It is demonstrated that a single ultrathin circumferential intradural electrode array can decode motor intent, classify sensory inputs, and discriminate visceral afferent signals across rodent and porcine models.
Abstract
Spinal cord injury affects over 2.5 million people worldwide, yet current neuroprosthetic strategies remain fragmented, addressing motor, sensory, or autonomic function in isolation. Here we show that a single ultrathin circumferential electrode array, conforming to the spinal cord without penetrating neural tissue, can simultaneously decode motor intent, classify sensory inputs, and discriminate visceral sensory inputs. In freely moving rats during short-term implantation (up to three days), deep learning decoders achieved robust motor intent decoding (R² = 0.97) by exploiting low-frequency spinal oscillations aligned with central pattern generator rhythms. The same interface classified eight sensory modalities with 94.4% accuracy. In acutely anaesthetized pigs, cross-species validation confirmed translational scalability and reliably distinguished visceral sensory inputs. Uniquely, the two-row electrode configuration resolved directional propagation within spinal tracts while electrode-dense one-row devices enabled high-precision intraspinal source localization. By consolidating motor, sensory, and visceral afferent decoding within a single conformal interface, this approach positions the spinal cord as a target for multifunctional neuroprosthetic interfacing, offering a path toward integrated restoration of physiological function after neurological injury. Spinal cord injury disrupts motor, sensory, and autonomic functions. Here, the authors demonstrate that a single ultrathin circumferential intradural electrode array can decode motor intent, classify sensory inputs, and discriminate visceral afferent signals across rodent and porcine models.
This work demonstrates for the first time in humans that clinical-grade lumbosacral epidural paddle arrays capture sufficient fine-scale spatiotemporal structure to decode these overlapping inputs from highly overlapping, volume-conducted epidural fields.
Alexander G. Steele, M. Candela, Gracie Hufft et al.· Research Square· 0 citations
Recording single-unit neural activity in the spinal cord in freely moving rodents is crucial for understanding spinal network dynamics but remains challenging due to specific biomechanical constraints. So far, the vast majority of studies have been conducted in anesthetized or restrained animals, limiting the correlation of neuronal activity with naturalistic behaviours. Here we introduce the Spinotrode, a vertebral implant able to stably record signals (over several weeks) at the single-cell level in the bilateral dorsal horns of adult mice. It is engineered to minimize postural constraints and interrogate spinal activity during sensory stimulation and motor behaviours. No functional impairment or tissue damage was apparent. Spinotrode recordings allowed identification of distinct functional types of neurons associated with paw withdrawal, revealed dorsal horn activity during locomotor behaviour distinct from proprioception and touch, and uncovered contralateral sensory activation upon nociceptive reflex responses, which forces reassessment of data obtained in anesthetized animals.
J. Viellard, L. Brochoire, Michelle Janusz et al.· bioRxiv· 0 citations
AbstractBackground. Cervical spinal cord injury disrupts descending motor commands, sensory information, and autonomic regulation, producing tetraplegia incompletely reversible with rehabilitation. Brain–spine interfaces seek to re-establish communication across the lesion by decoding cortical intention and delivering patterned spinal stimulation, while bidirectional systems restore somatosensory feedback.Materials and methods. A structured narrative review integrated trials, prospective cohorts, first-in-human studies, neurophysiological investigations, regulatory documents, and reports published through July 2026. Evidence was organized by biological plausibility, neural-signal acquisition, decoding performance, spinal-target selectivity, sensorimotor integration, safety, durability, rehabilitation dependence, and translational readiness. A prospective multicenter comparative protocol was developed for adults with chronic cervical injury and neurological stability.Results. The synthesis addresses cortical decoding, epidural and transcutaneous spinal neuromodulation, upper-limb and locomotor restoration, artificial sensory feedback, autonomic effects, and activity-dependent neuroplasticity. Implanted brain–spine systems have enabled intention- driven standing and walking in an individual with chronic tetraplegia, whereas non-invasive cervical stimulation combined with task practice has improved hand strength and sensation in cohorts. A double neural bypass reported in 2026 integrated intracortical decoding, patterned spinal stimulation, and cortical sensory stimulation, producing immediate assistance together with persistent motor and sensory gains in one participant. The proposed primary endpoint is functional independence measured by SCIM III and task-specific upper-extremity performance without device-related morbidity. Secondary endpoints include GRASSP, CUE-T, gait capacity, decoding latency, sensory localization, autonomic stability, quality of life, caregiver burden, durability, and cost-effectiveness.Conclusion. Brain–spine interfaces are evolving from proof-of-concept bypasses toward restorative neuroprosthetic systems. Multicenter validation, standardized endpoints, transparent algorithms, and long-term surveillance remain essential.Keywords: cervical spinal cord injury, tetraplegia, brain–spine interface, brain–computer interface, epidural spinal cord stimulation, transcutaneous stimulation, neural bypass, sensory feedback, neuroplasticity.
