Cortical source estimation of ssVEF responses
Before the ssVEF responses were extracted from the averaged MEG data epochs, these evoked MEG responses were transformed into cortical source space by a weighted minimum norm estimation (wMNE) approach (Gramfort et al. 2014) using the brainstorm toolbox (https://neuroimage.usc.edu/brainstorm/). The source reconstruction of evoked MEG data epochs was based on a forward model based on an overlapping spheres head model (Huang et al. 1999) using a canonical cortical mesh (3003 vertices) derived from the ICBM152 template (Mazziotta et al. 2001). Thereby, the individual head and sensor positions of each subject were co-registered with the template brain by realigning the individual with the template brain’s fiducials and minimizing the mean distance between the individual head shape points and the template brain scalp surface (Moratti et al. 2011). Thereafter, the forward model was calculated by using a head model based on overlapping spheres (Huang et al. 1999). Finally, a weighted Minimum Norm Estimation (wMNE) (Gramfort et al. 2014) was used to calculate the current density of the evoked MEG averages (see above) on the cortical surface for later analysis. Finally, for illustration purposes and later statistical analysis, regions of interests (ROIs) were defined by overlaying a co-registered Brodmann Atlas as implemented in Brainstorm (Tadel et al., 2011;https://neuroimage.usc.edu/brainstorm/). The atlas contains V1, V2, and hMT+ areas and were extended by adding inferior-temporal visual cortex and the occipito-parietal brain area (see Figure 1 below).
In order to extract the ssVEF responses in cortical source space, the time series of the cortical current densities at each vertex of the cortical surface were demeaned and the 12 Hz and 15 Hz Fourier components were determined by a Fast Fourier Transform (FFT) using the fieldtrip toolbox (Oostenveld et al. 2011);https://www.fieldtriptoolbox.org). For the pre-cue baseline conditions (attend the central fixation cross vs. attend both peripheral rings), the 1 s epochs were zero padded to 10 s in order to obtain a frequency resolution of 0.1 Hz for the FFT. For the post-cue 2.9 s epochs overlapping four 1 s intervals were extracted (0.5 s to 1. 5 s, 1 s to 2 s, 1.5 s to 2.5 s and 2 s to 2.9 s). Each interval was zero padded to 10 s in order to obtain a 0.1 Hz frequency resolution and submitted to FFT analysis. Then, the obtained FFTs were averaged. Finally, for all FFTs (pre-cue baseline and post-cue intervals) the 12 Hz and 15 Hz power of the corresponding Fourier components were determined at each vertex of the canonical brain surface mesh.