Adaptive Control Methods Transform Neuroprosthetics

by James Cavuoto, editor

Currently, most of the systems on the market for neural prostheses and neurorehabilitation make use of electronic controllers that send regular signals to surface or implanted electrodes. Neural engineers at several research institutions and commercial firms are developing new systems that will be able to modify control signals automatically based on sensory feedback, discerned user intent, or adaptive algorithms that are modified on the fly.

One of the first categories of neuroprosthetics to employ adaptive control is foot-drop stimulators. The NESS L300 system from Bioness incorporates a gait sensor that adjusts control signals to accommodate uneven surfaces and changes in elevation. Victhom’s implantable NeuroStep device incorporates a sensing cuff electrode that is used to detect afferent signals generated by mechanoreceptors in the foot.

The Center for Adaptive Neural Systems at Arizona State University is currently pursuing several research projects that may have commercial applications in neuroprosthetics and neurorehabilitation. One project involves adaptive electrical stimulation for locomotor retraining, which seeks to improve outcomes for locomotor therapy following spinal cord injury. The research team, led by Jimmy Abbas and Ranu Jung, wants to use more appropriately timed muscle contractions and generate more repeatable movement patterns on a treadmill.

Abbas has spun off a commercial firm called customKYnetics Inc., which is developing a portable home-based stimulation system incorporating adaptive control. The ASU researchers make use of the entrainment of oscillations phenomenon to design algorithms and electronic circuits that mimic the functionality of neuromotor control systems.

A second commercial spinoff from the lab, Advensys LLC, received a $1 million grant from the U.S. Army to develop powered limb splints that can be used in combat settings to provide bipedal mobility to injured soldiers. Founded in 2004 by Ranu Jung, the company seeks to commercialize a suite of products to provide new orthotic and prosthetic options for people with lower limb dysfunction or lower limb amputation. The products will integrate biologically-inspired adaptive neuromorphic control systems technology with biomorphic compliant actuators, sensor systems, and lightweight orthotic/prosthetic components. Advensys’ first product, NOCS (neuromorphic orthotic control system), is a prototype control system based on neuromorphic principles that senses and controls an active powered orthosis for the lower extremity.

The Center for Adaptive Neural Systems recently received a $3.3 million bioengineering research partnership award from the National Institute for Biomedical Imaging and Bioengineering for neural-enabled prostheses that make use of sensorimotor integration.

The project proposes to develop a prosthetic system with electrodes implanted within the fascicles of peripheral nerves to provide upper extremity amputees with sensory feedback and active volitional control. The ASU team and its partners at the Mayo Clinic Arizona and two commercial firms seek to elicit meaningful sensations of hand opening and grip force. The team will then focus on using the neural interface to provide both sensation and control. A key feature will be bidirectional communication at speeds that enable real-time sensorimotor control of the prosthesis.