Computational protocol: Test-retest reliability of fMRI experiments during robot-assisted active and passive stepping

Similar protocols

Protocol publication

[…] All fMRI datasets were analysed using SPM8 (Wellcome Department of Cognitive Neurology, London, UK, running on Matlab 2012b (Mathworks, Inc., Natick, MA, USA, The first three volumes prior to the first task-block were removed from each run. In spatial preprocessing the remaining 90 volumes were firstly realigned to their mean image and unwarped to remove residual head motion related variance and image distortions along air-tissue boundaries []. Secondly, all data from t2 was coregistered to the mean image of the respective condition at t1. Thirdly, all images were normalised into standard MNI space using to the EPI-template provided by the Montreal Neurological Institute, re-sliced to a voxel size of 2 × 2 × 2 mm3, and finally all data was spatially smoothed (FWHM = 8 mm). The estimated realignment parameter data from the realignment step were filtered using the discrete cosine transform matrix filter (cut off at 128 s) incorporated in SPM8, to remove linear baseline drifts. Only data from participants whose estimated head motion parameters were below a stringent threshold of ½ voxel size after filtering in every spatial dimension in both conditions and at both experimental sessions were included in the subsequent statistical analysis. In the 1st-level analysis the data from t1 and t2 were modelled as two separate task regressors in the same general linear model (GLM) for each movement condition individually []. Two additional regressors were added to the model for each session to account for the T1-decay along consecutive volumes []. A high pass filter (cut off at 128 s) was used to remove slow signal drifts. To account for the sparse-sampling fMRI scheme, data taken during each trial was modelled using a boxcar function (1st-order, window length 3 x TR (i.e., 9.075 s)) []. Contrast images for each task regressor were calculated to reveal task-related activation at t1 and t2. To estimate the task-related effects at the group level, all contrast images of a specific task from the 1st-level analysis were subject to individual one-sample t-tests. Planned comparisons were computed, in order to test for significant differences between t1 and t2. The resulting activation maps were limited to a cluster-corrected voxel threshold of p < 0.001 (spatial extent: k ≥ 42 contiguous voxels) [, ]. The cluster threshold method was applied to control for the overall type I error. Anatomical correlates of clusters of activation were determined with the help of probabilistic cytoarchitectonic maps implemented in the Anatomy toolbox []. This toolbox was also used to define bilateral anatomical regions of interest (ROI) in the primary motor cortex (M1), primary somatosensory cortex (S1), secondary somatosensory cortex (S2) and the cerebellum Table . The ROI covering M1 was built by combining BA 4a and 4b []. BAs 1, 2, 3a and 3b served to create the ROI in S1 [–]. The ROI covering S2 was built by combining areas Operculum (OP) 1, OP2, OP3 and OP4 in the parietal operculum [, ]. The ROI located in the cerebellum was created by combining the lobules I to X (lobes and vermis) included in the Anatomy toolbox []. A ROI covering SMA was built from the automated anatomical labeling atlas [] using the WFU_pickatlas toolbox []. These specific ROIs were selected as these areas have repeatedly been reported to be involved in lower limb motor control in previous studies [, , –]. […]

Pipeline specifications

Software tools SPM, AAL
Applications Magnetic resonance imaging, Functional magnetic resonance imaging
Organisms Homo sapiens
Diseases Head and Neck Neoplasms, Metabolism, Inborn Errors, Spinal Cord Injuries, Stroke