| 제목 |
Treatment Targets for Central Post-stroke Pain Based on the Network Effects of Brain Lesions |
| 소속 |
Yongin Severance Hospital, Department of Rehabilitation Medicine1, Yonsei University College of Medicine, Department and Research Institute of Rehabilitation Medicine2, Brigham and Women’s Hospital, Department of Neurology3 |
| 저자 |
Na Young Kim1*, Joseph J. Taylor3, David Borsook3, Michael D. Fox3†, Yong Wook Kim2† |
| 사사 |
This work was supported by a National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. 2019R1C1C1006980) and by a faculty research grant of Yonsei University College of Medicine (6-2019-0100). None of the funding agencies had a role in the design and conduct of the study, in the collection, management, analysis and interpretation of the data, in the preparation, review or approval of the manuscript, nor in the decision to submit the manuscript for publication. |
Objective : Neuropathic pain is prevalent, debilitating, and in need of new treatments that avoid the side effects and addiction potential of medications. Brain stimulation could represent such a treatment, but development has been limited by trial-and-error searches for therapeutic targets. Central post-stroke pain (CPSP) can provide unique insight into the causal neuroanatomy of neuropathic pain in the human brain, and thus may help to identify therapeutic targets. We hypothesized that lesions causing CPSP have remote effects on functionally connected brain regions, and that these regions will align with the reported efficacy of brain stimulation.
Method : Two independent datasets were collated: 1) subcortical lesions from published case reports (N=63, 23 of which were associated with CPSP) and, 2) thalamic lesions with metabolic imaging using 18F- fluorodeoxyglucose PET-CT (N=43, 20 of which were associated with CPSP). Functional connectivity between each lesion location and the rest of the brain was assessed using a normative connectome (resting state functional MRI, N = 1000) and connections specific to CPSP were identified. Metabolic changes specific to CPSP were also identified and related to differences in lesion connectivity.
Results : We found that lesion locations associated with CPSP showed a specific pattern of brain connectivity (FWE p < 0.05) that was highly consistent across our two independent datasets (spatial r = 0.82, p < 0.0001). Metabolic differences between CPSP and control lesions were also identified (FWE p < 0.05) and metabolism was correlated with connectivity to the lesion location (r = -0.57, p < 0.0001). Ipsilesional primary motor cortex was the only published neuromodulation target significantly associated with CPSP across all analyses, including dataset one connectivity (p < 0.000005), dataset two connectivity (p < 0.05), and dataset two glucose metabolism (p < 0.005). These findings corroborate empirical clinical data suggesting that motor cortex is the most effective neuromodulation target for pain tested thus far. A voxelwise search for alternative therapeutic targets identified somatosensory cortex and contralesional visual cortex, a therapeutic target used for migraine pain.
Conclusion : Our results support the hypothesis that CPSP lesions have remote effects on functionally connected brain regions, and that the location of these remote effects can identify therapeutic targets for neuromodulation. |
|
Figure 1. Methods for identifying brain networks for pain based on focal brain lesions. (A) Two independent datasets of lesion locations (red) associated with central post stroke pain (CPSP) or not associated with CPSP (control) were analyzed (row 1). Functional connectivity between each lesion location and the rest of the brain was computed using a publicly available connectome dataset of 1000 healthy controls (row 2). Connections specific to CPSP versus control lesions were identified (row 3) and statistical significance was assessed through permutation (PTFCE FWE-corrected < 0.05, row 4). Positive correlations with CPSP lesions are shown in hot colors, and negative correlations (anticorrelations) are shown in cool colors. (B) In dataset 2, each lesion patient (row 1) underwent metabolic PET imaging approximately 2 months after their stroke (row 2). Differences in glucose metabolism between CPSP and controls were identified (row 3) and statistical significance was assessed using permutation (PTFCE FWE-corrected < 0.05, row 4). Hypermetabolism in the CPSP group is shown in hot colors, and hypometabolism is shown in cool colors.
Figure 3. Differences in metabolism after stroke were related to differences in lesion connectivity. Brain regions exhibiting significant decreases in glucose metabolism post-stroke (A) overlapped the brain regions showing significant differences in lesion connectivity (B). A conjunction analysis (C) identified voxels that were significant in both the metabolism and connectivity analyses, with one cluster in the ipsilesional sensorimotor cortex and another in the contralesional occipital cortex. (D) Connectivity from each lesion location (n=43) to the ipsilesional sensorimotor cortex (upper row) and the contralesional occipital cortex (bottom row) was significantly correlated with post-lesion metabolism. Abbreviations: I = ipsilesional; C = contralesional.
Figure 4. Relationship to brain stimulation targets for CPSP. (A) Repetitive transcranial magnetic stimulation targets (upper row) and deep brain stimulation targets (lower row) previously used for CPSP. Primary motor cortex (M1) is highlighted as the target with the most consistent evidence of efficacy. The electric field model for a figure-of-8 TMS coil targeting M1 (B) overlaps our lesion-based pain network (C). The electric field model for a figure-of-8 TMS coil positioned more medial and posterior (D) better overlaps our lesion-based pain network (E). In panels B and D the color represents the absolute magnitude of electric field. In panels C and E the electric field is shown thresholded at 1mV/m (yellow) along with our lesion-based pain network (purple). Abbreviations: ACC = anterior cingulate cortex, DLPFC = dorsolateral prefrontal cortex, dpreM = dorsal premotor cortex, M1 = primary motor cortex, PAG = periaqueductal gray, PCC = posterior parietal cortex, S1 = primary somatosensory cortex, S2 = secondary somatosensory cortex, SMA = supplementary motor area.
|
