Computational protocol: Altered attentional control over the salience network in complex regional pain syndrome

Similar protocols

Protocol publication

[…] All brain imaging data were acquired using a 3.0 Tesla MR scanner (Skyra, Siemens, Erlangen, Germany). Resting-state fMRI data were obtained with a T2*-weighted echo planar imaging sequence using the following parameters: repetition time (TR) = 3,000 ms; echo time (TE) = 20 ms; flip angle (FA) = 90°; field of view (FOV) = 192 mm2; slice thickness = 3 mm; 120 volumes; 48 slices. During the resting-state fMRI scan, participants were instructed to keep their eyes closed, not to fall asleep, think of nothing in particular, and let their mind wander freely. For co-registration with the fMRI data, high-resolution T1-weighted structural images were obtained with the following acquisition parameters: TR = 1,900 ms; TE = 2.49 ms; FA = 9°; FOV = 230 mm2; slice thickness = 0.9 mm; 208 contiguous sagittal slices.Functional image data preprocessing was performed using the modules contained within the FMRIB Software Library tools (FSL, The standard preprocessing steps consisted of motion correction using multi-resolution rigid body co-registration, brain extraction using the FSL Brain Extraction Tool (BET), spatial smoothing with a Gaussian kernel of full width at half maximum of 5 mm, and high-pass filtering at 0.01 Hz. Functional image data of each individual was first co-registered to the corresponding T1-weighted image. These co-registered images were further linearly registered to the Montreal Neurological Institute (MNI) 152 template using affine transformation with 12 degrees of freedom. There were no differences in head motion parameters between the two groups (absolute head motion, the CRPS group 0.145 ± 0.054 mm, the control group, 0.125 ± 0.045, t = 1.56, P = 0.12; relative head motion, the CRPS group 0.086 ± 0.044 mm, the control group, 0.076 ± 0.032, t = 1.01, P = 0.32).Single-subject independent component analysis (ICA) was applied to identify the structural artifacts in each functional image data as implemented in the Multivariate Exploratory Linear Optimized Decomposition into Independent Components (MELODIC),. Afterwards, FMRIB’s ICA-based Xnoiseifier (FIX) was used to remove components corresponding to structural artifacts from each functional image data set.In order to obtain group-level RSNs, group ICA - a model-free and data-driven approach - was implemented to decompose the preprocessed four-dimensional functional images into three-dimensional spatial maps and one-dimensional time series,. In the current study, functional image data was decomposed into 25 independent components with a temporal concatenation approach. Consequently, 18 components were classified as anatomically and functionally meaningful RSNs corresponding to the functional networks previously described, and 7 components were classified as artifacts by visual inspection of an experienced researcher (S. Y.). Of these 18 identified RSNs, we selected 5 RSNs of interest, which were the salience, sensorimotor, and default mode (anterior and posterior) networks to represent the pain-related RSNs and the right frontoparietal network as the attention network. The abovementioned RSNs of interest were used in subsequent analyses. Component information and spatial maps of all available components, which were thresholded at a level of z = 3.0 (P = 0.001) are presented in Supplementary Figure.A dual regression algorithm was applied to estimate subject-specific time courses and spatial maps, which corresponded to the RSNs of interest derived from the initial group ICA. In the first step of dual regression, the average subject-specific time series of the RSNs of interest was derived using a linear model fit of the RSNs of interest against each individual’s functional data. The second step provided the subject-specific spatial maps for the RSNs of interest, which reflect the degree of synchronization by the temporal regression against each individual’s functional data.Using the FSLNets (, the temporal correlation coefficients between the attention network and pain-related RSNs, including the salience, sensorimotor, and default mode networks were computed based on individual time series of the RSNs. Correlation coefficients between the RSNs, which reflect the strength of inter-network connections, were Fisher z-transformed and were then used for subsequent analyses. […]

Pipeline specifications

Software tools FSL, BET
Application Magnetic resonance imaging
Organisms Homo sapiens