To investigate cerebral and cerebellar structural changes and cervical spinal cord cross-sectional area atrophy after human spinal cord injury (SCI).
Central nervous system injuries at the spinal cord level cause tissue loss and altered neurological functions. Beyond tissue loss and scarring at the injury site, these injuries disrupt neuroanatomical networks, leading to structural changes far from the injury. While many studies have separately noted supraspinal and spinal atrophy after SCI, the degree of structural changes varies between studies and analytical methods. Moreover, persons with SCI and concomitant traumatic brain injury (TBI) have typically been excluded from analyses.
101 persons with spinal cord injury and 58 healthy controls from three sites underwent clinical and brain magnetic resonance imaging (MRI) on 3T scanners. Data was analyzed using FreeSurfer and CAT12 for brain structure, SUIT and CERES for cerebellar structure, and Spinal Cord Toolbox for spine structure. Region of interest (ROI) data from cervical spinal cord cross-sectional area, subcortical, and brainstem structures were analyzed and compared using R.
The SCI group exhibited thinning in multiple cortical areas (pFWE<0.05) with CAT12 and threshold-free cluster enhancement. No significant cortical thinning was observed in SCI group compared to controls using Freesurfer. Two clusters in the right posterior insula showed reduced gray matter VBM signal (pFWE<0.05). Both SCI and SCI-TBI diminished volume in the medulla oblongata and thalamus. No significant changes were found in the brainstem and cerebellum after re-analysis using SUIT and VBM. CERES volumetric segmentation and SUIT with VBM analysis revealed no significant differences (pFWE<0.05). Spinal cord atrophy in SCI group was evident at the C2/C3 level.
We corroborate earlier findings that spinal cord injuries result in supraspinal and supralesional structural alterations. TBI may further increase atrophy. However, the subtle changes in cortical gray matter might necessitate larger samples for reproducibility.