1716
Phenotyping assay of neuropathic pain models using resting state functional connectivity MRI and Graph theoretical analysis1Laboratory Animal Research Department, Central Institute for Experimental Animals, Kawasaki, Japan, 2Keio University, Tokyo, Japan, 3Brain Science Institute, RIKEN, Saitama, Japan, 4Okinawa Institute of Science and Technology Graduate University (OIST), Japan
Synopsis
Resting state functional connectivity MRI was performed with neuropathic pain model mice.The functional network was constructed by temporal correlation analysis at the whole brain level based on the Allen brain atlas. Graph theoretical analysis was conducted to evaluate the feature of constructed networks. Compared with the intact model, degree and eigenvector centrality of neuropathic pain model showed a significant reduction in the primary somatosensory area. The clustering coefficient and local efficiency were significantly increased in the ACA. Significantly higher betweenness centrality was observed in the VPL. These results indicate that amount of information about connection to S1 was decreased. Neuropathic pain disrupt the pain matrix and the pain matrix that includes ACA and VPL may construct the complicated network. Integration of resting state functional connectivity MRI and graph theoretical analysis can evaluate the interactive complex networks of each region, not only existence or non-existence of activation region.
INTRODUCTION
Neuropathic pain is typically caused by traumatic injury of the nervous system. Neuropathic pain has historically been evaluated by behavioral analyses. Subjective measures of neuropathic pain can be influenced by several factors that involve both the subject and the observer. To overcome these problems, MRI may enable the objective and quantitative evaluation of pain. Recent research in brain imaging, which involves the integration of resting state functional connectivity MRI (rs-fc MRI) and graph theory, has revealed some fundamental aspects of brain-network organization in neurological disorders. rs-fc MRI is a novel approach that examines spontaneous brain function by using blood oxygen level-dependent contrast in the absence of a task. In this study, fMRI scans were conducted using mouse models of neuropathic pain. With the use of rs-fc MRI, we evaluated the properties of brain networks in previously reported areas in task fMRI.METHODS
This study was approved by the local Animal Experiment Committee and was conducted in accordance with the Guidelines for Conducting Animal Experiments of the Japanese Central Institute for Experimental Animals (approval number: 12014).
Twelve intact adult mice (C57BL/6, male, approximately 10 weeks old; CLEA Japan Inc., Tokyo, Japan) were used in this study, together with six adult mice for the peripheral nerve injury model, six adult mice as a control group.
fMRI was performed using a 7.0 tesla MRI system(Biospec; 70/16 Bruker BioSpin, Ettlingen, Germany) with a cryogenic surface coil (CryoProbe; Bruker BioSpin AG, Fällanden, Switzerland). rs-fc MRI was applied in the allodynia model mice before the operation and seven days after the operation. single shot GRE-EPI; TE 20 ms, TR 1000ms, flip angle 55°, number of averages 1, spatial resolution 200 × 200 × 500 (μm)3, and number of slices 16, scan repetition 600 times.
A functional connectivity analysis was performed using CONN1. A temporal band-pass filter was applied in the 0.009 Hz to 0.1 Hz range2. Five hundred seventy-six predefined regions based on connectional and architectonic subdivisions in the mouse brain atlas were used by combining atlases as provided by the Allen Institute3,4. The node coordinate was determined by calculating centroids of the volume which each region has. The brain networks were visualized with the BrainNet Viewer5.
A graph theory analysis based on a undirected, weighted, connection matrix was performed with the Brain Connectivity Toolbox6. Graph theory indicators were analyzed to determine the inter-subject standard error, and groups were compared by applying a Paired t-test.
RESULTS and DISCUSSION
A connection matrix was expressed as the correlation between architectonic subdivisions of each whole brain region (Figure 1), and visualized in terms of nodes and edges (Figure 2). The characteristics of the brain network were calculated by a graph theory approach (Figure 3). Degree and eigenvector centrality showed a significant reduction in the contralateral Primary somatosensory area of lower limbs (SSp-ll) and upper limbs (SSp-ul) of peripheral nerve injury model mice compared to those in pre-operative animals. The clustering coefficient and local efficiency were significantly increased in the Anterior Cinglate area (ACA). Significantly higher betweenness centrality was observed in the ventral postero-lateral nucleus of the thalamus (VPL). The degree reflects the number of significant functional connections of a node, Eigenvector centrality is a self-referential measure of centrality, Betweenness centrality is the fraction of all shortest paths in the network that contain a given node, and the clustering coefficient is the fraction of triangles around a node and the local efficiency is the global efficiency computed in the neighborhood of the node. These results indicate that the sensory information that reaches from the neospinothalamic tract and the medial lemniscus of the posterior column-medial lemniscus pathway to become somatosensory through the VPL is reduced, and a pain matrix that includes the ACA and VPL may form a complicated circuit network. Future studies will be needed to clarify the direction of the transfer and to validate in detail the indicators of graph theory.CONCLUSION
fMRI studies have previously been used to identify brain regions that correspond to tasks. Recently, graph theory has been applied to understanding functional connectivity in MRI, particularly with regard to the resting state or endogenous fluctuations.In this study, in Neuropathic pain model mice, the degree and eigenvector centrality were significantly decreased in the contralateral SSp-ll and SSp-ul, the clustering coefficient and local efficiency were significantly increased in the ACA, and the betweenness centrality was significantly higher in the VPL. These results indicate that normal connection was decreased in S1, and the pain matrix that includes ACA and VPL is complicated by neuropathic pain.This can be used to improve our understanding of pathophysiological mechanisms and improve therapeutic efficacy.
Acknowledgements
This research is partially supported by the program for Brain Mapping by Integrated Neurotechnologies for Disease Studies (Brain/MINDS) from Japan Agency for Medical Research and development, AMED.This work was supported by JSPS KAKENHI Grant Number JP16K10842.The authors would like to thank prof. Olaf Sporns for graph theoretical analysis. And we also thank Ms. C. Yamana for technical assistance with the experiments.References
1. Whitfield-Gabrieli, S. & Nieto-Castanon, A. Conn: a functional connectivity toolbox for correlated and anticorrelated brainnetworks. Brain Connect 2, 125–141, doi: 10.1089/brain.2012.0073 (2012).
2. Stafford, J. M. et al. Large-scale topology and the default mode network in the mouse connectome. Proceedings of the NationalAcademy of Sciences 111, 18745–18750, doi: 10.1073/pnas.1404346111 (2014).
3. Oh, S. W. et al. A mesoscale connectome of the mouse brain. Nature 508, 207–214, doi: 10.1038/nature13186 (2014).
4. Lein, E. S. et al. Genome-wide atlas of gene expression in the adult mouse brain. Nature 445, 168–176, doi: 10.1038/nature05453(2007).
5. Xia, M., Wang, J. & He, Y. BrainNet Viewer: A Network Visualization Tool for Human Brain Connectomics. PloS one 8, doi: 10.1371/journal.pone.0068910 (2013).
6. Rubinov, M. & Sporns, O. Complex network measures of brain connectivity: uses and interpretations. Neuroimage 52, 1059–1069,doi: 10.1016/j.neuroimage.2009.10.003 (2010).
Figures
