Fig 1.
A. The model predicts brain activity in MEG and EEG sensors based on simulations of propagating activity in primary visual area (right and left V1). First, brain activity is modeled in V1 using participants’ individual retinotopic maps obtained with anatomical and functional MRI. Second, this activity is projected onto MEG and EEG sensors. B. To test the model, we recorded simultaneous MEG-EEG activity when participants were presented with stimuli designed to elicit specific patterns in V1. There were three types of stimuli: traveling out, traveling in and standing. C. Experimental protocol. Each stimulus was presented for 2 s, following a 2s-grey screen (8 trials per run). At the end of each run, participants were asked to report if they had seen the fixation point change color. VM: vertical meridian. HM: horizontal meridian.
Fig 2.
Visual stimuli designed to induce traveling waves in V1.
A. Each stimulus corresponds to the multiplication of a static, black and white pattern (carrier) with a contrast modulator. The modulator is adjusted for cortical magnification, i.e., higher spatial frequency at the fovea (see zoomed area in blue). B. Example stimuli presented at four time points for each condition (rows: traveling out, traveling in, standing), with the corresponding patterns simulated in V1, used as model inputs (bottom right insets). A is the maximal amplitude of the simulated signal (A = 0.5).
Fig 3.
Measured and predicted time courses.
A. Time courses of magnetometer (MAG) data (top) and model (bottom) for each condition (traveling out in yellow, traveling in in red and standing in blue) for an occipital sensor (highlighted in white in the topomaps). Topomaps for each condition are plotted at the time point indicated by the vertical dashed line below the topomaps. Note that amplitude values in the model are arbitrary (see Methods). B. Time courses of EEG signal. Same convention as in A.
Fig 4.
The model captures phase and amplitude relations between sensors.
A. Topomaps of amplitude and phase at 5 Hz for MAG (top) and EEG sensors (bottom), for both data (left) and model (right). Polar plots show the phase of selected sensors, along horizontal and vertical meridians. B. All datasets are compared either with the “match” model (e.g., traveling out model for traveling out condition) or with “cross” (unmatched) models. The correlation coefficients for matched comparisons (R match) are plotted against the coefficient for crossed comparisons (R cross), for MAG (top) and EEG sensors (bottom). Each black dot is one participant. The red dot is the average across participants, error bars are standard error of the mean. Dots above the bisector (black dashed line) mean that the matched models are more accurate than the cross models. The red dashed line corresponds to the permutation-based null threshold.
Fig 5.
The model is specific to the stimulus-induced traveling waves.
A. Using paired corrected t-tests, we compared the correlation coefficients between matched and crossed comparisons, for different values of temporal and spatial frequency. The bold outline bars correspond to the temporal (Ft) and spatial (Fs) frequency values used in the experimental conditions. Stars indicate significant p-values (*** p < 0.001, ** p < 0.01, * p < 0.05). Error bars represent the Standard Error of the Mean (SEM). B. Paired t-tests on coefficients between matched and crossed comparisons were performed separately for each condition.