Providing proprioceptive information through biomimetic multi-electrode stimulation patterns

Sombeck, Joseph T

doi:https://doi.org/10.21985/n2-x8sd-f648

Work

Providing proprioceptive information through biomimetic multi-electrode stimulation patterns

Public

Download PDF

Download All Files (.zip)

Persons with spinal cord injuries can use state-of-the-art brain-computer interfaces to control robotic arms. Despite this high-tech solution, their movements are slow and imprecise, much like those made by individuals who have lost proprioception, the sense of body position and movement. Intracortical microstimulation (ICMS) used to reactivate neural circuits in the somatosensory cortex is a promising approach for providing artificial proprioceptive feedback. While tactile interfaces have advanced to the point where ICMS can provide force and contact location feedback to a spinal cord injured patient, proprioceptive interfaces have proven more difficult to develop. Previous proprioceptive interfaces either required months of training to use or evoked unreliable sensations. Part of the difficulty in designing these interfaces is the complicated somatotopy in proprioceptive cortical areas, where even simple limb movements evoke a complex spatial pattern of neural activity. It may be that stimulation patterns that evoke neural activity more nearly like that of limb movements will elicit naturalistic sensations and reduce the training time required to use proprioceptive interfaces. This dissertation presents my work to develop such biomimetic stimulation patterns. By quantifying the spatiotemporal pattern of neural activity evoked by ICMS in Chapter 2, I show that stimulation through many electrodes with small amplitudes will be needed to recreate the complex spatial pattern of activity evoked by limb movements. In Chapter 3, I show that multi-electrode ICMS (mICMS) can replicate the rapid feedback provided by natural proprioception, something that single electrode stimulation cannot do. By modeling the evoked sensation with an artificially generated cortical map, I find that mICMS can produce effects as large as normal limb movements and in predictable directions in Chapter 4. Together, these results suggest that mICMS will be necessary to provide proprioceptive feedback in future afferent interfaces.

Creator