EEG can be used to generate control signals for Brain Computer Interface (BCI) applications. In this paper Artificial Neural Networks (ANN) were used to classify mental tasks. Primary data was decomposed by the help of discrete wavelet transformation to extract features. Out of control events were identifiedusing statistical process control. These out of control events were removed from the original dataset. Two data sets were created, one containing all the events and other without the out of control events. Each dataset was divided into training dataset (80%) and testing dataset (20%). The classification efficiency of statistically controlled dataset was found to be 85.71%, as compared to 67.74% from the dataset without statistical process control.