Computational protocol: Neural Correlates of Mirror Visual Feedback-Induced Performance Improvements: A Resting-State fMRI Study

Similar protocols

Protocol publication

[…] Analysis of the rs-fMRI data was performed with the FMRIB Software Library (FSL 5.0), utilizing the independent component analysis (ICA)-AROMA pipeline, the fastECM toolbox (Wink et al., ) and SPM12 running in MATLAB version 8.6. First, the T1-weighted images were segmented using SPM12, and a skull-stripped version was created with fslmaths. Second, standard motion correction, spatial smoothing with a Gaussian kernel of 6 mm FWHM and linear registration of functional and T1-weighted images to each other and to the MNI space were performed. To further detect and remove motion-related artifacts ICA-AROMA (Pruim et al., ,) was used to perform an ICA on the functional data to identify and remove head motion related components by employing predefined temporal (high frequency content and maximum correlation) as well as spatial features (edge and cerebrospinal fluid fraction). Subsequently, additional nuisance correction was performed by regressing out signal from white matter and cerebrospinal fluid (physiological noise; Fox and Raichle, ) and by high pass filtering (0.01 Hz). Within the preprocessing, the functional data was also resampled at a resolution of 2 × 2 × 2 mm3 as this is the standard in most fMRI analyses.For the analysis of functional connectivity, the eigenvector centrality maps (ECM) approach was used. Eigenvector centrality can quantify the relative importance, or centrality, of an individual node on a network as a whole. The centrality is high if a node is connected to many nodes that themselves are “central” (Lohmann et al., ). Hereby, the importance of points in brain networks can be measured and visualized (Lohmann et al., ). Centrality analyses are supposed to enable the interpretability of connectivity matrices used in graph analyses combined with the high spatial resolution of voxel-based methods (Wink et al., ). ECM, in particular, were used for our data analysis for its exploratory whole brain approach independently from predefined seed regions (Lohmann et al., ). Using the preprocessed functional data, an ECM analyses was performed for each subject and each scan time point (rs-fMRI_pre and rs-fMRI_post) separately using the fastECM toolbox (Wink et al., ). The fastECM algorithm was chosen for its shorter computation times and lower storage requirements for high-resolution fMRI data. Within the preprocessed functional images, the voxel-wise connectivities between all pairs of voxels were computed with the fastECM program within a study-specific gray matter mask. To create this study-specific mask, the individual gray matter masks derived from segmenting the T1-weighted images were added up. Then a threshold was applied to include only voxels that contained gray matter from all participants (total of 202519 voxel). Hence, a 3D voxel-wise ECM per subject and scan time point was created and used for further statistical analyses.All statistical analyses were performed using SPM12. To assess the effect of MVF on the rs-FC single-subject ECM images were compared using a 2 × 2 flexible factorial design with the factors TIME (rs-fMRI_pre vs. rs-fMRI_post) and GROUP (MG vs. CG), and complemented with subsidiary t-tests. To analyze a potential relationship between changes in functional connectivity and behavioral improvements, a correlation analysis was performed with SPM12. First, ECM difference images (rs-fMRI_post minus rs-fMRI_pre) were created on the single subject level using the ImCalc-toolbox in SPM12. We then performed a correlation analysis per group correlating the difference images with the performance improvement of the untrained LH.All stated findings are significant at p < 0.05 with cluster-wise (p = 0.001) FWE correction for multiple comparisons. […]

Pipeline specifications

Software tools FSL, SPM
Application Magnetic resonance imaging
Organisms Homo sapiens
Diseases Brain Diseases