Spinal interneurons are central to sensorimotor integration. Even in the absence of overt sensory inputs or motor output, spinal interneurons remain active, yet their functional organization across spatial and temporal scales remains poorly understood. We characterized the discharge properties and functional connectivity of 2,847 lumbar spinal interneurons recorded in vivo from quiescent decerebrate cats (3 males, 5 females) using microelectrode arrays inserted into rostral (L3-4) and caudal (L5-7) spinal segments at superficial (0-1500 µm) and deep (1500-3000 µm) depths. Superficial neurons generally discharged at higher and less variable rates than deeper neurons, forming a consistent depth-dependent gradient at both segments. Functional connectivity was sparse yet structured, with distinct depth profiles across timescales. Short-timescale correlations (±10 ms) occurred in ∼2% of neuron pairs, which were predominantly excitatory and prominent caudally at ∼1500 µm depth. Longer-timescale correlations (∼400 ms) were similarly rare but showed equal proportions of excitation and inhibitory and were more pronounced at deeper depths. Graph-theoretic network analysis showed that timescale-specific correlations are architecturally distinct and non-randomly organized. Short-timescale networks exhibited depth-specific hubs near 1500 µm and lower clustering, whereas longer-timescale networks showed more clustering. At the ensemble level, neurons shared low-frequency oscillatory drive, particularly in rostral recordings. Separately, a subset of ensembles with dense short-timescale connectivity were consistently observed near 1500 µm in the caudal array, linking slow and fast interactions. Together, these findings show that spinal interneuron circuitry exhibits spatially and temporally structured organization even at rest, establishing a critical baseline for studies of spinal injury and disease.Significance Statement Spinal interneurons are essential for sensorimotor processing, yet their functional organization remains poorly defined. By recording the resting discharge behavior of 2,847 lumbar interneurons from in vivo cats, we show that spinal interneuron networks are sparse but systematically organized both spatially and temporally. Neurons exhibited depth-dependent discharge properties and formed distinct short- and longer-timescale functional networks with different architectures and hub locations. Shared low-frequency activity links distributed ensembles, particularly in rostral segments, while highly-connected ensembles emerge from specific layers in caudal segments. These findings reveal a multilayered, temporally distinct organization of spinal interneuron circuitry and provide a foundational reference for understanding how spinal networks adapt following injury or disease.
Mohamed H. Mousa, M. Zaback, Michel A. Lemay et al.· Journal of Neuroscience· 0 citations
Voluntary human movement emerges from dynamic interactions between the brain, spinal cord, and sensory afferents. Spinal cord function has proven particularly difficult to study non-invasively in humans due to its deep anatomical location and narrow diameter. As a result, sensorimotor interactions are often studied using cortico-muscular coupling, which captures interactions between brain and muscle but cannot reveal how these signals propagate through the spinal cord. Here, we use concurrent brain and spinal cord imaging with optically pumped magnetometers (OPMs) and electromyography (EMG) to test whether endogenous synchronization within the human sensorimotor system can be detected and characterized non-invasively across cortical, spinal, and muscular levels. Participants (n = 9) performed a sustained isometric hand contraction while we recorded cortical and spinal cord magnetic fields using OPMs and hand muscle activity using EMG, and we compared rhythmic activity propagating through the cortico-spinal-muscular loop with the state at rest. We show that the spinal contribution to cortico-muscular coupling manifests as coherent 10-35 Hz activity linking the contralateral sensorimotor cortex, cervical spinal cord, and muscle. This synchronization forms a distributed, bidirectionally interacting network with physiologically plausible temporal delays. The spatial organization of this activity accords with known sensorimotor anatomy, localizing to cervical segments appropriate for upper-limb control. Our results provide the first direct, non-invasive magnetic field evidence in humans of coherent brain and spinal cord oscillations and demonstrate the feasibility of concurrent, spatiotemporally resolved imaging of the central nervous system to open new avenues for studying human sensorimotor physiology.
Meaghan E. Spedden, Maike Schmidt, George C. O’Neill et al.· Current Biology· 1 citation
The spinal cord serves as a crucial relay for motor commands, yet the role of its local circuitry in sensorimotor integration remains poorly understood. Most non-invasive cortical stimulation studies, rely on electrophysiological readouts or inferred spinal function from corticospinal anatomy, leaving the downstream impact of cortical stimulation on spinal circuitry largely uncharted in vivo. Advances in spinal cord functional MRI (SC-fMRI) now enable spatially resolved imaging of segmental gray and white matter and their interactions with descending cortical inputs. Here, we introduce a multimodal framework that combines single-pulse transcranial magnetic stimulation (TMS) of the primary motor cortex with SC-fMRI to probe TMS-evoked spinal activity in humans. Using graded TMS intensities, we examined blood oxygenation level-dependent (BOLD) responses in the cervical spinal cord and asked how spinal activation depends on effective engagement of the descending motor system. Our findings reveal robust, intensity-dependent spinal BOLD responses aligned with descending pathways, with activation concentrated in expected territories such as the lateral corticospinal tract and ventral horn at segments innervating the stimulated hand muscle. By linking peripheral output to segment- and pathway-resolved spinal signals, these results demonstrate that concurrent TMS–SCfMRI can capture, in vivo, how cortical drive is expressed within human spinal circuitry and provide a new framework to measure spinal contributions to sensorimotor control with spatial specificity beyond traditional peripheral readouts. Graphical Abstract
E. Sareen, Rebecca Jones, E. Raffin et al.· bioRxiv· 0 citations