Paweł Niedbalski

Martyna Poziomska, Marian Dovgialo, Przemysław Olbratowski et al.

This study investigates the impact of quantity and diversity of data on the performance of various machine-learning models for detecting general EEG pathology. We utilized an EEG dataset of 2,993 recordings from Temple University Hospital and a dataset of 55,787 recordings from Elmiko Biosignals sp. z o.o. The latter contains data from 39 hospitals and a diverse patient set with varied conditions. Thus, we introduce the Elmiko dataset - the largest publicly available EEG corpus. Our findings show that small and consistent datasets enable a wide range of models to achieve high accuracy; however, variations in pathological conditions, recording protocols, and labeling standards lead to significant performance degradation. Nonetheless, increasing the number of available recordings improves predictive accuracy and may even compensate for data diversity, particularly in neural networks based on attention mechanism or transformer architecture. A meta-model that combined these networks with a gradient-boosting approach using handcrafted features demonstrated superior performance across varied datasets.

Lukasz Czekaj, Wojciech Ziembla, Pawel Jezierski et al.

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.