Human ability in identification of location and pulse number for electrocutaneous stimulation applied on the forearm

Abstract

Background: The need of a sensory feedback system that would improve users' acceptance in prostheses is generally recognized. Feedback of hand opening and position are among the most important concerns of prosthetic users. To address the two concerns, this study investigated the human capability to identify pulse number and location when electrical stimulation applied on the forearm skin. The pulse number may potentially be used to encode the opening of prosthetic hands and stimulation location to encode finger position.

Methods: Ten able-bodied subjects participated in the study. Three electrodes were placed transversely across the ventral forearm spatially encoding three fingers (i.e., thumb, index, and middle finger). Five different pulse numbers (1, 4, 8, 12, and 20) encoded five levels of hand opening. The study consisted of three experiments. In the three experiments, each after a training session, the subjects were required to identify among: (a) five stimulation locations, (b) five pulse numbers, or (c) ten paired combinations of location and pulse number, respectively. The subjects' performance in the three identification tasks was evaluated.

Results: The main results included: 1) the overall identification rate for stimulation location was 92.2 ± 6.2%, while the success rate in two-site stimulation was lower than one-site stimulation; 2) the overall identification rate for pulse number was 90.8 ± 6.0%, and the subjects showed different performance in identification of the five pulse numbers; 3) the overall identification rate decreased to 80.2 ± 11.7% when the subjects were identifying paired parameters.

Conclusions: The results indicated that the spatial (location) and temporal (pulse number) identification performance are promising in electrocutaneous stimulation on the forearm. The performance degraded when both parameters had to be identified likely due to increased cognitive load resulting from multiple tasks. Utilizing the proposed coding strategy in practical prosthetic hands remains to be investigated for clinical evaluation of its feasibility.

Figure 1

Electrode placement. Three self-adhesive solid gel surface electrodes (D1, D2, D3) were transversely placed 5 cm distant from the elbow crease on the ventral side of the left forearm. The three electrode sites were intended to encode the position of three fingers.

Figure 2

Stimulation locations and fingers to be mentally associated. The subjects were instructed to mentally associate the stimulation applied at five locations (D1, D2, D3, D1 & D2, D1 & D3) to one or two particular fingers. The shadowed electrodes and fingers represent the association.

Figure 3

Pulse numbers and hand opening levels to be mentally associated. The subjects were instructed to mentally associate five pulse numbers (1, 4, 8, 12, 20) to five levels of hand opening.

Figure 4

Just noticeable difference (JND) for pulse number. The curves of JND for pulse number were drawn based on the measurement at three pulse rates: 10, 20, 40 pps in one subject. A group of five pulse numbers was selected for each pulse rate, ensuring that the ‘spacing’ between two successive pulse numbers equal to or larger than the JND. The three groups of selected pulse numbers were marked on the three curves, respectively.

Figure 5

Identification rate (mean ± standard deviation, n = 10) for stimulation location. The three error bars show the overall identification rate, the identification rate in one-electrode stimulation, and in two-electrode stimulation, respectively. ** indicates a significant difference (p < 0.01).

Figure 6

Identification rate (mean ± standard deviation, n = 10) for pulse number. The error bars show the identification rates for five individual pulse numbers. ** indicates a significant difference (p < 0.01). * indicates a significant difference (p < 0.05).

Figure 7

Identification rate (mean ± standard deviation, n = 10) for paired parameters. The error bars show the identification rate for paired parameters, as well as their marginal success rates.

Figure 8

Identification rate for 10 individual subjects. The bar plot shows the identification rate obtained from the three experiments for 10 individual subjects.