An automated algorithm for cardiovascular spinal cord epidural stimulation (CV-scES) achieved a mean correlation coefficient of 0.94 with researcher-controlled stimulation paradigms across eight individuals with spinal cord injury (SCI), successfully maintaining systolic blood pressure within a 110–120 mmHg target range. The sigmoid-based real-time implementation — which dynamically adjusts response speed based on how far blood pressure deviates from target — ultimately outperformed both the original algorithm and manual researcher control on several blood pressure stability metrics, though postural tilt recovery remained slower than expert-guided intervention.
Orthostatic hypotension affects roughly 40–80% of individuals with cervical or high thoracic SCI and represents a serious daily health burden, increasing fall risk, cognitive impairment, and reducing rehabilitation capacity. Epidural stimulation has shown genuine promise for autonomic cardiovascular control, but its clinical adoption has been bottlenecked by the need for constant expert oversight and highly individualized parameter tuning — exactly the gap this algorithm targets. The shift toward closed-loop, automated neuromodulation mirrors broader trends in deep brain stimulation and cardiac rhythm management.
Limitations are substantial: the real-time validation involved only a single participant, the offline dataset covered just eight individuals, and the study remains a preprint not yet peer-reviewed — findings and conclusions may change substantially after expert scrutiny. The lag in postural recovery also signals meaningful clinical gaps before unsupervised home use. Still, the approach is directionally significant, potentially democratizing access to a therapy currently confined to specialized research centers.
