Key points are not available for this paper at this time.
We investigated whether automatic control of a hand grasp neuroprosthesis by means of signals from natural sensors in the skin of the index finger can mimic the natural control of grasp force in an important task of daily living, namely eating. We designed a simulated eating task with the same ratio of rest and activity as was found on average in a video analysis of three meals consumed in a social environment. An instrumented fork measured grasp force as well as the force in the long axis and perpendicular to the long axis at the tip of the fork. The simulated eating task was performed by a tetraplegic volunteer using a hand grasp neuroprosthesis both with and without use of feedback from the natural sensors. Further, 10 able-bodied volunteers performed the task with the same (lateral) grasp as the tetraplegic volunteer to obtain measures for improving the control strategy of the hand grasp neuroprosthesis. We have shown that a hand grasp neuroprosthesis incorporating natural sensory feedback can to some extent mimic the natural application of grasp force on a fork during simulated eating. The mean grasp force during active phases was higher than the mean grasp force during inactive phases. The mean grasp force applied during a simulated eating task was reduced by using the system with sensory feedback compared to using the system without sensory feedback.
Inmann et al. (Tue,) studied this question.