The advent of high quality multi-channel EEG recording systems and the significant advance in the field of machine learning enabled decoding complex motion features, such as the direction of hand movement in real (Waldert et al. 2008) and imaginary conditions (Ofner and Muller-Putz 2015). These decoding algorithms usually use classifiers like linear discriminant analysis (LDA) and support vector machine (SVM) on signal features to decode classes of movements, while other algorithms such as multivariate regression (MVR) are used to decode hand movement trajectories (Rickert et al. 2005). Currently, the discrimination of actual or imagined pointing movements to different targets with the same limb is modest. This is due to the fact that these motor tasks activate essentially the same motor-related neural networks for all targets, thus, the discrimination between different actual or imagined movements has to rely more heavily on differences in temporal and spectral features of the neural activity rather than on differences in spatial activation patterns. In this study, we suggest that forming a distinct, target-specific neural activation pattern can be enhanced by segregating the internal representation of the different conditions (targets) and that this may be pursued by forming a strong recollection of each of the hand movements (muscle/kinesthetic memory) (Krakauer and, Shadmehr 2006). As kinesthetic memory involves consolidating a specific motor task into memory through repetition (Karni et al. 1995, Shadmehr and Holcomb 1997), a block design task was utilizedin which each target is pointed at for several times consecutively.