DATA_TASK_3DMOV_HP_CSF_WD

Briefly, data comes from five functional runs consisting of a resting-state measurement (eyes closed), four individual task measurements including a visual n-back (n=2) task (Kirchner, 1958), the Eriksen flanker task (Eriksen & Eriksen, 1974), a mental rotation task (Shepard & Metzler, 1971), and a verbal odd-man-out task (Flowers & Robertson, 1985). All runs comprise 192 data points with tasks being continuously performed during this period. For the n-back and flanker task, stimuli were presented at a rate of 0.5 Hz; for the mental rotation and odd-man out tasks they were presented at a rate of 0.25 Hz. Task sequence was counterbalanced across participants with the exception that the resting state functional run was always acquired first to prevent carry-over effects (Grigg & Grady, 2010). The data were acquired using a 3 Tesla Siemens Prisma Fit (upgraded Tim Trio) scanner and a 64-channel head coil. Initial preprocessing was performed using BrainVoyager QX (v2.6; Brain Innovation, Maastricht, the Netherlands). This includes slice scan time correction, 3D-motion correction, high-pass filtering with a frequency cutoff of .01 Hz, and registration of functional and anatomical images. Subsequently, using MATLAB (2013a, The MathWorks,Natick, MA), signals were cleaned by performing wavelet despiking (Patel & Bullmore, 2015) and regressing out a global noise signal given by the first principal component of signals observed within the cerebrospinal fluid of the ventricles. Next, voxels were uniquely assigned to one of the 68 cortical ROIs specified by the DK atlas and an average BOLD time-series was computed for each region as the mean time-series over all voxels of that region.