Abstract:Brain–computer interface (BCI) is an active domain which has attracted attention of the research community in recent years. It offers huge potential as a technology which can estimate the intention of a user by analysis of brain signals and establish a communication channel directly between a human brain and an external device. Electroencephalography (EEG) is the most popular signal acquisition technique due to its ease of use and simplicity. In EEG-based BCI systems, electrodes are placed on specific positions on the scalp of the subject to record electrical activity. The BCI system consists of sequential stages of signal acquisition, its preprocessing, feature extraction and feature classification. It is an active research area which has a focus on improving classification accuracy in motor imagery-based BCI systems. The first stage in a BCI system is to acquire EEG signals from different positions of the scalp of the human subject. The acquired brain signals are preprocessed to remove artifacts before these are fed to feature the extraction stage. In this paper, independent component analysis (ICA) technique is used to remove artifacts from acquired signals. Filter bank common spatial pattern (FBCSP) technique is then used for feature extraction and feature selection. A feature classification approach based on support vector machine (SVM) is proposed in this work and its performance is enhanced by optimizing its polynomial kernel parameters. Selection of kernel parameters is done by grid search method using the fivefold cross-validation procedure. The proposed approach is then executed on publicly available data set 2a of BCI Competition IV. Results show that the proposed approach offers higher classification accuracy and lower misclassification rate as compared to other methods executed on the same dataset, as reported in literature.

Feature Extraction And Parameters Selection Of Classification Model On Brain-Computer Interface

Simultaneously Optimizing Spatial Spectral Features Based on Mutual Information for Eeg Classification

Feature Extraction and Pattern Classification for EEG in Brain-Computer Interface

Study on Classification Algorithms of a Brain—Computer Interface Based on Multichannel Linear Descriptors

Evaluation of Classical Features and Classifiers in Brain-Computer Interface Tasks

Feature Extraction of Brain–Computer Interface Electroencephalogram Based on Motor Imagery

Temporal and Spatial Features of Single-Trial EEG for Brain-Computer Interface

Channel Selection and Feature Extraction of Ecog-Based Brain-Computer Interface Using Band Power

Research on Feature Extraction Algorithm Commonly Used in Brain-computer Interface Technology

An Automatic Optimum Data Selection Method for EEG-based Brain-computer Interface

Application of Tripolar Concentric Electrodes and Prefeature Selection Algorithm for Brain–Computer Interface

Research of Classification Methods and Evaluation Criteria in Brain-Computer Interface System

Research on the Multi-Classifier Features of the Motor Imagery EEG Signals in the Brain Computer Interface

Feature Extraction and Pattern Classification on Mining Electroencephalography Data for Brain-Computer Interface

Optimization of parameters for improving the performance of EEG-based BCI system

Brain-computer Interface Design Based on Slow Cortical Potentials Using Matlab/simulink

Wavelet and Common Spatial Pattern for EEG Signal Feature Extraction and Classification

A Pattern Recognition Method in Brain-Computer Interface

Research on Heuristic Feature Extraction and Classification of EEG Signal Based on BCI Data Set

Feature Extraction of Brain-Computer Interface Based on Second Order Blind Identification and ICA

Feature Extraction and Classification of EEG in Online Brain-Computer Interface