BIOMIMETIC SPATIOTEMPORAL MICROSTIMULATION OF THE HUMAN SOMATOSENSORY CORTEX FOR RESTORING TOUCH IN BRAIN-CONTROLLED BIONIC HANDS

Restoring touch in people with sensory loss is crucial for dexterous hand control. In people with paralysis, intracortical microstimulation (ICMS) of the somatosensory cortex (S1) can evoke tactile percepts. While we have excellent control over the location or intensity of the percept by changing the electrode or stimulus parameters, the quality of the sensations is less informative than natural touch.
Efforts to sensitize bionic hands by nerves or S1 stimulation have shown that sensory feedback that mimics natural tactile signals (biomimetic) better supports objects interactions than non-biomimetic feedback, even reducing interference on motor decoding in closed-loop BCI. Despite these successes, biomimetic feedback has yet to produce fully-realistic sensations comparable to natural touch.
To better understand how to encode more natural and informative sensations, a deeper understanding of how S1 encodes tactile stimuli is required. In this study, we recorded neural activity from S1 and M1 in a participant with residual sensation on his contralateral hand, chronically implanted with intracortical microelectrode arrays, while mechanical indentations with different depths and rates were applied to his digits.
First, we characterized cortical responses, identifying a precise spatiotemporal structure. Neural activity modulated according to indentation parameters, indicating S1's role in encoding tactile stimuli. Then, we designed and implemented ICMS patterns resembling the recorded patterns using multiple encoding strategies.
In early testing, we found that ICMS strategies that mimic spatiotemporal S1 dynamics activity were perceived as more realistic than conventional approaches. These findings support biologically-inspired encoding patterns to improve the naturalness and informativeness of artificial touch.