Mapping the network underpinnings of central poststroke pain and analgesic neuromodulation

Abstract

Central poststroke pain (CPSP) is a debilitating and often treatment-refractory condition that affects numerous stroke patients. The location of lesions most likely to cause pain and the identity of the functional brain networks that they impinge upon remain incompletely understood. We aimed to (1) elucidate which lesion locations are most frequently accompanied by pain; (2) explore CPSP-associated functional networks; and (3) examine how neuromodulation interacts with these networks. This multisite study investigated 17 CPSP patients who received deep brain stimulation (DBS; n = 12) or motor cortex stimulation (MCS; n = 5). Pain-causing lesions were manually segmented and normalized to standard space. To identify areas linked to high risk of pain, the locations of CPSP lesions and 220 control lesions were compared using voxelwise odds ratio mapping. The functional connectivity of pain-causing lesions was obtained using a large (n = 1000) normative resting-state functional MRI connectome and compared to that of control lesions and therapeutic DBS activation volumes. Brain regions most associated with CPSP risk (highest value = 63 times) were located along the ascending somatosensory pathways. These areas and the majority of individual CPSP lesions were functionally connected to anterior/middle cingulate cortex, insula, thalamus, and inferior parietal lobule (P Bonferroni < 0.05). The extent of connectivity to the thalamus, inferior parietal lobule, and precuneus also differed between CPSP and control lesions (P Bonferroni < 0.05). Posterior insula and thalamus shared connectivity with both CPSP lesions and pain-alleviating DBS activation volumes (P Bonferroni < 0.05). These findings further clarify the topography and functional connectivity of pain-causing brain lesions, and provide new insights into the network-level mechanism of CPSP neuromodulation.

Untapped Neuroimaging Tools for Neuro-Oncology: Connectomics and Spatial Transcriptomics

“…The functional (i.e., rsfMRI) and structural (i.e., dMRI tractography) connectomes can be studied using normative, atlas-based connectome data assembled using high-quality acquisitions from a large number of subjects [5,26,33]. The use of these normative connectomes allows the investigation of brain-wide circuits implicated in clinical symptoms in the absence of patient-specific rsfMRI or dMRI acquisitions [24,29,[34][35][36] (Figure 1). For example, if tumors are found to cause a specific cognitive impairment, one could use the normative functional connectome to map the key implicated regions and optimize patient counseling and surgical timing of subsequent tumors with similar characteristics.…”

A Network-Based Approach to Glioma Surgery: Insights from Functional Neurosurgery

“…Such an approach has been used to map the underlying connectivity of DBS-associated flashback phenomena following forniceal DBS, panic attacks induced by inferior thalamic peduncle DBS, and seizures following subcallosal cingulate DBS for refractory anorexia nervosa [ 30 , 31 , 32 ]. Similarly, connectivity-derived models can guide effective DBS for less well-understood pathologies, such as central post-stroke pain and neuropsychiatric indications, including obsessive-compulsive disorder [ 8 , 33 ]. Similar analyses were undertaken to investigate the perturbed networks leading to post-operative morbidity in glioma surgery.…”

An exploratory study into the influence of laterality and location of hippocampal sclerosis on seizure prognosis and global cortical thinning

“…Greater seizure severity may be related to greater brain-wide connectivity of the seizure focus. Because the patient's included in our study did not have native functional imaging acquisitions, patterns of functional connectivity associated with seizure focus laterality were explored using an established functional connectivity mapping method that has been leveraged in numerous previous studies [19][20][21][22][23] . This method uses a large-scale, high-quality normative resting-state fMRI (rsfMRI) dataset constructed from 1000 healthy subjects (http://neuro infor matic s.harva rd.edu/gsp).…”

“…This method uses a large-scale, high-quality normative resting-state fMRI (rsfMRI) dataset constructed from 1000 healthy subjects (http://neuro infor matic s.harva rd.edu/gsp). Detailed information about the preprocessing and aggregation of this dataset have been previously described [18][19][20][21][22][23][24] . Briefly, each healthy subject in the normative dataset was scanned once or twice (1.7 times per subject on average) with a 6.2 min-long echo-planar imaging sequence (124 time points; 3 × 3 × 3 voxel size, TR 3000 ms, TE 30 ms, flip angle 85°) in order to acquire rsfMRI data.…”