The Neurophysiological Representation of Imagined Somatosensory Percepts in Human Cortex

Intracortical microstimulation (ICMS) in human primary somatosensory cortex (S1) has been used to successfully evoke naturalistic sensations. However, the neurophysiological mechanisms underlying the evoked sensations remain unknown. To understand how specific stimulation parameters elicit certain sensations we must first understand the representation of those sensations in the brain. In this study we record from intracortical microelectrode arrays implanted in S1, premotor cortex, and posterior parietal cortex of a male human participant performing a somatosensory imagery task. The sensations imagined were those previously elicited by ICMS of S1, in the same array of the same participant. In both spike and local field potential recordings, features of the neural signal can be used to classify different imagined sensations. These features are shown to be stable over time. The sensorimotor cortices only encode the imagined sensation during the imagery task, while posterior parietal cortex encodes the sensations starting with cue presentation. These findings demonstrate that different aspects of the sensory experience can be individually decoded from intracortically recorded human neural signals across the cortical sensory network. Activity underlying these unique sensory representations may inform the stimulation parameters for precisely eliciting specific sensations via ICMS in future work.<b>SIGNIFICANCE STATEMENT</b> Electrical stimulation of human cortex is increasingly more common for providing feedback in neural devices. Understanding the relationship between naturally evoked and artificially evoked neurophysiology for the same sensations will be important in advancing such devices. Here, we investigate the neural activity in human primary somatosensory, premotor, and parietal cortices during somatosensory imagery. The sensations imagined were those previously elicited during intracortical microstimulation (ICMS) of the same somatosensory electrode array. We elucidate the neural features during somatosensory imagery that significantly encode different aspects of individual sensations and demonstrate feature stability over almost a year. The correspondence between neurophysiology elicited with or without stimulation for the same sensations will inform methods to deliver more precise feedback through stimulation in the future.