| Part 1 Alterations of resting-state neural networks in patients with migraine in MEG sensor levelPurposeAlthough alterations in resting-state neural network have been previously reported in migraine using functional magnetic resonance imaging (fMRI), whether this atypical neural network is frequency dependent remains unknown. The aim of this study was to investigate the alterations of the functional connectivity of neural networks and their frequency specificity in sensor level in patients with migraine as compared with healthy controls by using magnetoencephalography (MEG) and concepts from graph theory.MethodsResting-state data from 23 patients with migraine and 23 age- and gender-matched healthy controls were collected in the sensor level using MEG. Neural networks from low (0.1-1 Hz) to high (80-250 Hz) frequency ranges were analyzed with functional connectivity patterns and quantified with graph theory. Functional connectivity networks from low (0.1-1 Hz) to high (80-250 Hz) frequency ranges were analyzed with functional connectivity patterns to determine whether there are differences between the migraine group and the healthy control group and further determine the abnormal functional connectivity brain regions; and neural network parameters from low (0.1-1 Hz) to high (80-250 Hz) frequency ranges were quantified with graph theory to determine whether there are differences between the migraine group and the healthy control group and further determine the correlation between the abnormal network parameters and the migraine clinical characteristics.ResultsThe topographic patterns of neural networks showed that the patients with migraine had significantly increased functional connectivity in slow wave (0.1-1 Hz) band in the frontal area as compared with controls. Compared with the patients with migraine without aura (MwoA), the patients with migraine with aura (MwA) had significantly increased functional connectivity in theta (4-8 Hz) band in the occipital area. Graph theory analysis revealed that the patients with migraine had significantly increased connection strength in the slow wave (0.1-1 Hz) band, increased path length in the theta (4-8 Hz) and ripple (80-250 Hz) bands, and increased clustering coefficient in the slow wave (0.1-1 Hz) and theta (4-8 Hz) bands.ConclusionsResults indicate that functional connectivity of neural network in migraine is significantly impaired in both low- and high-frequency ranges. The alteration of neural network may imply that migraine is associated with functional brain reorganization.Part 2 Alterations of resting-state neural networks in patients with migraine in MEG source levelPurposeThis study aimed to characterise the alterations of functional brain networks in patients with migraine through MEG source level in low- and high-frequency ranges.MethodsResting-state MEG data from 20 patients with migraine and 20 age- and gender-matched healthy controls were collected at the source level through functional connectivity analysis. MEG data were recorded at a sampling rate of 6000 Hz for 120 seconds. Neural network parameters from low (0.1-1 Hz) to high (80-250 Hz) frequency ranges were analyzed with the graph theory to determine whether there are differences between the migraine group and the healthy control group, and further determine the correlations between these abnormalities and the migraine clinical characteristics.ResultsGraph theory analysis revealed that compared with the controls, patients with migraine showed a dysfunctional network organization. The abnormal networks in patients with migraine were characterised by increased global connectivity strength in the 8-12Hz,30-80 Hz and 80-250Hz bands, as well as increased nodal connectivity strength in the orbitofrontal cortex, thalamus and some regions of the default mode network, including the prefrontal cortex, inferior parietal cortex and posterior cingulate cortex in the 12-30Hz and 30-80Hz bands; increased global degree in the 8-12Hz,12-30 Hz and 30-80Hz bands, as well as increased nodal degree in brain regions related to pain-processing in the 30-80Hz band, including the orbitofrontal cortex, prefrontal cortex, supplementary motor area and thalamus; increased path length in the 0.1-1Hz and 4-8Hz bands; and increased clustering coefficient in the 8-12Hz,12-30Hz and 30-80Hz bands. Furthermore, the abnormal functional network was positively correlated with number of years with migraine.ConclusionsThese results, for the first time, indicate that inherent brain networks are significantly impaired in low- and high-frequency ranges in patients with migraine and may contribute to the clinical manifestations of this disorder. |
PDF Full Text Request