In recent years, several researches and developments have been made in biomedical engineering, which aim to improve the healthcare diagnosis and treatment. Signal analysis methods such as Fourier transform and Wavelet transform are widely used in this field. However, these methods are limited to the linear and stationary signal analysis. As a result, they are not well adaptive methods for biomedical signals, which in fact are mostly nonlinear and non-stationary signals. Empirical Mode Decomposition (EMD) is a novel technique that has been widely applied in the signal processing. EMD has been demonstrated as a method for the data processing of nonlinear and non-stationary signals. This method is to decompose a signal into a sum of finite number of “intrinsic mode functions” (IMF). In this thesis, EMD scheme is applied to analyze the steady-state visually evoked potentials (SSVEP) in electroencephalogram (EEG). Based on the EMD, the oscillatory activities of the decomposed SSVEP signal are analyzed. It drives us to focus on the investigation of the very low frequency of the SSVEP signal. Based on the observation of the IMF, this thesis proposes a method to detect the transition response when a person turns from an attentively focusing stage into an unfocused attention stage during the experiment. As a consequence, this may be used for detecting the idle period of a SSVEP based Brain-Computer Interface (BCI) system. After the evaluation, the attention-to-rest transition is detected with an accuracy of 82.6%. The occurrence of the very low frequency can be explained that the attention-to-rest transition will enhance the delta wave in EEG.