Quantitative EEG features and machine learning classifiers for eye-blink artifact detection: A comparative study

Ocular artifact, namely eye-blink is an inevitable and one of the most destructive noises EEG signals. Many solutions detecting artifact were proposed. Different subsets features Machine Learning (ML) classifiers used for this purpose. But no comprehensive comparison these ML was presented. This paper presents twelve five classifiers, commonly in existing studies detection artifacts. An dataset, containing 2958 epochs eye-blink, non-eye-blink, eye-blink-like (non-eye-blink) activities, study. The performance each feature classifier has been measured using accuracy, precision, recall, f1-score. Experimental results reveal that scalp topography potential among selected best performing Artificial Neural Network (ANN) classifiers. combination ANN performed as powerful feature-classifier combination. However, it expected findings study will help future researchers to select appropriate building models.