Over the last 2 decades, advances in microfabrication and digital signal processing technologies have enabled the development of machines that interface directly with neurons in the brain, spinal cord and periphery. These so-called “neural interfaces” serve as bi-directional communication channels, allowing information to be read-out from the brain in the form of neural recordings or written-in via patterned electrical stimulation. We are exploiting these technologies for two purposes: 1) to advance our understanding of how the nervous system senses and controls limb motion, and 2) to develop advanced prosthetic devices that interface directly with the nervous system for control. The ultimate goal is to create neuroprosthetic limbs that literally look, feel, and function naturally, with user intention and state feedback communicated by a neural interface linking the prosthesis to the brain. One key to enabling this communication is to understand how neurons encode information. Fortunately, neuroscientists and engineers have made great progress in deciphering the language of the nervous system, and we are now able to “decode” information in real-time from measurements of neural activity. Conversely, electrical stimulation pulses can be patterned appropriately to activate sensory neurons, for example, to restore hearing to persons with profound deafness. My talk will focus on research in my lab that is aimed at understanding how somatosensory neurons encode information about touch and proprioception (i.e. sense of body position). I will also highlight some of our work in the recording and decoding of neural signals from the brain for expression of motor commands.