Implementing a Microneurography Setup for Online Denoising of Peripheral Motor Activity: Preliminary Results

Microneurography has been proposed, since its introduction in the 1960s, as a valuable tool for the study of peripheral neural control of movement, which could drastically improve the current development of neuroprosthetic limbs. The current work on robotic neuroprostheses is predominately performed with amputees surgically implanted with neural electrodes, a procedure whose complexity is currently mastered by very few groups all around the world. The reduced number of reported experiments resolves in poor availability of databases of human peripheral nerve signals, which are needed to fully test the interfacing algorithms, so far limited to animal testing. On the other hand, microneurography is a fully safe and little invasive procedure which can be applied to healthy subjects as well as to amputees and which permits to access peripheral neural motor activity. In order to be implemented as a neuroprosthesis interface, though, the microneurographic data needs to undergo online analysis for motion artifacts removal and white Gaussian noise suppression, features currently missing from the commercial devices. In this paper we report the instrumentation we have been developing to satisfy these requirements. In particular, we currently equipped our setup with an online wavelet denoising filter which substantially reduces white Gaussian noise. Here we present our preliminary results.