Eye Blink Artifact Detection With Novel Optimized Multi-Dimensional Electroencephalogram Features

Abstract

Accurate eye blink artifact detection is essential for electroencephalogram (EEG) analysis and auxiliary analysis of nervous system diseases, especially in the presence of the frontal epileptiform discharges. In this paper, we develop a novel eye blink artifact detection algorithm based on optimally selected multi-dimensional EEG features. Specific efforts have been paid to filtering the frontal epileptiform discharges, where an unsupervised learning exploiting the EEG signal physiological characteristics and smooth nonlinear energy operator (SNEO) based on the K-means clustering has been firstly proposed. Multiple statistical EEG features derived from the frontal electrodes and other electrodes are then extracted to characterize eye blink artifacts. Discriminative feature selection scheme based on the variance filtering and Relief algorithms has been respectively studied, and the average correlation coefficient (ACC) is applied for feature optimization evaluation. The eye blink artifact detection is finally achieved based on the support vector machine (SVM) trained on the optimized EEG features. The effectiveness of the proposed algorithm is demonstrated by experiments carried out on the EEG database of 11 subjects recorded from the Children's Hospital, Zhejiang University School of Medicine (CHZU). Comparisons to several state-of-the-art (SOTA) eye blink artifact detection methods are also presented.

Multilevel Feature Learning Method for Accurate Interictal Epileptiform Spike Detection

“…Interpreting the raw EEG is challenging because of the encoded information in EEG signals [12], [22]. Therefore, mathematical transformations were conventionally applied to EEG signals to quantify their properties using the domain knowledge in the concrete-level as shown in Fig.…”

“…Additionally, the correlation of information between channels is useful [5], [17], in other words the spatial information of inter-channels can provide useful differentiation information, because the spike usually occurs simultaneously in multiple related channels [18]- [20]. Moreover, nonlinear dynamic feature extraction techniques consider that the energy of spike is greater than that of the noise within the same time interval [21], [22], which can be used to distinguish the spike from the background signals. Therefore, in our preliminary study for spike detection [12], the aforementioned feature representation in the concrete-level, i.e., spatial-temporal-frequency multidomain features were integrated and could comprehensively represent the effective mimicry representational information of the spike to improve detection performance.…”

Deep-Learning Approach for Efficient Eye-blink Detection with Hybrid Optimization Concept

“…We're feeding it data and then training this to understand from it in order to make predictions. In the end, the algorithm produces whether any of the frames may be categorized as a feature or not by defining a minimal threshold [13][14][15][16].…”

“…The frequency and duration are the prime factors for blinking. For an adult, the Average blinking frequency is 15-20 blinks/min, while the blinking frequency of the children is lower [12][13][14][15].…”

Childhood epilepsy syndromes classification based on fused features of electroencephalogram and electrocardiogram

“…Although fruitful achievements have been presented in using EEG, ECG and other physiological signals for epilepsy analyses [15][16][17][18][19][20][21][22][23], it is noteworthy that existing research mainly focussed on epilepsy seizure onset detection while little attention has been paid to the childhood epileptic syndrome classification. The current classification of childhood epilepsy syndrome relies on expert experience and clinical characteristics, and the diagnostic accuracy rate was only about 50%.…”