Stimulation details

The technique for evoking a tapping sensation from the nerves was similar to what has been described previously.^10,11 Participants were instructed to wash the area where the electrodes would be applied (cubital fossa or lateral knee) with soap and water and were seated in an upright chair. Rubbing alcohol was used to clean the skin before applying the electrodes. Electrodes were placed over the median, ulnar and/or peroneal nerve as described previously. ¹¹

Stimulation was supplied via voltage-controlled, charge-balanced, biphasic, non-symmetric square pulses. Pulses were non-symmetric in that the anodic phase was set to a maximum value of 4 V with a width as needed to balance the charge. ¹⁶ Voltage-controlled stimulation was used to decrease the risk of high current density in the case of reduced adhesion from the surface electrodes. The stimulation waveforms were created in MATLAB (2013a) and delivered using a National Instruments USB DAQ (NI USB-6229, Austin, TX) and an isolated biostimulator (Coulbourn Instruments model A13-75, Pittsburg, PA). Adhesive electrodes were cut to 30 mm by 17 mm (ValuTrode, Axelgaard Manufacturing, Fallbrook, CA).

Stimulation parameters were chosen as described previously. ¹¹ Briefly, voltage amplitude was slowly increased at a frequency of 50 Hz to determine an approximate threshold and set a maximum comfort level (pulse durations used were 100 μs and 500 μs). A more complete threshold was determined using Parameter Estimation by Sequential Testing (PEST)^17,18 with four threshold points. ¹⁰ Stimulation between 1 and 4 Hz had been found to produce a tapping sensation in a majority of participants ¹¹ so the frequency was initially set to 2 Hz and voltages at 75% and 90% of the range between threshold and maximum were tested. The participant identified the parameter combination that felt the most like a tap and then this voltage-pulse width combination was tested at 1, 2 and 3 Hz to choose the final frequency.

A four degree of freedom robotic arm (Lynxmotion AL5D) was used to automate the tapping and allow synchronization with the stimulation. The robot was positioned over the rubber extremity to match where the participant reported the sensation (Figure 1). Attachments of different sizes and shapes were available to allow a better match between a participant’s reported sensation area and what they saw happening on the rubber extremity. The timing between when the stimulation was sent and when the robot moved could be adjusted for each participant to provide simultaneous visual and somatosensory information (default delay for the stimulation was 0.1345 s).