The field of invasive Brain Machine Interfaces (iBMI) has during the last ten years proven its enormous potential in restoring movements in paralyzed patients. The present doctoral thesis introduces a new dimension to this field by using complex cognitive behavior to drive an iBMI. In this respect, visual processes including spatial attention and perception are of special interest. This thesis project has three principal objectives: first, show the feasibility of decoding cognitive information in an offline setup. Second, evaluate the decoding of cognitive information in a real time experimental setup and third, investigate the impact of this setup in a changing environment, this both from the perspective of driving real time brain-Machine interfaces and that of understanding distributed populational neuronal codes. In line with the first objective of this thesis, an evaluation of several different classification techniques has been carried out in order to choose the best suited method for reading out cognitive information. The study provides evidence that visual information can be read out with similar performance as cognitive information. This study is the first study aiming at explicitly comparing the read out of sensory and cognitive information. The two last objectives of the present thesis are carried out on data from a new real-Time experimental setup. First we demonstrate the feasibility of real-Time readout of spatial attention and perception and we bring about a novel understanding about these two cognitive processes. Second, we show that in a changing environment, remarkable reconfigurations of prefrontal neural populations occur under certain contexts while left unaffected by other contexts. This Ph.D. thesis has taken the field of cognitive brain-Machine interfaces one step further by establishing the impact on spatial attention and perception of a changing environment. Facing the many neurological and neurodegenerative pathologies existing today, this thesis provides a steady ground for the continuation of research in this area