I. Introduction
People interact with the environment through verbal and nonverbal means. However, neurological disorders like motor neuron diseases may impair this interaction as they prevent voluntary muscular activities [1]. Provided that the cognitive ability is not affected by these disorders, a direct communication channel between a brain and an external device enables nonmuscular interaction with the environment using only cognitive functions. Brain computer interface (BCI), which involves the acquisition of the brain activity using a neuroimaging technology followed by its analysis using preprocessing, feature extraction and classification operations, provides such a channel. Electroencephalogram (EEG) is a neuroimaging technology commonly used to acquire brain activity for BCI [2], [3] due to its noninvasive nature, low cost and ease of use [4]. One of the most popular features of EEG utilized in BCI is the P300 event related potential (P300 ERP) [5], [6]. An advantage of the utilization of the P300 ERP over other EEG features in BCI is that the P300 ERP does not require any training since it is an automated response [7].