Recent research in BCI focuses not only on developing a new communication channel for severely handicapped people but also on applications for rehabilitation, multimedia, communication, virtual reality, entertainment and relaxation. Most of them fall in the domain of human-computer interfaces (HCIs) designed for interaction between brain, eyes, body and computer or robot. For brain signal acquisition several technologies have been applied, for example electroencephalography (EEG), magneto encephalography (MEG), functional magnetic resonance imaging (fMRI) and near infrared spectroscopy (NIRS). Portability and cost effectiveness problems channeled BCI systems to exploit EEG signals mostly. This paper presents a methodology and recommended parameter setting, for representation in time-frequency scale of EEG signals. It refines the detection of event-related changes in the signals, revealing specific patterns of rhythms, for actual and intended physical movement. The result shows that, with well defined window length it is possible to improve localization of specific frequencies within the brain activity. This lead to the fact that actual muscle activity form could be identified from EEG signals. Using the referenced methodology a wide range of HCIs systems can be designed to perform specific tasks for the benefit of the end-user.