In this work, we propose the correction of univariate, order techniques can be used to decompose the signal into single channel EEGs using projective subspace techniques. The uncorrelated components. The multidimensional signal is then biomedical signals which often represent one dimensional time projected to the most significant directions computed using series, need to be transformed to multi-dimensional signal vectors projecte to tecmositsignificant d rections compedun for the latter techniques to be applicable. The transformation s ingular vposti ( ) pncip componen can be achieved by embedding an observed signal in its delayed analysis (PCA). Singular spectrum analysis (SSA) [4] used coordinates. We propose the application of two non-linear subspace in climatic, meteorologic and geophysics data analysis is the techniques to the obtained multidimensional signal. One of the most widely used technique that follows this strategy. The techniques consists in a modified version of Singular Spectrum general purpose of SSA is to decompose the embedded signal Analysis (SSA) and the other is kernel Principal Component Analvectors into additive components. This can be used to separate ysis (KPCA) implemented using a reduced rank approximation of the kernel matrix. Both nonlinear subspace projection techniques noise contributons from a recorded signal by estmating those are applied to an electroencephalogram (EEG) signal recorded directions, corresponding to the L largest eigenvalues, which in the frontal channel to extract its prominent electrooculogram can be associated with the eigenvectors spanning the signal (EOG) interference.subspace. The remaining orthogonal directions then can be associated with the noise subspace. Reconstructing the signal Keywords -Subspace Techniques, local SSA, KPCA, EOGs, EEG using only those L dominant components then can result in a substantial noise reduction of the recorded signals. The time