Deep learning based motor imagery recognition model for brain computer interfaces

Abstract

Brain Computer Interface (BCI) is widely employed for connecting brain and external environment by the recognition of brain activity and transforming it to messages. Electroencephalography (EEG) based on BCI, principally using Motor Imagery (MI) data becomes a common tool for recognition process. BCI allows the brain signals to control the external devices and also helps a disabled person suffering from neuromuscular disorders. Motor Imaging (MI) has been widely employed in the domains of neurological rehabilitation and robot control as an essential model of spontaneous Brain-Computer Interfaces (BCIs). The activities for motor Imagery (MI) movements in Electroencephalography (EEG) are still interesting and challenging. Although several approaches for feature extraction and classification based on MI signals have recently been presented by researchers, BCI remains a popular research topic due to its applications in various domains like neuro-rehabilitation, neuroprosthetics, and gaming. In any BCI system, the two most essential steps are feature extraction and classification method. The training time is reduced and non-stationary problem is managed by applying Empirical Mode Decomposition (EMD) and mixing their Intrinsic Mode Functions (IMFs) in feature extraction technique. However, in this thesis work, the MI classification is improved by the performance of Deep Learning (DL) concept. In proposed system two-moment imagination of right hand and right foot from the BCI competition datasets has been taken and classification methods utilizing Conventional Neural Network (CNN) and Generative Adversarial Network (GAN) are developed. The proposed GAN classification technique achieves better classification accuracy in terms of 95.29%. newline