Neural Noise

Modeling strategies in non-invasive spinal stimulation

3 min read

  • Non-invasive spinal stimulation (NISS) is used for spinal dysfunctions (e.g. spasticity, chronic pain, hypotonia) when implants or drugs are not suitable.1
  • Trans-spinal direct current stimulation (tsDCS) delivers low-intensity DC (1–4 mA) via large electrodes over the vertebral column; modeling approaches for NISS/tsDCS are under development.1 1

Weekly enrichment (2026-07-20)

  • The 2026 Frontiers in Human Neuroscience review by Sofia Rita Fernandes frames NISS modeling as inherited from non-invasive brain stimulation pipelines, tracing tsDCS back to Cogiamanian et al. (2008), whose anodal spinal electrode produced a 25% reduction in the cervico-medullary P30 component of tibial-nerve somatosensory evoked potentials, an effect sustained ~20 min post-stimulation.2
  • tsDCS delivers low-intensity direct current of 1–4 mA through large 8–25 cm² electrodes over the vertebral column, inducing weak electric fields typically below 1 V/m whose main mechanism is persistent inward currents mediated by calcium channels.2
  • Realistic spinal models demand MRI resolution roughly an order of magnitude finer than head modeling because spinal white matter is only 6–12 mm in diameter; the ITIS Virtual Population models were segmented at 0.5 × 0.5 × 0.5 mm³ voxels, and early 1.5 T acquisitions of full-body datasets required about 6 h of scanning.2
  • Finite element models of the trunk and spine can reach 10⁷–10⁸ mesh elements; in the author’s prior COMSOL work a tsDCS simulation with 2.6 × 10⁷ degrees of freedom took ~150 min to solve on dual quad-core Xeon hardware with 48 GB RAM, and tetrahedral volume meshing alone took 3–6 h.2
  • No single software pipeline covers segmentation, meshing, and field simulation; the review tabulates Spinal Cord Toolbox and ITK-SNAP (open-source segmentation), Gmsh/MeshLab (meshing), and commercial COMSOL, Abaqus, SEMCAD X, and Sim4Life for electric-field solving, with SimNIBS/ROAST adaptable only via mesh-format conversion.2
  • Modeling suggests neuromodulatory tsDCS effects appear at spinal electric-field magnitudes above ~0.20 V/m, compared with the >0.15 V/m estimated in motor cortex for 1 mA tDCS, but the review stresses that effect direction depends on field-vector orientation relative to neural targets, not magnitude alone.2
  • Animal validation remains scarce: de Oliveira Pires et al. (2025) measured voltage gradients in anesthetized SOD1 mouse spinal cords during tsDCS and found experimental fields 2–4 fold smaller than MRI-based model predictions, attributed to conductivity assumptions and measurement uncertainty.2
  • Broader clinical reviews report that tsDCS combined with adjunctive therapies such as robotic gait training and neuromuscular electrical stimulation augments motor outcomes in stroke and spinal cord injury, while noting substantial heterogeneity in intensity, duration, and electrode montage that currently limits translation; recent double-blind RCTs pairing transcutaneous spinal stimulation with Lokomat robotic gait training in incomplete SCI reported gains in lower-extremity motor score and walking recovery.3 4

Footnotes

  1. https://www.frontiersin.org/articles/10.3389/fnhum.2026.1763470 ↩ ↩2 ↩3

  2. https://doi.org/10.3389/fnhum.2026.1763470 ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7

  3. https://journals.lww.com/bnam/fulltext/2025/01000/transcutaneous_spinal_direct_current_stimulation.2.aspx ↩

  4. https://link.springer.com/article/10.1186/s12984-025-01545-8 ↩

Graph View