2017, Version of Record 30 June 2017.

论文官网地址：https://doi.org/10.1016/j.eswa.2017.05.055