Computational protocol: Magnetoencephalographic study of event‐related fields and cortical oscillatory changes during cutaneous warmth processing

Similar protocols

Protocol publication

[…] For analysis of the MEG data, we used Fieldtrip for data preprocessing and time‐frequency analysis (Oostenveld, Fries, Maris, & Schoffelen, ; and Brainstorm for source analysis (Tadel, Baillet, Mosher, Pantazis, & Leahy, ). Open source toolboxes were run using a Matlab (The Mathworks, Natick, Massachusetts, USA) environment. Continuous MEG data were band‐pass filtered between 0.02 and 100 Hz and segmented from −3 to 5 s with respect to the stimulus onset. To reject artifacts due to eye blinks, eye movements, and heart beats, we applied an independent component analysis method (“runica” implemented in FieldTrip, Independent components representing the ocular and cardiac activities were identified by visual inspection based on their topography and their time‐course. The artifact‐rejected waveforms were recovered by back‐projection of the rest independent components into the signal‐space after eliminating these artifact components. We selected trials where the participants pressed the “yes” button and applied the 40‐Hz low‐pass filter. The trials with muscle artifacts were rejected. Further, we discarded data from 6 participants who had <120 trials. Data from 24 participants in total (13 women and 11 men) were analyzed. The relative head positions of the MEG sensors were different across sessions. The MEG data were interpolated to standard sensor locations, which was obtained by averaging the sensor positions over all the sessions across subjects by using the “ft_megrealign” function from the Fieldtrip.To obtain ERFs evoked by the warmth stimulation, we averaged all the trials from the three successive sessions, preprocessed artifact‐free trials for each participant, and examined the grand‐average for the entire participants. The baseline was selected from −1.1 to −0.1 s prior to the stimulus onset. To obtain the source activity of warmth‐related ERFs in the brain region, we performed the weighted minimum norm estimates (wMNE) (Hamalainen & Ilmoniemi, ; Hauk, ; Lin et al., ), which is implemented in Brainstorm toolbox (Tadel et al., ). We used the ICBM152 template anatomy, which was scaled according to each participant's individual head shapes, in the Brainstorm toolbox. The baseline period (−1.1 to −0.1 s) was used to estimate noise‐covariance for each session of each subject. The wMNE source analysis was performed on an overlapping‐sphere head model with standard Tikhonov regularization (λ = 0.1).Before the time–frequency analysis, we applied the planar transformation for easier interpretation of MEG signals with complicated spatial patterns. Preprocessed data based on axial gradiometer was transformed to planar gradient sensors, using “ft_megplanar” and “ft_combineplanar” functions from the fieldtrip toolbox, to locate oscillatory brain activation more easily. We calculated time–frequency representations (TFRs) at 2–40 Hz, using a sliding Hanning‐window Fourier transform approach with a fixed 500‐ms time window, moving in steps of 10 ms. The results of TFRs were expressed as percent power change relative to baseline (i.e., −1.1 to −0.1 s). The TFRs of each sensor were grand‐averaged across participants.For cortical mapping of oscillatory power changes, we filtered the wMNE source data in specific frequency ranges (i.e., 1–4 Hz, 4–8 Hz, 8–13 Hz, and 18–23 Hz, for delta, theta, alpha, and beta, respectively). Thereafter, we applied the Hilbert transformation to obtain power and phase information in the cortical sources. […]

Pipeline specifications

Software tools FieldTrip, Brainstorm, EEGLAB
Application Clinical electrophysiology
Organisms Homo sapiens