Analysis of The Inter-subject and Inter-session Variability Phenomena Effect on the MI-EEG Datasets using The Brain Topographic Map

Abstract

The development of existing diagnostic technology has resulted in the developing the Brain Computer Interface (BCI). BCI is expected to connect the human nervous system with the outside world to provide more efficient communication and control in human body. Motor Imagery (MI) is the type of brain signal that commonly used in BCI practice. MI allows us to be able to control the movement of certain limbs only by imaginative processes. Although the MI-EEG signal has great potential, MI signal processing is still difficult to be done. MI signal has abstract pattern, making it difficult to distinguish one type of movement from another, especially if the MI signal used is not only from one subject and at the same recording time. This phenomenon is called inter-subject and inter-session variability. Based on this problem, authors conducted a study using the WPT-CSP method. This method will decompose signals into multiple frequency bands, and then filtered using spatial filter to obtain a better temporal and spatial resolution for the MI signal. The results of this method are then displayed on the Brain Topographic Map (Topomap) to show the activity level of the brain regions. The dataset used in this study is dataset 2a from Brain-Computer Interface Competition (BCIC) IV. The results of the research show that the phenomenon of inter-subject and inter-session variability can be observed more easily using the Topomap. These results also indicate that a new method is needed to overcome this phenomenon.