Systems based on Brain Computer Interface (BCI) have been developed from the past three decades for assisting locked-in state patients. Researchers across the globe are developing new techniques to increase the BCI accuracy. In 1924 Dr. Hans Berger recorded the first EEG signal. The number of experimental measurements of brain activity has been done using human control commands. The main function of BCI is to convert and transmit human intentions into appropriate motion commands for the wheelchairs, robots, devices, and so forth. BCI allows improving the quality of life of disabled patients and letting them interact with their environment. Since the BCI signals are non-stationary, the main challenges in the non-invasive BCI system are to accurately detect and classify the signals. This paper reviews the State of Art of BCI and techniques used for feature extraction and classification using electroencephalogram (EEG) and highlights the need of adaptation concept.