Understanding neural plasticity for programming brain-machine systems

The rapidly emerging field of brain-machine interfaces (BMI) establishes a spectacular convergence between literature and neuroscience. But brain-machine interfaces emerge from two well-defined practical goals: creating more powerful computers and giving new hope to a broad segment of the disabled population. During the last century, the study of computers and of the brain have evolved in a reciprocal metaphor: the brain is investigated as an organ that processes information and computers have been developed starting from the dream of creating an artificial brain. Brain-machine interfaces provide us with the perspective of moving beyond metaphor and considering the possibility of accessing the computational power of neural tissue. Despite the speed with which today´s computers execute billions of operations, our brains have still unsurpassed performance when it comes to recognizing a face or controlling the complex dynamics of the arm. The computational power of biological systems has sparked research aimed at mimicking neurobiological processes in artificial system. Perhaps the most well-known outcomes of this research are artificial neural-networks, which attempt to emulate some of the key features of neural information processing. More recently, a new idea has begun to take shape: the idea of constructing hybrid computers in which neurons are grown over a semiconductor substrate.