Modeling of neural networks based on transient response analysis of EEG signals from Broca's area

In recent years, brain's electrical activity is being extensively studied by researchers and application developers to realize more reliable Brain Computer interfacing (BCI) technology which can control and manipulate the physical objects in the real world. In this paper, we have extracted electroencephalography (EEG) signals from Broca's area with an EEG headband which consists of two inexpensive stainless steel electrodes without employing any noise treating hardware. Broca's area is a center in brain that stores language related knowledge. Sensory signals from skin act as interferences while a user is operating a BCI in real time environment. In this paper, the characterization of those unwanted sensory signals is done in a much simpler way. By passing the amplitude normalized neural signals through low pass filters with different cut off frequencies, it is found that transient response of neural networks in Broca's area is an under damped response. From the results of this analysis, corresponding neural network is modeled as a control system with second order transfer function by using natural frequency and damping ratio values.