Reproducing tactile and proprioception based on the human-in-the-closed-loop conceptual approach

Abstract

Prosthetic limb embodiment remains a significant challenge for many amputees due to traditional designs' lack of sensory feedback. To address this challenge, the effectiveness of non-invasive neuromuscular electrical stimulation (NMES) controlled by a hybrid proportional-differential (PD)-Fuzzy logic system was evaluated for providing real-time proprioception and tactile feedback. The study used a human-in-the-closed-loop approach with ten participants: five upper limb amputees and five non-disabled individuals as the control group. An applied force, the joint angle of a prosthetic hand's finger, and surface electromyography signals generated by the biceps muscle all regulate the intensity of sensory feedback. Additionally, the C6 and C7 myotomes were selected as elicitation sites. The average threshold for detecting action motion and force was around 21° and 1.524N, respectively. The participants successfully reproduced desired joint angles within the range of 0°-110° at five separate intervals. In the weight recognition experiment, the amputee participant's minimum number of false predictions was four. The highest accuracy achieved was 80.66% in detecting object size and stiffness. Additionally, unpaired t-tests were performed for the means of the results of the experiments to determine statistically significant differences between groups. The results suggest that stimulation of myotomes by NMES is an effective non-invasive method for delivering rich multimodal sensation information to individuals with disabilities, including upper limb amputees, without needing visual or auditory cues. These findings contribute to the development of non-invasive sensory substitution in prostheses.