The method involves acquiring a set of samples of neural signals by a set of electrodes, and estimating a covariance matrix of the neural signals. The neural signals are classified, where the classification of the neural signals is realized in Riemannian geometry of positive definite symmetric matrices of dimensions equal to Riemannian geometry of the covariance matrix and tangent space of the Riemannian geometry. Independent claims are also included for the following: (1) a method for direct neuronal control (2) a system for classifying neuronal signals (3) a method for selecting acquisition electrodes of neuronal signal for direct neuronal control.
