Specific brain morphometric changes in spinal cord injury with and without neuropathic pain

Abstract

Why only certain patients develop debilitating pain after spinal chord injury and whether structural brain changes are implicated remain unknown. The aim of this study was to determine if patients with chronic, neuropathic below-level pain have specific cerebral changes compared to those who remain pain-free. Voxel-based morphometry of high resolution, T1-weighted images was performed on three subject groups comprising patients with pain (SCI-P, n = 18), patients without pain (SCI-N, n = 12) and age- and sex-matched controls (n = 18). The SCI-P group was first compared directly with the SCI-N group and then subsequently with controls. Overall, grey and white matter changes dependent on the presence of pain were revealed. Significant changes were found within the somatosensory cortex and also in corticospinal tracts and visual-processing areas. When the SCI-P group was directly compared with the SCI-N group, reduced grey matter volume was found in the deafferented leg area of the somatosensory cortex bilaterally. This region negatively correlated with pain intensity. Relative to controls, grey matter in this paracentral primary sensory cortex was decreased in SCI-P but conversely increased in SCI-N. When compared with controls, discrepant corticospinal tract white matter reductions were found in SCI-P and in SCI-N. In the visual cortex, SCI-N showed increased grey matter, whilst the SCI-N showed reduced white matter. In conclusion, structural changes in SCI are related to the presence and degree of below-level pain and involve but are not limited to the sensorimotor cortices. Pain-related structural plasticity may hold clinical implications for the prevention and management of refractory neuropathic pain.

Keywords: 10.020: Structural MRI; 40.100: Systems Pain; 40.130: Systems Somatosensory; 50.140: Pain syndromes Other; 50.160: Spinal cord; 50.180: Trauma; ASIA, American Spinal Injury Association; BAI, Beck Anxiety Inventory; BDI, Beck Depression Inventory; GM, grey matter; SCI, spinal cord injury; SCI-N, spinal cord injury without pain; SCI-P, spinal cord injury with pain; VBM, voxel-based morphometry; WM, white matter.

Fig 1

Direct focal volume comparison of SCI-N with SCI-P: S1 leg area. Averaged S1 volumes were extracted from the supra-threshold cluster formed at a threshold of p < 0.001 from the whole brain SCI-P versus SCI-N contrast. Divergent changes in S1 (grey matter) were found in SCI-P and SCI-N groups compared to healthy controls. The SCI-P subgroup had significantly decreased grey matter in S1 leg area bilaterally (A–C) compared with the SCI-N subgroup (E*). Focal grey matter volumes for this S1 region correlated negatively to pain scores in the SCI-P group (D). Focal grey matter volumes for the illustrated bilateral S1 cluster varied across the three study groups (E). In whole-brain analyses, relative to controls, there was no significant increase in S1 grey matter in the SCI-N group (E**). Notably, however, from the a priori S1 region of interest with its increased sensitivity, a statistically significant increase was found in the left S1 leg area in the SCI-N group (p = 0.048) and also an increase in the right S1, though this did not reach significance (p = 0.101). Image shows the cluster formed at a threshold of p < 0.001 (uncorrected).

Fig 2

SCI-N and SCI-P groups compared with controls: averaged ROI volumes were extracted for the supra-threshold clusters identified from whole brain contrasts (p < 0.001 uncorrected) in Table 2. Changes were found in (i) the sensorimotor tract and (ii) visual processing areas. Areas associated with the corticospinal tract showed reduced volume (white matter). Compared with controls, the SCI-P group showed reduced white matter (WM) volume in the pyramids and also in two further regions: deep to S2 (A), and in the area of the posterior corona radiata (B). The SCI-N group showed white matter changes also in the pyramids but not in any other locations (C) (ii). Mixed visual stream changes were found (both grey matter and white matter). The SCI-P group showed a decrease of white matter medially in the cuneus (D). The SCI-N subgroup showed increased grey matter in the adjacent lateral cuneus extending into the precuneus (E). The anatomical relationship between visual cortex clusters is shown in (F). When tissue volumes were extracted for each region and compared across the three groups, variation between groups was revealed (G). Note that error bars indicate 95% confidence intervals for the mean and are based on raw extracted tissue volumes. Asterisks represent t-tests that showed significant group differences of averaged extracted supra-threshold cluster volumes (p < 0.05). All images are displayed on an averaged template based on all study participants, thresholded at p < 0.001 uncorrected.