Labeler-hot Detection of EEG Epileptic Transients
Authors:
Lukasz Czekaj,
Wojciech Ziembla,
Pawel Jezierski,
Pawel Swiniarski,
Anna Kolodziejak,
Pawel Ogniewski,
Pawel Niedbalski,
Anna Jezierska,
Daniel Wesierski
Abstract:
Preventing early progression of epilepsy and so the severity of seizures requires an effective diagnosis. Epileptic transients indicate the ability to develop seizures but humans overlook such brief events in an electroencephalogram (EEG) what compromises patient treatment. Traditionally, training of the EEG event detection algorithms has relied on ground truth labels, obtained from the consensus…
▽ More
Preventing early progression of epilepsy and so the severity of seizures requires an effective diagnosis. Epileptic transients indicate the ability to develop seizures but humans overlook such brief events in an electroencephalogram (EEG) what compromises patient treatment. Traditionally, training of the EEG event detection algorithms has relied on ground truth labels, obtained from the consensus of the majority of labelers. In this work, we go beyond labeler consensus on EEG data. Our event descriptor integrates EEG signal features with one-hot encoded labeler category that is a key to improved generalization performance. Notably, boosted decision trees take advantage of singly-labeled but more varied training sets. Our quantitative experiments show the proposed labeler-hot epileptic event detector consistently outperforms a consensus-trained detector and maintains confidence bounds of the detection. The results on our infant EEG recordings suggest datasets can gain higher event variety faster and thus better performance by shifting available human effort from consensus-oriented to separate labeling when labels include both, the event and the labeler category.
△ Less
Submitted 9 July, 2019; v1 submitted 11 March, 2019;
originally announced March 2019.
