Development of a Robust Encoding Scheme for Delivering Artificial Sensory Information through an ICMS Electrical Interface

Abstract

Current users of brain-computer interface (BCI) technology must rely on visual feedback of cursor or robotic arm movement. The inherently long delays of visual processing likely contribute to relatively slow and unnatural control of BCIs. Despite increasing numbers of electrode sites and ever growing complexity of control algorithms, BCI technology has yet to achieve rapid, dexterous control comparable to an intact human system. We believe the lack of tactile perception and proprioceptive input imposes a fundamental limit on speed and accuracy of BCI-controlled prostheses or re-animated limbs. By artificially recreating this high-resolution pathway via Intra-Cortical Microwire Stimulation (ICMS), BCI stability and control may be substantially improved. Towards this aim, we are exploring cortical sensitivity of modulating stimulation parameters to further understand the critical stimulation features which best encode sensory intensity. Given the high dimensional parameter space of electrical stimulation, neural interface designers have many options for presenting graded sensation. Parameters such as amplitude, frequency, pulse-width, stimulation train duration, and electrode site could be modulated to vary the intensity of a percept. Using our novel center-out task for rodents as a behavioral metric of perceived intensity, we compare the perceptual resolution of modulating key parameters, both individually and combinatorically. Rodents perform discrimination tasks to measure psychometric curves and just-noticeable-differences (JNDs). Our results show that parameters which modulate the charge-per-pulse (CPP), have the highest resolution while parameters which modulate the frequency of presentations of those pulses have much lower resolution. Amplitude and pulse-width are critical parameters in understanding the mapping between parameter modulation and sensory perception of intensity. Our overarching goal is to formulate a general pattern for providing a high resolution feedback signal which can be mapped to any sensory modality. Based on our comprehensive exploration of electrical stimulation parameters, we present the state of the art encoding scheme for modulating the sensory perception of intensity.