AUTHOR=Wei Xiaoya , Wang Liqiong , Yu Fangting , Lee Chihkai , Liu Ni , Ren Mengmeng , Tu Jianfeng , Zhou Hang , Shi Guangxia , Wang Xu , Liu Cun-Zhi 

TITLE=Identifying the neural marker of chronic sciatica using multimodal neuroimaging and machine learning analyses

JOURNAL=Frontiers in Neuroscience

VOLUME=Volume 16 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/neuroscience/articles/10.3389/fnins.2022.1036487

DOI=10.3389/fnins.2022.1036487

ISSN=1662-453X

ABSTRACT=Sciatica is a neuropathic pain disorder often caused by the herniated disc compressing the lumbosacral nerve roots. Neuroimaging studies have identified functional abnormalities in patients with chronic sciatica (CS). However, few studies have investigated the neuropathology of CS using brain structural or topological properties, and the classification value of multidimensional neuroimaging features in CS patients is unclear. Here, structural and resting-state functional MRI was acquired for 34 CS patients and 36 matched healthy controls (HC). We analyzed cortical surface area, cortical thickness, low-frequency fluctuation amplitude (ALFF), regional homogeneity (REHO), degree centrality (DC) of topological properties, functional connectivity (FC) among networks, and assessed the correlation between neuroimaging measures and clinical scores. Finally, the multimodal neuroimaging features were used to differentiate the CS patients and HC individuals by support vector machine (SVM) algorithm.  CS patients had a larger cortical surface area in the right banks of the superior temporal sulcus and rostral anterior cingulate; higher ALFF value in the left inferior frontal gyrus and middle frontal gyrus, lower REHO value in the left thalamus; reduced DC in the anterior prefrontal cortex and thalamus; weaker FC in the frontoparietal network and enhanced FCs between somatomotor and ventral attention network. Only DC and FC (basal ganglia and precentral) values were associated with clinical pain or mental scores. Furthermore, the above four types of multimodal neuroimaging features and SVM algorithm could classify CS patients and HC with an accuracy of 94.29%. Together, our findings revealed extensive reorganization of local functional properties, surface area, and network metrics in CS patients. The success of the patient identification highlights the potential of using artificial intelligence and multimodal neuroimaging markers in chronic pain research.