Single-Trial Evoked Potentials Extraction Based on Sparsifying Transforms

Evoked potentials are widely used to diagnose diseases and disorders in the central nervous system. It is thus essential to develop fast algorithms which can track the variations of evoked potentials for a variety of clinical applications. The sparsity of signals in a certain transform domain or dictionary has been exploited in the extraction of noisy signal. However, it isn’t effective enough to extract the evoked potentials because the signal-to-noise ratio is extremely low. In this paper, we present a novel approach to solving evoked potentials extracting problem. Before the sparsifying the observations of evoked potentials, the observations are transformed to enhance the signal-to-noise ratio and sparsity. Then we can use the sparse representation algorithm to extract the evoked potentials. The alternating minimization algorithms are applied to calculate the transformation matrix and the sparse coefficients. We show the superiority of our approach over some filtering and sparse representation methods.