Objective:The first objective of this paper is to establish a relationship between brain activities and motor imagery (MI) movements through mapping functions. The second objective is to distinguish left and right hands MI movements of the subjects. Approach:In this paper, the dynamical behavior of brain activities is well explained by the discrete wavelet transform (DWT) and phase space reconstruction (PSR) techniques. The DWT is employed to identify μ and β frequency bands from raw electroencephalogram (EEG) signals. In order to extract rest and MI features and their corresponding brain patterns, a PSR-based common spatial pattern (CSP) technique is applied to both the frequency bands. The obtained MI features used to train a support vector machine (SVM) model to detect MI movements. Results:The proposed method is tested on benchmark BCI competition (II, III and IV) datasets. The mentioned technique yields a higher value of classification accuracy (%CA), Cohen’s kappa coefficients (K) and information transfer rate (ITR) in bits per trial. Significance:The μ band is highly responsive than the β band in MI period. The PSR is a powerful graphical technique to investigate the brain activities.