Automatic determination of the optimal shape of a surface electrode: Selective stimulation

Abstract

We present a method for automatic determination of the shape and position of the surface electrode for selective control of fingers extension and flexion by means of electrical stimulation. The multi-pad electrodes used in the experiments comprised 24 pads (1 cm diameter) distributed over an area (7 cm × 10 cm) positioned over dorsal and volar aspects of the forearm. The four-channel stimulation system for grasping comprised also an oval reference electrode over the carpal tunnel, and two oval electrodes over the thenar and thumb extensor muscles. We measured seven angles: proximal inter-phalangeal and metacarpal phalangeal index and ring finger joint rotations, wrist extension/flexion and ulnar/radial rotation, and pronation/supination of the forearm. The optimal electrode was determined as the combination of pads that led to fingers, wrist and forearm rotations being similar to the trajectories of healthy individuals when grasping. The similarity of trajectories was assessed by analyzing the aggregate error defined as the sum of squares of differences between the angles measured when stimulating the forearm in tetraplegics and the angles measured in healthy individuals. The aggregate errors were determined from measurements during sequential stimulation of each of the 24 pads. The analysis comprised hand opening and closing for palmar and lateral grasps. The time for determining the optimal electrode was about 10 min. The optimal electrodes had different branched shapes in each of the six tetraplegics; however, once determined they remained unchanged when tested on different days.

Introduction

A spinal cord injury at the cervical level or a cerebro-vascular accident (stroke) often results in life-long paralysis of one or both upper extremities. One method for providing the compromised function is the application of a Neural Prosthesis (NP) that activates muscles by Electrical Stimulation (ES). Originally, a NP with surface electrodes was considered mainly as an orthotic necessity to improve daily functioning, but today the NP with surface electrodes is mostly envisioned for therapy (Popović et al., 2003, Popović et al., 2004). The therapeutic NP requires an interface that is easy to use. Almost 30 years ago, Nathan et al. began studying how to simplify their use (Nathan, 1979, Nathan, 1990, Nathan and Ohry, 1990, Sagi-Dolev et al., 1995). The major difficulty was due to the interference of non-agonist muscles when surface electrodes were applied. This interference (e.g., wrist radial/ulnar rotation and flexion/extension) follows the spread of the electrical current over many motor nerves that should not be activated. The solution suggested was to apply a customized splint to limit the non-functional movement, provide additional support, and house the customized insets of best positioned electrodes (http://www/nessltd.com, 2008; IJzerman et al., 1996).

The development of technologies for the fabrication of electrodes in various shapes and sizes, as well as the availability of smart electronic stimulators and conductive textiles, allows better steering of stimulation currents through target tissues. There are only few studies presented in scientific journals on the effects of multi-pad surface electrodes which include individuals with disability. Nathan (1979) presented a detailed map of skin positions which corresponds to various muscle innervation points. He found a substantial variation between the positions within the population which participated in his study. In the work that followed Sagi-Dolev et al. (1995) analyzed current density under the surface of stimulation electrodes. The study combined software simulation and hardware phantom simulation, aimed at developing a method that would eliminate the typical problem of activation of the Flexor Carpi m. when trying to activate the Flexor Digitorum Superificialis m. and the Flexor Digitorum Profundus m. which occurs during externally controlled grasping, as well as the activation of afferent pathways when trying to control prehension. The validated hypothesis was that stimulation overflow can be reduced by controlling the current density distribution under the surface stimulating electrodes (Sagi-Dolev et al., 1995). The density distribution depends on the size, orientation, geometry, and impedance of the electrodes. Similar findings from simulation by distributed electrical fields were reported by Livshitz et al. (2001). Fuji and colleagues (2004) proposed an electrical stimulation system with trained super-multichannel surface electrodes, but there was no follow up of this work. Elsaify et al. (2004a) suggested a method of using the twitch response and feedback sensors (Elsaify et al., 2004b) for selecting an optimal electrode which was created from several conventional single pad electrodes. Lawrence et al. (2004) and Keller et al., 2007a, Keller et al., 2007b presented a procedure for selective stimulation based on detailed tissue modeling. This research provides evidence regarding the best size of the pad and the distance between the pads within a multi-pad electrode positioned on the skin. The quantitative experimental results came mostly from healthy individuals. O’Dwyer et al. (2006) demonstrated that a multi-pad electrode resulted in better selectivity during stimulation. They also presented a selection method based on signals from the sensors. Our own research suggests that the use of multi-pad electrode is efficient; however, the methodology of creating an optimal shape and position remained an open question (Bijelić et al., 2004, Popović-Bijelić et al., 2005).

In this paper, we present a method for automatic determination of the optimal shape of a multi-pad surface electrode for the functional activation of finger flexors and extensors. The goal of the optimization procedure was to select pads on the multi-pad electrode, which when activated, result in joint excursions of the fingers of individuals with tetraplegia that are minimally different from the joint excursions measured in healthy individuals during palmar and lateral grasps.