High-Dimensional Information Transfer from the Spinal Cord to the Brain using Spinal Cord Stimulation

Advancements in high-channel count electrodes and decoding methods have enhanced the ability of neural interfaces to "read" from the brain. However, it is equally crucial to be able to "write" meaningful information to the brain. Spinal cord stimulation (SCS) offers a novel approach to target ascending sensory pathways that transmit information to the brain. Yet, the ability of SCS to encode high-dimensional information in supraspinal structures through temporal, spatial, and spatiotemporal patterns has not been demonstrated. Here, we introduce a method to send high-dimensional information to the brain using high-density SCS in anesthetized rats. Using high-density recordings in somatosensory cortex (S1) and ventral posterolateral thalamic nucleus (VPL), we quantified the mutual information (MI) between SCS parameters and neural responses. Neural activity during stimulation exhibited a clear and significant elevation in MI but declined during pre- and post-stimulation periods. Deep neural networks (DNNs) trained to decode temporal (5 frequencies), spatial (15 cathodes), and spatiotemporal (10 unique sequences and combinations) dimensions from the brain responses achieved significantly above-chance classification accuracy. These results demonstrate that supraspinal neural activity contained SCS stimulus-specific information that can be reliably decoded. This work positions SCS as a viable channel for high-dimensional information transfer to the brain.