TY - JOUR T1 - Shaping the Dynamics of a Bidirectional Neural Interface A1 - Vato, Alessandro A1 - Semprini, Marianna A1 - Maggiolini, Emma A1 - Szymanski, Francois D. A1 - Fadiga, Luciano A1 - Panzeri, Stefano A1 - Mussa-Ivaldi, Ferdinando A. Y1 - 2012/07/19 N2 - Author Summary Brain-machine interfaces establish new communication channels between the brain and the external world with the goal of restoring sensory and motor functions for people with severe paralysis or sensory impairments. Current methodologies are based on decoding the motor intent from the recorded neural activity and transforming the extracted information into motor commands to control external devices as robotic arms. We developed a novel computational approach, based on the concept of programming dynamical behaviors trough the bi-directional sensory-motor interaction between the brain and the connected external device. This approach is based on the emulation of some control features of a biological interface, the spinal cord. The first prototype of our interface controls the state of motion of a simulated point mass in a viscous medium. The position of the point mass is encoded into a stimulus to the somatosensory cortex of an anesthetized rat. The evoked activity of a population of motor cortical neurons is decoded into a force vector applied to the point mass. The parameters of the encoder and of the decoder are set to approximate a desired force field. In the first test of the interface, we obtained a family of trajectories that converged upon a stable attractor. JF - PLOS Computational Biology JA - PLOS Computational Biology VL - 8 IS - 7 UR - https://doi.org/10.1371/journal.pcbi.1002578 SP - e1002578 EP - PB - Public Library of Science M3 - doi:10.1371/journal.pcbi.1002578 ER -