Open dataset: lumbosacral spinal cord task fMRI now on OpenNeuro

15 Sep 2026 . category: Research . Comments
#neuroimaging #fMRI #open science

I am happy to share that the dataset underlying our Imaging Neuroscience study on lumbosacral spinal cord task fMRI is now openly available on OpenNeuro:

ds008803: Towards personalized mapping through lumbosacral spinal cord task fMRI

DOI: 10.18112/openneuro.ds008803.v1.0.0

Corresponding to: Hernandez-Charpak et al., Imaging Neuroscience, Vol. 3 (2025)

---

What’s in the dataset

The release covers the full acquisition from the study: anatomical, functional, timings, and physiological (cardiac, respiration) data for all 12 healthy participants across four sessions each:

• High-resolution anatomical scans: T2-weighted structural volumes of the lumbosacral spinal cord, including a reduced field-of-view ZOOMit T2w sequence for improved local resolution.
• Task-based functional MRI, spanning 14 conditions targeting the extensor and flexor muscles of the right ankle, knee, and hip:
  • Ankle (4): active extension, active flexion, passive extension, passive flexion
  • Knee (2): active extension, passive flexion
  • Hip (2): active flexion, passive extension
  • Tendon vibration (6): extension and flexion vibration at the ankle, knee, and hip, isolating proprioceptive afferent input without active or passive joint movement
• Event timing files for every run, marking stimulus/movement onsets and durations for first-level analysis.
• Physiological recordings (cardiac and respiratory traces) acquired alongside each functional run, enabling retrospective physiological noise correction (e.g. RETROICOR) — a particularly important step for spinal cord fMRI given the region’s sensitivity to cardiac-related pulsatility and respiration.

Sharing the raw data alongside the timing and physiological files means the full pipeline — from preprocessing through activation mapping — can be reproduced or re-purposed, whether to benchmark new spinal cord fMRI analysis methods, pool with other lumbosacral datasets, or revisit the active/passive/vibration contrasts with alternative models.

If you use this dataset, please cite both the OpenNeuro release and the original article linked above.

---