Brain computer interfaces (BCI) systems utilize the neuronal activities to design control or communication devices. Motor signals, executed or imagined, modify the sensorimotor rhythms such as mu and beta oscillations in contralateral regions of the brain. Motor Imagery (MI) signals are described as signals generated during the planning stage of movement execution. There are large number of studies that confirms MI signals as robust signals for controlling BCI systems. We use EEG recordings obtained by using sensors placed on the scalp of user based on international 10-20 system and apply common-spatial pattern (CSP) based projection to classify imagined movements. We aim to study the learning behavior of BCI users while they some standard complex motor task by evaluating their performance over the time.