Curry 7

Scalp topographic maps of LEPs were analyzed using MATLAB (The Mathworks, Inc., Natick, MA) and the open source toolbox EEGLAB (Swartz Center for Computational Neuroscience, La Jolla, CA). Contour plots of the grand average of LEPs were created with intervals of 30 ms.
To further understand the possible source distribution of LEPs, the spatiotemporal source model was used to assess equivalent current dipoles of LEPs using Curry 7 software (NeuroScan, Inc., USA) [18 (link)]. Regional dipole fitting was used to fit the signal. The boundary element method for a realistic head model based on the MNI averaged structure image was used. According to previous reports on the dipole source analysis of LEPs [18 (link), 19 (link)], a four-dipole model was calculated. Residual variance is the percentage of data that cannot be explained by the model. During the optimization process for the four-dipole model of LEPs, a criterion for the residue variance of <10 % was used [19 (link)]. In the present study, 4-W LEPs were selected for dipole analysis because they were the greatest and consistently appeared for all subjects.

The MMII workflow was carried out on FDA-approved software platform Curry 7 (Compumedics Neuroscan^™, Charlotte, NC, USA). The data structure was organized as databases, with each modality stored as one data folder/file within the database. DICOM images with different scanners were in general compatible with the platform, making it the most universal image format for communication among different vendors/modalities. When each modality was imported for the first time, an initialization process was needed to set the parameters, followed by image coregistration. The parameters and procedures optimized for each modality are detailed in the following sections. The workflow runs on a regular PC; the 3D rendering of cortical surface runs more smoothly with advanced graphics card.

Acquisition of the EEG and EMG data was performed in Curry 7 (Compumedics Neuroscan, Charlotte, NC). The following procedures were applied online to the continuous EEG data: re-referencing to the bilateral mastoid electrodes; high- and low-pass filters at 1 and 100 Hz, respectively; a notch filter at 60 Hz; and baseline correction (using the first 3 s of data acquired). Artifact reduction was also performed online via principal component analysis (PCA) as implemented in Curry 7, using a threshold of ± 360 mV at both vertical and horizontal occular electrodes to identify eye blinks and movements, attenuating the first component within a window of −200 to 500 ms relative to the peak of the detected artifact.

Participants were seated in a comfortable chair in a dimly lit and electrically shielded room. Participants were instructed to stay relaxed and to avoid any body movements throughout the experiment. EEG data were recorded at a sampling rate of 1000 Hz with an online filter of 0.05–100 Hz using 64 Ag/AgCl electrodes based on the modified international 10–20 system (NeuroScan SynAmps2 (Compumedics USA, El Paso, TX, USA)). The electrodes at the linked mastoid sites served as reference electrodes, and the ground electrode was placed between the FPz and Fz electrode sites. Horizontal electrooculogram (EOG) signals were recorded with electrodes at the outer canthus of each eye, and vertical EOG signals were obtained from above and below the left. Gross artifacts, such as movement artifacts, were visually monitored and removed by an expert. Eye movements and eye blinks were removed using a regression procedure implemented in Curry 7 (Compumedics, Charlotte, NC) [35 (link)]. Epochs with EEG signals exceeding ±100 μV were excluded from further analysis. Electrode resistance was maintained below 5 kΩ. The schematic flow of this study is shown in Figure 1.