Advancing spine connectomics and neural integration through machine learning and neuroengineering: a narrative review

Spinal connectomics is increasingly shifting understanding of the spinal cord from a simple reflex relay toward an active system that contributes to sensorimotor integration and adaptive motor control. This narrative review summarizes recent advances in the study of spinal circuitry and examines how these networks may contribute to flexible, context-dependent motor behavior.

We reviewed experimental and computational studies focusing on high-density electrophysiology, advanced imaging, circuit mapping, and computational modeling, with an emphasis on recent and landmark studies.

Our review suggests that spinal circuits may implement principles consistent with predictive coding, Bayesian integration, and adaptive gain control, though much of the direct mechanistic evidence for these computations originates in cortical and psychophysics literature; spinal-specific empirical validation remains an active research frontier. High-density recording and imaging techniques permit laminar-specific analysis of spinal activity, while computational models link circuit organization to function and plasticity. These advances are beginning to inform the development of closed-loop neuromodulation, targeted rehabilitation strategies, and brain-machine interface approaches aimed at restoring movement and sensory feedback following spinal cord injury.

Together, these findings are consistent with the emerging view of the spinal cord as a dynamic computational system rather than a passive relay. Integrating connectomic data with computational modeling and neuromodulation provides a framework for understanding spinal function and developing more precise therapeutic interventions. Continued progress in neural interface technologies and data-driven modeling has the potential to further advance spinal systems neuroscience and, over the coming years, to improve the treatment of neurological disorders.