Electroencephalography (EEG) is a technique used to measure the electrical impulses generated by the brain. EEG analysis is a promising methodology for brain-computer interfaces (BCIs). It allows for a broader range of input signals than conventional, mechanically controlled prosthetics, or myoelectric prosthetics and is also non-invasive. However, the processing of EEG data presents a serious challenge due to a low signal to noise ratio and a tremendous degree of signal complexity. The main method used to resolve this challenge involves digital signal processing and machine learning algorithms. A wide range of algorithms have been applied with varying degrees of success, and a successful BCI will use these algorithms to increase our understanding of neurophysiology, then apply that knowledge in a multi-step classification algorithm which combines a variety of techniques and signal inputs. The focus of this project was to attempt to classify individual finger movements, based on EEG signals collected from over the motor cortex. This presents a significant signal processing hurtle, which must be overcome in any fully functioning prosthetic. This was accomplished by analyzing slow cortical potentials as well as sensory motor rhythms, and comparing the successful classification rate of a linear discriminant analysis to that of a CWT-based template algorithm.