EEG-FES-Force-MMG closed-loop control systems of a volunteer with paraplegia considering motor imagery with fatigue recognition and automatic shut-off

Abstract

People with spinal cord injury (SCI) may have their paralyzed muscles activated through functional electrical stimulation (FES). This neuromodulation technique has been used frequently to assist in controlling the movement of neuroprostheses. Electroencephalography (EEG) is able to trigger FES from the motor imagery captured through movements intentions. This research presents an isometric neuromuscular control system of the quadriceps muscle activated by EEG. Additionally, the detection of neuromuscular fatigue through the mechanomyography (MMG) technique is proposed, which is used to shut-off the system. A pilot study was performed on a chronic 42-year-old paraplegic (no voluntary contraction below the spinal cord injury level T8) volunteer. To do so, the training procedure for EEG signals was divided into the calibration and feedback phases. In the first one, four EEG channels and the Linear Discriminant Analysis (LDA) classifier were used to classify between motor imagery of the right leg and remain at rest. The maximum accuracy obtained during this stage was 77%. In the feedback phase, the volunteer was able to activate FES through brain–computer interface (BCI) in two tests (defined as Test 1 and Test 2) with the same procedure in different days. The closed-loop force control was tested with the setpoint of 2 kgf and 2.5 kgf and proved to be stable on both tests, successfully turning off the FES using the fatigue threshold from the MMG signal, being the main contribution of this work.